[
  {
    "context": "The CSV file stent365.csv summarizes an experiment that studies effectiveness of stents in treating patients at risk of stroke with some unexpected results. These data represent the results 365 days after stroke.",
    "question": "Compute the proportion of patients in the treatment group who had a stroke by the end of their first year. Please round to the nearest hundredth.",
    "answer": "0.20",
    "data": [
      "/data/qrdata/data/stent365.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 1.1",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A migraine is a particularly painful type of headache, which patients sometimes wish to treat with acupuncture. To determine whether acupuncture relieves migraine pain, researchers conducted a randomized controlled study where 89 females diagnosed with migraine headaches were randomly assigned to one of two groups: treatment or control. 43 patients in the treatment group received acupuncture that is specifically designed to treat migraines. 46 patients in the control group received placebo acupuncture (needle insertion at non-acupoint locations). 24 hours after patients received acupuncture, they were asked if they were pain free. The research data is in the CSV file migraine.csv.",
    "question": "What proportion of patients in the treatment group were pain free 24 hours after receiving acupuncture? Please round to the nearest hundredth.",
    "answer": "0.23",
    "data": [
      "/data/qrdata/data/migraine.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 1.1a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Researchers studying the effect of antibiotic treatment for acute sinusitis compared to symptomatic treatments randomly assigned 166 adults diagnosed with acute sinusitis to one of two groups: treatment or control. Study participants received either a 10-day course of amoxicillin (an antibiotic) or a placebo similar in appearance and taste. The placebo consisted of\nsymptomatic treatments such as acetaminophen, nasal decongestants, etc. At the end of the 10-day period, patients were asked if they experienced improvement in symptoms. The research data is in the CSV file sinusitis.csv.",
    "question": "What proportion of patients in the treatment group experienced improvement in symptoms? Please round to the nearest hundredth.",
    "answer": "0.78",
    "data": [
      "/data/qrdata/data/sinusitis.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 1.2a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Researchers studying the effect of antibiotic treatment for acute sinusitis compared to symptomatic treatments randomly assigned 166 adults diagnosed with acute sinusitis to one of two groups: treatment or control. Study participants received either a 10-day course of amoxicillin (an antibiotic) or a placebo similar in appearance and taste. The placebo consisted of\nsymptomatic treatments such as acetaminophen, nasal decongestants, etc. At the end of the 10-day period, patients were asked if they experienced improvement in symptoms. The research data is in the CSV file sinusitis.csv.",
    "question": "What proportion of patients in the control group experienced improvement in symptoms? Please round to the nearest hundredth.",
    "answer": "0.80",
    "data": [
      "/data/qrdata/data/sinusitis.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 1.2b",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Researchers studying the effect of antibiotic treatment for acute sinusitis compared to symptomatic treatments randomly assigned 166 adults diagnosed with acute sinusitis to one of two groups: treatment or control. Study participants received either a 10-day course of amoxicillin (an antibiotic) or a placebo similar in appearance and taste. The placebo consisted of\nsymptomatic treatments such as acetaminophen, nasal decongestants, etc. At the end of the 10-day period, patients were asked if they experienced improvement in symptoms. The research data is in the CSV file sinusitis.csv.",
    "question": "In which group did a higher percentage of patients experience improvement in symptoms? Please answer with \"treatment group\" or \"control group\".",
    "answer": "control group",
    "data": [
      "/data/qrdata/data/sinusitis.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 1.2c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "treatment group",
        "control group"
      ]
    }
  },
  {
    "context": "A survey was conducted to study the smoking habits of UK residents. The data is in the CSV file smoking.csv.",
    "question": "How many participants were included in the survey?",
    "answer": "1691",
    "data": [
      "/data/qrdata/data/smoking.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 1.10b",
      "keywords": [
        "Statistics",
        "Data analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Daily air quality is measured by the air quality index (AQI) reported by the Environmental Protection Agency. This index reports the pollution level and what associated health effects might be a concern. The index is calculated for five major air pollutants regulated by the Clean Air Act and takes values from 0 to 300, where a higher value indicates lower air quality. AQI was reported for a sample of 91 days in 2011 in Durham, NC. The data is in the CSV file pm25_2011_durham.csv.",
    "question": "Estimate the median AQI value of this sample. Please round to the nearest integer.",
    "answer": "30",
    "data": [
      "/data/qrdata/data/pm25_2011_durham.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.11a",
      "keywords": [
        "Statistics",
        "Median"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Daily air quality is measured by the air quality index (AQI) reported by the Environmental Protection Agency. This index reports the pollution level and what associated health effects might be a concern. The index is calculated for five major air pollutants regulated by the Clean Air Act and takes values from 0 to 300, where a higher value indicates lower air quality. AQI was reported for a sample of 91 days in 2011 in Durham, NC. The data is in the CSV file pm25_2011_durham.csv.",
    "question": "Estimate the interquartile range of AQI in this sample. Please round to the nearest hundredth.",
    "answer": "19.50",
    "data": [
      "/data/qrdata/data/pm25_2011_durham.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.11c",
      "keywords": [
        "Statistics",
        "Interquartile range"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "910 randomly sampled registered voters from Tampa, FL were asked if they thought workers who have illegally entered the US should be (i) allowed to keep their jobs and apply for US citizenship, (ii) allowed to keep their jobs as temporary guest workers but not allowed to apply for US citizenship, or (iii) lose their jobs and have to leave the country. The results of the survey by political ideology are in the CSV file immigration.csv.",
    "question": "What proportion of these Tampa, FL voters identify themselves as conservatives? Please round to the nearest hundredth.",
    "answer": "0.41",
    "data": [
      "/data/qrdata/data/immigration.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.22a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "910 randomly sampled registered voters from Tampa, FL were asked if they thought workers who have illegally entered the US should be (i) allowed to keep their jobs and apply for US citizenship, (ii) allowed to keep their jobs as temporary guest workers but not allowed to apply for US citizenship, or (iii) lose their jobs and have to leave the country. The results of the survey by political ideology are in the CSV file immigration.csv.",
    "question": "What proportion of these Tampa, FL voters identify themselves as conservatives and are in favor of the citizenship option? Please round to the nearest hundredth.",
    "answer": "0.06",
    "data": [
      "/data/qrdata/data/immigration.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.22c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "910 randomly sampled registered voters from Tampa, FL were asked if they thought workers who have illegally entered the US should be (i) allowed to keep their jobs and apply for US citizenship, (ii) allowed to keep their jobs as temporary guest workers but not allowed to apply for US citizenship, or (iii) lose their jobs and have to leave the country. The results of the survey by political ideology are in the CSV file immigration.csv.",
    "question": "What proportion of these Tampa, FL voters who identify themselves as conservatives are also in favor of the citizenship option? Please round to the nearest hundredth.",
    "answer": "0.15",
    "data": [
      "/data/qrdata/data/immigration.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.22d",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Rosiglitazone is the active ingredient in the controversial type 2 diabetes medicine Avandia and has been linked to an increased risk of serious cardiovascular problems such as stroke, heart failure, and death. A common alternative treatment is pioglitazone, the active ingredient in a diabetes medicine called Actos. In a nationwide retrospective observational study of 227,571 Medicare beneficiaries aged 65 years or older, it was found that 2,593 of the 67,593 patients using rosiglitazone and 5,386 of the 159,978 using pioglitazone had serious cardiovascular problems. The data is in the CSV file avandia.csv.",
    "question": "Determine if the following statement is true or false. Be careful: The reasoning may be wrong even if the statement's conclusion is correct. In such cases, the statement should be considered false. Answer with \"true\" or \"false\".\nStatement: The data suggest that diabetic patients who are taking rosiglitazone are more likely to have cardiovascular problems since the rate of incidence was (2,593 / 67,593 = 0.038) 3.8% for patients on this treatment, while it was only (5,386 / 159,978 = 0.034) 3.4% for patients on pioglitazone.",
    "answer": "true",
    "data": [
      "/data/qrdata/data/avandia.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.25a2",
      "keywords": [
        "Statistics",
        "Data analysis"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "true",
        "false"
      ]
    }
  },
  {
    "context": "Rosiglitazone is the active ingredient in the controversial type 2 diabetes medicine Avandia and has been linked to an increased risk of serious cardiovascular problems such as stroke, heart failure, and death. A common alternative treatment is pioglitazone, the active ingredient in a diabetes medicine called Actos. In a nationwide retrospective observational study of 227,571 Medicare beneficiaries aged 65 years or older, it was found that 2,593 of the 67,593 patients using rosiglitazone and 5,386 of the 159,978 using pioglitazone had serious cardiovascular problems. The data is in the CSV file avandia.csv.",
    "question": "What proportion of all patients had cardiovascular problems? Please round to the nearest thousandth.",
    "answer": "0.035",
    "data": [
      "/data/qrdata/data/avandia.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.25b",
      "keywords": [
        "Statistics",
        "Data analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Rosiglitazone is the active ingredient in the controversial type 2 diabetes medicine Avandia and has been linked to an increased risk of serious cardiovascular problems such as stroke, heart failure, and death. A common alternative treatment is pioglitazone, the active ingredient in a diabetes medicine called Actos. In a nationwide retrospective observational study of 227,571 Medicare beneficiaries aged 65 years or older, it was found that 2,593 of the 67,593 patients using rosiglitazone and 5,386 of the 159,978 using pioglitazone had serious cardiovascular problems. The data is in the CSV file avandia.csv.",
    "question": "If the type of treatment and having cardiovascular problems were independent, about how many patients in the rosiglitazone group would we expect to have had cardiovascular problems? Please round to the nearest integer.",
    "answer": "2370",
    "data": [
      "/data/qrdata/data/avandia.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.25c",
      "keywords": [
        "Statistics",
        "Data analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The Stanford University Heart Transplant Study was conducted to determine whether an experimental heart transplant program increased lifespan. Each patient entering the program was designated an official heart transplant candidate, meaning that he was gravely ill and would most likely benefit from a new heart. Some patients got a transplant and some did not. The variable transplant indicates which group the patients were in; patients in the treatment group got a transplant and those in the control group did not. Of the 34 patients in the control group, 30 died. Of the 69 people in the treatment group, 45 died. Another variable called survived was used to indicate whether or not the patient was alive at the end of the study. The data is in the CSV file heart_transplant.csv.",
    "question": "How much higher is the proportion of deaths in the control group than in the treatment group? Please round to the nearest hundredth.",
    "answer": "0.23",
    "data": [
      "/data/qrdata/data/heart_transplant.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.26c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "This CSV file infmortrate.csv gives the number of deaths of infants under one year old in 2012 per 1,000 live births in the same year. This rate is often used as an indicator of the level of health in a country.",
    "question": "Estimate the first quartile (Q1) of the infant death rate. Please round to the nearest hundredth.",
    "answer": "6.51",
    "data": [
      "/data/qrdata/data/infmortrate.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.28a",
      "keywords": [
        "Statistics",
        "Quartile"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "This CSV file infmortrate.csv gives the number of deaths of infants under one year old in 2012 per 1,000 live births in the same year. This rate is often used as an indicator of the level of health in a country.",
    "question": "Estimate the third quartile (Q3) of the infant death rate. Please round to the nearest hundredth.",
    "answer": "42.14",
    "data": [
      "/data/qrdata/data/infmortrate.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 2.28a",
      "keywords": [
        "Statistics",
        "Quartile"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "This CSV file loans_full_schema.csv represents thousands of loans made through the Lending Club platform, which is a platform that allows individuals to lend to other individuals. The homeownership variable describes whether the borrower rents, has a mortgage, or owns her property.",
    "question": "Compute the probability a randomly selected loan from the data set is for someone who has a mortgage or owns her home. Please round to the nearest hundredth.",
    "answer": "0.61",
    "data": [
      "/data/qrdata/data/loans_full_schema.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.8c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "This CSV file loans_full_schema.csv represents thousands of loans made through the Lending Club platform, which is a platform that allows individuals to lend to other individuals. The homeownership variable describes whether the borrower rents, has a mortgage, or owns her property.",
    "question": "Determine the probability a randomly drawn loan from the data set is from a joint application where the couple had a mortgage. Please round to the nearest thousandth.",
    "answer": "0.095",
    "data": [
      "/data/qrdata/data/loans_full_schema.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.15a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Volunteer patients were randomized into one of two experiment groups where they would receive an experimental vaccine or a placebo. They were subsequently exposed to a drug-sensitive strain of malaria and observed to see whether they came down with an infection. The vaccine trial data is described in the CSV file malaria.csv.",
    "question": "What is the difference in infection rates between the placebo and vaccine groups? Please round to the nearest thousandth.",
    "answer": "0.643",
    "data": [
      "/data/qrdata/data/malaria.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 2.34",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file smallpox.csv represents a sample of 6,224 individuals from the year 1721 who were exposed to smallpox in Boston. Some of them had received a vaccine (inoculated) while others had not. Doctors at the time believed that inoculation, which involves exposing a person to the disease in a controlled form, could reduce the likelihood of death.",
    "question": "What is the probability a randomly selected person who was not inoculated died from smallpox? Please round to the nearest thousandth.",
    "answer": "0.141",
    "data": [
      "/data/qrdata/data/smallpox.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.31",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file smallpox.csv represents a sample of 6,224 individuals from the year 1721 who were exposed to smallpox in Boston. Some of them had received a vaccine (inoculated) while others had not. Doctors at the time believed that inoculation, which involves exposing a person to the disease in a controlled form, could reduce the likelihood of death.",
    "question": "Determine the probability that an inoculated person died from smallpox. Please round to the nearest thousandth.",
    "answer": "0.025",
    "data": [
      "/data/qrdata/data/smallpox.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.32",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A Pew Research poll asked 1,306 Americans \"From what you've read and heard, is there solid evidence that the average temperature on earth has been getting warmer over the past few decades, or not?\". The data is in the CSV file global_warming_pew.csv",
    "question": "What is the probability that a randomly chosen respondent believes the earth is warming or is a liberal Democrat? Please round to the nearest hundredth.",
    "answer": "0.62",
    "data": [
      "/data/qrdata/data/global_warming_pew.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.15b",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A Pew Research poll asked 1,306 Americans \"From what you've read and heard, is there solid evidence that the average temperature on earth has been getting warmer over the past few decades, or not?\". The data is in the CSV file global_warming_pew.csv",
    "question": "What is the probability that a randomly chosen respondent believes the earth is warming given that he is a liberal Democrat? Please round to the nearest hundredth.",
    "answer": "0.90",
    "data": [
      "/data/qrdata/data/global_warming_pew.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.15c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A Pew Research poll asked 1,306 Americans \"From what you've read and heard, is there solid evidence that the average temperature on earth has been getting warmer over the past few decades, or not?\". The data is in the CSV file global_warming_pew.csv",
    "question": "What is the probability that a randomly chosen respondent believes the earth is warming given that he is a conservative Republican? Please round to the nearest hundredth.",
    "answer": "0.33",
    "data": [
      "/data/qrdata/data/global_warming_pew.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.15d",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A Pew Research poll asked 1,306 Americans \"From what you've read and heard, is there solid evidence that the average temperature on earth has been getting warmer over the past few decades, or not?\". The data is in the CSV file global_warming_pew.csv",
    "question": "What is the probability that a randomly chosen respondent is a moderate/liberal Republican given that he does not believe that the earth is warming? Please round to the nearest hundredth.",
    "answer": "0.18",
    "data": [
      "/data/qrdata/data/global_warming_pew.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.15f",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A 2010 SurveyUSA poll asked 500 Los Angeles residents, \"What is the best hamburger place in Southern California? Five Guys Burgers? In-N-Out Burger? Fat Burger? Tommy's Hamburgers? Umami Burger? Or somewhere else?\". The data is in the CSV file burger.csv",
    "question": "What is the probability that a randomly chosen male likes In-N-Out the best? Please round to the nearest hundredth.",
    "answer": "0.65",
    "data": [
      "/data/qrdata/data/burger.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.17b",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A 2010 SurveyUSA poll asked 500 Los Angeles residents, \"What is the best hamburger place in Southern California? Five Guys Burgers? In-N-Out Burger? Fat Burger? Tommy's Hamburgers? Umami Burger? Or somewhere else?\". The data is in the CSV file burger.csv",
    "question": "What is the probability that a randomly chosen female likes In-N-Out the best? Please round to the nearest hundredth.",
    "answer": "0.72",
    "data": [
      "/data/qrdata/data/burger.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.17c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A 2010 SurveyUSA poll asked 500 Los Angeles residents, \"What is the best hamburger place in Southern California? Five Guys Burgers? In-N-Out Burger? Fat Burger? Tommy's Hamburgers? Umami Burger? Or somewhere else?\". The data is in the CSV file burger.csv",
    "question": "What is the probability that a randomly chosen person likes Umami best or that person is female? Please round to the nearest hundredth.",
    "answer": "0.51",
    "data": [
      "/data/qrdata/data/burger.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.17e",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Assortative mating is a nonrandom mating pattern where individuals with similar genotypes and/or phenotypes mate with one another more frequently than what would be expected under a random mating pattern. Researchers studying this topic collected data on eye colors of 204 Scandinavian men and their female partners. The data is in the CSV file assortative_mating.csv",
    "question": "What is the probability that a randomly chosen male respondent or his partner has blue eyes? Please round to the nearest hundredth.",
    "answer": "0.71",
    "data": [
      "/data/qrdata/data/assortative_mating.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.18a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Assortative mating is a nonrandom mating pattern where individuals with similar genotypes and/or phenotypes mate with one another more frequently than what would be expected under a random mating pattern. Researchers studying this topic collected data on eye colors of 204 Scandinavian men and their female partners. The data is in the CSV file assortative_mating.csv",
    "question": "What is the probability that a randomly chosen male respondent with blue eyes has a partner with blue eyes? Please round to the nearest hundredth.",
    "answer": "0.68",
    "data": [
      "/data/qrdata/data/assortative_mating.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.18b",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Assortative mating is a nonrandom mating pattern where individuals with similar genotypes and/or phenotypes mate with one another more frequently than what would be expected under a random mating pattern. Researchers studying this topic collected data on eye colors of 204 Scandinavian men and their female partners. The data is in the CSV file assortative_mating.csv",
    "question": "What is the probability of a randomly chosen male respondent with green eyes having a partner with blue eyes? Please round to the nearest hundredth.",
    "answer": "0.31",
    "data": [
      "/data/qrdata/data/assortative_mating.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 3.18c",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file fcid.csv contains a sample of heights from a survey in the USDA Food Commodity Intake Database. It approximates the height distribution in the US adult population.",
    "question": "What proportion of the sample is between 180 cm and 185 cm tall? Please round to the nearest thousandth.",
    "answer": "0.117",
    "data": [
      "/data/qrdata/data/fcid.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.72",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file fcid.csv contains a sample of heights from a survey in the USDA Food Commodity Intake Database. It approximates the height distribution in the US adult population.",
    "question": "Three US adults are randomly selected. What is the probability that all three are between 180 and 185 cm tall? Please round to the nearest thousandth.",
    "answer": "0.002",
    "data": [
      "/data/qrdata/data/fcid.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.73a",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file fcid.csv contains a sample of heights from a survey in the USDA Food Commodity Intake Database. It approximates the height distribution in the US adult population.",
    "question": "Three US adults are randomly selected. What is the probability that none are between 180 and 185 cm tall? Please round to the nearest hundredth.",
    "answer": "0.69",
    "data": [
      "/data/qrdata/data/fcid.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 3.73b",
      "keywords": [
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ami_occurrences.csv contains occurrences of acute myocardial infarction (AMI) on 365 days in New York City.",
    "question": "Use a Poisson distribution to approximate the data. What is the event rate of the Poisson distribution? Please round to the nearest tenth.",
    "answer": "4.4",
    "data": [
      "/data/qrdata/data/ami_occurrences.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 4.50",
      "keywords": [
        "Statistics",
        "Poisson distribution"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file pew_energy_2018.csv contains data of a US-based survey on support for expanding six different sources of energy, including solar, wind, offshore drilling, hydrolic fracturing (\"fracking\"), coal, and nuclear.",
    "question": "Calculate the lower bound of a 99% confidence interval for the level of American support for expanding the use of wind turbines for power generation. Please round to the nearest thousandth.",
    "answer": "0.819",
    "data": [
      "/data/qrdata/data/pew_energy_2018.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 5.15b",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In New York City on October 23rd, 2014, a doctor who had recently been treating Ebola patients in Guinea went to the hospital with a slight fever and was subsequently diagnosed with Ebola. Soon thereafter, an NBC 4 New York/The Wall Street Journal/Marist Poll asked New Yorkers whether they favored a \"mandatory 21-day quarantine for anyone who has come in contact with an Ebola patient\". This poll included responses of 1,042 New York adults between October 26th and 28th, 2014.",
    "question": "What is the point estimate in this case? Please round to the nearest hundredth.",
    "answer": "0.82",
    "data": [
      "/data/qrdata/data/ebola_survey.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 5.11",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In New York City on October 23rd, 2014, a doctor who had recently been treating Ebola patients in Guinea went to the hospital with a slight fever and was subsequently diagnosed with Ebola. Soon thereafter, an NBC 4 New York/The Wall Street Journal/Marist Poll asked New Yorkers whether they favored a \"mandatory 21-day quarantine for anyone who has come in contact with an Ebola patient\". This poll included responses of 1,042 New York adults between October 26th and 28th, 2014.",
    "question": "Estimate the standard error of the point estimate from the Ebola survey. Please round to the nearest thousandth.",
    "answer": "0.012",
    "data": [
      "/data/qrdata/data/ebola_survey.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 5.12",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In New York City on October 23rd, 2014, a doctor who had recently been treating Ebola patients in Guinea went to the hospital with a slight fever and was subsequently diagnosed with Ebola. Soon thereafter, an NBC 4 New York/The Wall Street Journal/Marist Poll asked New Yorkers whether they favored a \"mandatory 21-day quarantine for anyone who has come in contact with an Ebola patient\". This poll included responses of 1,042 New York adults between October 26th and 28th, 2014.",
    "question": "Construct a 95% confidence interval for p, the proportion of New York adults who supported a quarantine for anyone who has come into contact with an Ebola patient. Please output the lower bound of the confidence interval and round to the nearest thousandth.",
    "answer": "0.796",
    "data": [
      "/data/qrdata/data/ebola_survey.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 5.13",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In New York City on October 23rd, 2014, a doctor who had recently been treating Ebola patients in Guinea went to the hospital with a slight fever and was subsequently diagnosed with Ebola. Soon thereafter, an NBC 4 New York/The Wall Street Journal/Marist Poll asked New Yorkers whether they favored a \"mandatory 21-day quarantine for anyone who has come in contact with an Ebola patient\". This poll included responses of 1,042 New York adults between October 26th and 28th, 2014.",
    "question": "Construct a 95% confidence interval for p, the proportion of New York adults who supported a quarantine for anyone who has come into contact with an Ebola patient. Please output the upper bound of the confidence interval and round to the nearest thousandth.",
    "answer": "0.844",
    "data": [
      "/data/qrdata/data/ebola_survey.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 5.13",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Public health has improved and evolved, but has the public's knowledge changed with it? This data set rosling_responses.csv contains sample responses for two survey questions posed by Hans Rosling during lectures to a wide array of college-educated audiences.",
    "question": "Compute the upper bound of a 95% confidence interval for the fraction of college-educated adults who answered the children_with_1_or_more_vaccination question correctly. Please round to the nearest thousandth.",
    "answer": "0.358",
    "data": [
      "/data/qrdata/data/rosling_responses.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 5.20",
      "keywords": [
        "Statistics",
        "Sampling"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "We consider an experiment for patients who underwent cardiopulmonary resuscitation (CPR) for a heart attack and were subsequently admitted to a hospital. These patients were randomly divided into a treatment group where they received a blood thinner or the control group where they did not receive a blood thinner. The outcome variable of interest was whether the patients survived for at least 24 hours. The data is shown in cpr.csv.",
    "question": "Calculate the lower bound of the 90% confidence interval of the difference for the survival rates in the CPR study. Please round to the nearest thousandth.",
    "answer": "-0.026",
    "data": [
      "/data/qrdata/data/cpr.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.12",
      "keywords": [
        "Statistics",
        "Confidence interval"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The results in fish_oil_18.csv summarize each of the health outcomes for an experiment where 12,933 subjects received a 1g fish oil supplement daily and 12,938 received a placebo daily. The experiment's duration was 5-years. The first row represents the treatment group and the second row represents the placebo group.\n\nVariables\nmajor_cardio_event - Major cardiovascular event. (Primary end point.)\ncardio_event_expanded - Cardiovascular event in expanded composite endpoint.\nmyocardioal_infarction - Total myocardial infarction. (Heart attack.)\nstroke - Total stroke.\ncardio_death - Death from cardiovascular causes.\nPCI - Percutaneous coronary intervention.\nCABG - Coronary artery bypass graft.\ntotal_coronary_heart_disease - Total coronary heart disease.\nischemic_stroke - Ischemic stroke.\nhemorrhagic_stroke - Hemorrhagic stroke.\nchd_death - Death from coronary heart disease.\nmyocardial_infarction_death - Death from myocardial infarction.\nstroke_death - Death from stroke.\ninvasive_cancer - Invasive cancer of any type. (Primary end point.)\nbreast_cancer - Breast cancer.\nprostate_cancer - Prostate cancer.\ncolorectal_cancer - Colorectal cancer.\ncancer_death - Death from cancer.\ndeath - Death from any cause.\nmajor_cardio_event_after_2y - Major cardiovascular event, excluding the first 2 years of follow-up.\nmyocardial_infarction_after_2y - Total myocardial infarction, excluding the first 2 years of follow-up.\ninvasive_cancer_after_2y - Invasive cancer of any type, excluding the first 2 years of follow-up.\ncancer_death_after_2y - Death from cancer, excluding the first 2 years of follow-up.\ndeath_after_2y - Death from any cause, excluding the first 2 years of follow-up.",
    "question": "Calculate the length of the 95% confidence interval for the effect of fish oils on heart attacks for patients who are well-represented by those in the study. Please round to the nearest thousandth.",
    "answer": "0.006",
    "data": [
      "/data/qrdata/data/fish_oil_18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.13",
      "keywords": [
        "Statistics",
        "Confidence interval"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The results in fish_oil_18.csv summarize each of the health outcomes for an experiment where 12,933 subjects received a 1g fish oil supplement daily and 12,938 received a placebo daily. The experiment's duration was 5-years. The first row represents the treatment group and the second row represents the placebo group.\n\nVariables\nmajor_cardio_event - Major cardiovascular event. (Primary end point.)\ncardio_event_expanded - Cardiovascular event in expanded composite endpoint.\nmyocardioal_infarction - Total myocardial infarction. (Heart attack.)\nstroke - Total stroke.\ncardio_death - Death from cardiovascular causes.\nPCI - Percutaneous coronary intervention.\nCABG - Coronary artery bypass graft.\ntotal_coronary_heart_disease - Total coronary heart disease.\nischemic_stroke - Ischemic stroke.\nhemorrhagic_stroke - Hemorrhagic stroke.\nchd_death - Death from coronary heart disease.\nmyocardial_infarction_death - Death from myocardial infarction.\nstroke_death - Death from stroke.\ninvasive_cancer - Invasive cancer of any type. (Primary end point.)\nbreast_cancer - Breast cancer.\nprostate_cancer - Prostate cancer.\ncolorectal_cancer - Colorectal cancer.\ncancer_death - Death from cancer.\ndeath - Death from any cause.\nmajor_cardio_event_after_2y - Major cardiovascular event, excluding the first 2 years of follow-up.\nmyocardial_infarction_after_2y - Total myocardial infarction, excluding the first 2 years of follow-up.\ninvasive_cancer_after_2y - Invasive cancer of any type, excluding the first 2 years of follow-up.\ncancer_death_after_2y - Death from cancer, excluding the first 2 years of follow-up.\ndeath_after_2y - Death from any cause, excluding the first 2 years of follow-up.",
    "question": "Calculate the lower bound of the 95% confidence interval for the effect of fish oils on heart attacks for patients who are well-represented by those in the study. Please round to the nearest thousandth.",
    "answer": "-0.007",
    "data": [
      "/data/qrdata/data/fish_oil_18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.13",
      "keywords": [
        "Statistics",
        "Confidence interval"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A mammogram is an X-ray procedure used to check for breast cancer. Whether mammograms should be used is part of a controversial discussion, and it's the topic of our next example where we learn about 2-proportion hypothesis tests when H0 is p1 = p2. A 30-year study was conducted with nearly 90,000 female participants. During a 5-year screening period, each woman was randomized to one of two groups: in the first group, women received regular mammograms to screen for breast cancer, and in the second group, women received regular non-mammogram breast cancer exams. No intervention was made during the following 25 years of the study, and we'll consider death resulting from breast cancer over the full 30-year period. Results from the study are in mammogram.csv.\n\nIf mammograms are much more effective than non-mammogram breast cancer exams, then we would expect to see additional deaths from breast cancer in the control group. On the other hand, if mammograms are not as effective as regular breast cancer exams, we would expect to see an increase in breast cancer deaths in the mammogram group.",
    "question": "Is this study an experiment or an observational study? Please answer with \"experiment\" or \"observational\".",
    "answer": "experiment",
    "data": [
      "/data/qrdata/data/mammogram.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.14",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "experiment",
        "observational"
      ]
    }
  },
  {
    "context": "A mammogram is an X-ray procedure used to check for breast cancer. Whether mammograms should be used is part of a controversial discussion, and it's the topic of our next example where we learn about 2-proportion hypothesis tests when H0 is p1 = p2. A 30-year study was conducted with nearly 90,000 female participants. During a 5-year screening period, each woman was randomized to one of two groups: in the first group, women received regular mammograms to screen for breast cancer, and in the second group, women received regular non-mammogram breast cancer exams. No intervention was made during the following 25 years of the study, and we'll consider death resulting from breast cancer over the full 30-year period. Results from the study are in mammogram.csv.\n\nIf mammograms are much more effective than non-mammogram breast cancer exams, then we would expect to see additional deaths from breast cancer in the control group. On the other hand, if mammograms are not as effective as regular breast cancer exams, we would expect to see an increase in breast cancer deaths in the mammogram group.",
    "question": "Compute the absolute value of the point estimate of the difference in breast cancer death rates in the two groups. Please round to 5 decimal places.",
    "answer": "0.00012",
    "data": [
      "/data/qrdata/data/mammogram.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.17",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A mammogram is an X-ray procedure used to check for breast cancer. Whether mammograms should be used is part of a controversial discussion, and it's the topic of our next example where we learn about 2-proportion hypothesis tests when H0 is p1 = p2. A 30-year study was conducted with nearly 90,000 female participants. During a 5-year screening period, each woman was randomized to one of two groups: in the first group, women received regular mammograms to screen for breast cancer, and in the second group, women received regular non-mammogram breast cancer exams. No intervention was made during the following 25 years of the study, and we'll consider death resulting from breast cancer over the full 30-year period. Results from the study are in mammogram.csv.\n\nIf mammograms are much more effective than non-mammogram breast cancer exams, then we would expect to see additional deaths from breast cancer in the control group. On the other hand, if mammograms are not as effective as regular breast cancer exams, we would expect to see an increase in breast cancer deaths in the mammogram group.",
    "question": "Compute the standard error of the point estimate of the difference in breast cancer death rates in the two groups. Please round to 5 decimal places.",
    "answer": "0.00070",
    "data": [
      "/data/qrdata/data/mammogram.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.17",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A mammogram is an X-ray procedure used to check for breast cancer. Whether mammograms should be used is part of a controversial discussion, and it's the topic of our next example where we learn about 2-proportion hypothesis tests when H0 is p1 = p2. A 30-year study was conducted with nearly 90,000 female participants. During a 5-year screening period, each woman was randomized to one of two groups: in the first group, women received regular mammograms to screen for breast cancer, and in the second group, women received regular non-mammogram breast cancer exams. No intervention was made during the following 25 years of the study, and we'll consider death resulting from breast cancer over the full 30-year period. Results from the study are in mammogram.csv.\n\nIf mammograms are much more effective than non-mammogram breast cancer exams, then we would expect to see additional deaths from breast cancer in the control group. On the other hand, if mammograms are not as effective as regular breast cancer exams, we would expect to see an increase in breast cancer deaths in the mammogram group.",
    "question": "Calculate a p-value for the hypothesis that there was no difference in breast cancer deaths in the mammogram and control groups. Please round to 4 decimal places.",
    "answer": "0.8650",
    "data": [
      "/data/qrdata/data/mammogram.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.18",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A quadcopter company is considering a new manufacturer for rotor blades. The new manufacturer would be more expensive, but they claim their higher-quality blades are more reliable, with 3% more blades passing inspection than their competitor.\nThe quality control engineer collects a sample of blades, examining 1000 blades from each company, and the CSV file drone_blades.csv represents the quality control data set for quadcopter drone blades.",
    "question": "Given the hypothesis that the higher-quality blades will pass inspection 3% more frequently than the standard-quality blades, calculate the p-value for the hypothesis. Please round to 3 decimal places.",
    "answer": "0.012",
    "data": [
      "/data/qrdata/data/drone_blades.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.20",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A quadcopter company is considering a new manufacturer for rotor blades. The new manufacturer would be more expensive, but they claim their higher-quality blades are more reliable, with 3% more blades passing inspection than their competitor.\nThe quality control engineer collects a sample of blades, examining 1000 blades from each company, and the CSV file drone_blades.csv represents the quality control data set for quadcopter drone blades.",
    "question": "Given the hypothesis that the higher-quality blades will pass inspection 3% more frequently than the standard-quality blades, will you accept or reject the hypothesis with a significance level of 5%? Please answer with \"accept\" or \"reject\".",
    "answer": "reject",
    "data": [
      "/data/qrdata/data/drone_blades.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 6.20",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "accept",
        "reject"
      ]
    }
  },
  {
    "context": "The Stanford University Heart Transplant Study was conducted to determine whether an experimental heart transplant program increased lifespan. Each patient entering the program was designated an official heart transplant candidate, meaning that he was gravely ill and would most likely benefit from a new heart. Some patients got a transplant and some did not. The variable transplant indicates which group the patients were in; patients in the treatment group got a transplant and those in the control group did not. Of the 34 patients in the control group, 30 died. Of the 69 people in the treatment group, 45 died. Another variable called survived was used to indicate whether or not the patient was alive at the end of the study. The data is in the CSV file heart_transplant.csv.",
    "question": "Suppose we are interested in estimating the difference in survival rate between the control and treatment groups using a confidence interval. Can we construct the interval using the normal approximation? Please answer with \"yes\" or \"no\".",
    "answer": "no",
    "data": [
      "/data/qrdata/data/heart_transplant.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.18",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "yes",
        "no"
      ]
    }
  },
  {
    "context": "A survey asked 827 randomly sampled registered voters in California \"Do you support? Or do you oppose? Drilling for oil and natural gas off the Coast of California? Or do you not know enough to say?\" The survey data is in offshore_drilling.csv.",
    "question": "Given the hypothesis that the proportion of college graduates who do not have an opinion on this issue is equal to that of non-college graduates, will you accept or reject the hypothesis with a significance level of 5%? Please answer with \"accept\" or \"reject\".",
    "answer": "reject",
    "data": [
      "/data/qrdata/data/offshore_drilling.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.23b",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "accept",
        "reject"
      ]
    }
  },
  {
    "context": "A survey asked 827 randomly sampled registered voters in California \"Do you support? Or do you oppose? Drilling for oil and natural gas off the Coast of California? Or do you not know enough to say?\" The survey data is in offshore_drilling.csv.",
    "question": "Given the hypothesis that the proportion of college graduates who support off-shore drilling in California is equal to that of non-college graduates, calculate the p-value for the hypothesis. Please round to 4 decimal places.",
    "answer": "0.6966",
    "data": [
      "/data/qrdata/data/offshore_drilling.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.23b",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A news article reports that “Americans have differing views on two potentially inconvenient and invasive practices that airports could implement to uncover potential terrorist attacks.” This news piece was based on a survey conducted among a random sample of 1,137 adults nationwide, interviewed by telephone November 7-10, 2010, where one of the questions on the survey was “Some airports are now using ‘full-body’ digital x-ray machines to electronically screen passengers in airport security lines. Do you think these new x-ray machines should or should not be used at airports?” The responses are in full_body_scan.csv.",
    "question": "Conduct an appropriate hypothesis test evaluating whether there is a difference in the proportion of Republicans and Democrats who think the full-body scans should be applied in airports with a significance level of 5%. Please answer with \"there is difference\" or \"no difference\".",
    "answer": "no difference",
    "data": [
      "/data/qrdata/data/full_body_scan.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.26a",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "there is difference",
        "no difference"
      ]
    }
  },
  {
    "context": "The National Sleep Foundation conducted a survey on the sleep habits of randomly sampled transportation workers and a control sample of non-transportation workers. The results of the survey are shown in sleep_deprivation.csv.",
    "question": "Given the hypothesis that there is no difference between the proportions of truck drivers and non-transportation workers (the control group) who get less than 6 hours of sleep per day, calculate the p-value for the hypothesis. Please round to 4 decimal places.",
    "answer": "0.0989",
    "data": [
      "/data/qrdata/data/sleep_deprivation.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.27",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The data file jury.csv contains a random sample of 275 jurors in a small county. Jurors identified their racial group, and we would like to determine if these jurors are racially representative of the population. In the population, 72% are white, 7% are black, 12% are hispanic, and 9% are others.",
    "question": "Given the null hypothesis that the jurors are a random sample, calculate the chi-square (X2) statistic. Please round to the nearest hundredth.",
    "answer": "5.89",
    "data": [
      "/data/qrdata/data/jury.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Text 6.3.2",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In this experiment, each individual was asked to be a seller of an iPod (a product commonly used to store music on before smart phones...). The participant received $10 + 5% of the sale price for participating. The iPod they were selling had frozen twice in the past inexplicably but otherwise worked fine. The prospective buyer starts off and then asks one of three final questions, depending on the seller's treatment group. The experiment data is in the CSV file ask.csv.\n\nThe three possible questions:\nGeneral: What can you tell me about it?\nPositive Assumption: It doesn't have any problems, does it?\nNegative Assumption: What problems does it have?\nThe outcome variable is whether or not the participant discloses or hides the problem with the iPod.\n\nThe hypothesis test for the iPod experiment is really about assessing whether there is statistically significant evidence that the success each question had on getting the participant to disclose the problem with the iPod. In other words, the goal is to check whether the buyer's question was independent of whether the seller disclosed a problem.",
    "question": "If the questions were actually equally effective, meaning about 27.85% of respondents would disclose the freezing issue regardless of what question they were asked, about how many sellers would we expect to hide the freezing problem from the Positive Assumption group? Please round to the nearest hundredth.",
    "answer": "52.67",
    "data": [
      "/data/qrdata/data/ask.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.39",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In this experiment, each individual was asked to be a seller of an iPod (a product commonly used to store music on before smart phones...). The participant received $10 + 5% of the sale price for participating. The iPod they were selling had frozen twice in the past inexplicably but otherwise worked fine. The prospective buyer starts off and then asks one of three final questions, depending on the seller's treatment group. The experiment data is in the CSV file ask.csv.\n\nThe three possible questions:\nGeneral: What can you tell me about it?\nPositive Assumption: It doesn't have any problems, does it?\nNegative Assumption: What problems does it have?\nThe outcome variable is whether or not the participant discloses or hides the problem with the iPod.\n\nThe hypothesis test for the iPod experiment is really about assessing whether there is statistically significant evidence that the success each question had on getting the participant to disclose the problem with the iPod. In other words, the goal is to check whether the buyer's question was independent of whether the seller disclosed a problem.",
    "question": "Given the null hypothesis that the questions had no impact on the sellers in the experiment, calculate the chi-square (X2) statistic. Please round to the nearest hundredth.",
    "answer": "40.13",
    "data": [
      "/data/qrdata/data/ask.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Text 6.4.2",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In this experiment, each individual was asked to be a seller of an iPod (a product commonly used to store music on before smart phones...). The participant received $10 + 5% of the sale price for participating. The iPod they were selling had frozen twice in the past inexplicably but otherwise worked fine. The prospective buyer starts off and then asks one of three final questions, depending on the seller's treatment group. The experiment data is in the CSV file ask.csv.\n\nThe three possible questions:\nGeneral: What can you tell me about it?\nPositive Assumption: It doesn't have any problems, does it?\nNegative Assumption: What problems does it have?\nThe outcome variable is whether or not the participant discloses or hides the problem with the iPod.\n\nThe hypothesis test for the iPod experiment is really about assessing whether there is statistically significant evidence that the success each question had on getting the participant to disclose the problem with the iPod. In other words, the goal is to check whether the buyer's question was independent of whether the seller disclosed a problem.",
    "question": "Given the null hypothesis that the questions had no impact on the sellers in the experiment, how many degrees of freedom should be associated with the chi-square distribution used for X2?",
    "answer": "2",
    "data": [
      "/data/qrdata/data/ask.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Text 6.4.2",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Three treatments were compared to test their relative efficacy (effectiveness) in treating Type 2 Diabetes in patients aged 10-17 who were being treated with metformin. The primary outcome was lack of glycemic control (or not); lacking glycemic control means the patient still needed insulin, which is not the preferred outcome for a patient.\nEach of the 699 patients in the experiment was randomized to one of the following treatments: (1) continued treatment with metformin (coded as met), (2) formin combined with rosiglitazone (coded as rosi), or (3) a lifestyle-intervention program (coded as lifestyle). Each patient had a primary outcome, which was either lacked glycemic control (failure) or did not lack that control (success).",
    "question": "Given the null hypothesis that there is no difference in the effectiveness of the three treatments, compute the chi-square test statistic. Please round to the nearest hundredth.",
    "answer": "8.16",
    "data": [
      "/data/qrdata/data/diabetes2.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.43",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Three treatments were compared to test their relative efficacy (effectiveness) in treating Type 2 Diabetes in patients aged 10-17 who were being treated with metformin. The primary outcome was lack of glycemic control (or not); lacking glycemic control means the patient still needed insulin, which is not the preferred outcome for a patient.\nEach of the 699 patients in the experiment was randomized to one of the following treatments: (1) continued treatment with metformin (coded as met), (2) formin combined with rosiglitazone (coded as rosi), or (3) a lifestyle-intervention program (coded as lifestyle). Each patient had a primary outcome, which was either lacked glycemic control (failure) or did not lack that control (success).",
    "question": "Given the null hypothesis that there is no difference in the effectiveness of the three treatments, conduct a chi-square test. Will you accept or reject the hypothesis at the 5% significance level? Please answer with \"accept\" or \"reject\".",
    "answer": "reject",
    "data": [
      "/data/qrdata/data/diabetes2.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.44",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "accept",
        "reject"
      ]
    }
  },
  {
    "context": "Three treatments were compared to test their relative efficacy (effectiveness) in treating Type 2 Diabetes in patients aged 10-17 who were being treated with metformin. The primary outcome was lack of glycemic control (or not); lacking glycemic control means the patient still needed insulin, which is not the preferred outcome for a patient.\nEach of the 699 patients in the experiment was randomized to one of the following treatments: (1) continued treatment with metformin (coded as met), (2) formin combined with rosiglitazone (coded as rosi), or (3) a lifestyle-intervention program (coded as lifestyle). Each patient had a primary outcome, which was either lacked glycemic control (failure) or did not lack that control (success).",
    "question": "Given the null hypothesis that there is no difference in the effectiveness of the three treatments, conduct a chi-square test. Identify the p-value. Please round to 3 decimal places.",
    "answer": "0.017",
    "data": [
      "/data/qrdata/data/diabetes2.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 6.44",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A news article reports that “Americans have differing views on two potentially inconvenient and invasive practices that airports could implement to uncover potential terrorist attacks.” This news piece was based on a survey conducted among a random sample of 1,137 adults nationwide, interviewed by telephone November 7-10, 2010, where one of the questions on the survey was “Some airports are now using ‘full-body’ digital x-ray machines to electronically screen passengers in airport security lines. Do you think these new x-ray machines should or should not be used at airports?” The responses are in full_body_scan.csv.",
    "question": "The differences in each political group may be due to chance. Under the null hypothesis of independence between an individual's party affiliation and his support of full-body scans, how many Republicans would you expect to not support the use of full-body scans? Please round to the nearest hundredth.",
    "answer": "47.55",
    "data": [
      "/data/qrdata/data/full_body_scan.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.36a",
      "keywords": [
        "Statistics",
        "Categorial data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A survey asked 827 randomly sampled registered voters in California \"Do you support? Or do you oppose? Drilling for oil and natural gas off the Coast of California? Or do you not know enough to say?\" The survey data is in offshore_drilling.csv.",
    "question": "Given the null hypothesis that the opinion of college grads and non-grads is not different on the topic of drilling for oil and natural gas off the coast of California, compute the chi-square test statistic. Please round to the nearest hundredth.",
    "answer": "11.47",
    "data": [
      "/data/qrdata/data/offshore_drilling.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.37",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A survey asked 827 randomly sampled registered voters in California \"Do you support? Or do you oppose? Drilling for oil and natural gas off the Coast of California? Or do you not know enough to say?\" The survey data is in offshore_drilling.csv.",
    "question": "Given the null hypothesis that the opinion of college grads and non-grads is not different on the topic of drilling for oil and natural gas off the coast of California, conduct a chi-square test. Identify the p-value. Please round to 3 decimal places.",
    "answer": "0.003",
    "data": [
      "/data/qrdata/data/offshore_drilling.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 6.37",
      "keywords": [
        "Statistics",
        "Categorial data",
        "Chi-square"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ucla_textbooks_f18.csv contains a sample of courses collected from UCLA from Fall 2018, and the corresponding textbook prices collected from the UCLA bookstore and also from Amazon.",
    "question": "Given the null hypothesis that there is no difference in the average textbook price from the UCLA bookstore and from Amazon, conduct a hypothesis test. Identify the standard error associated with the price difference. Please round to 2 decimal places.",
    "answer": "1.63",
    "data": [
      "/data/qrdata/data/ucla_textbooks_f18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 7.18",
      "keywords": [
        "Statistics",
        "Numerical data",
        "Standard error"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ucla_textbooks_f18.csv contains a sample of courses collected from UCLA from Fall 2018, and the corresponding textbook prices collected from the UCLA bookstore and also from Amazon.",
    "question": "Given the null hypothesis that there is no difference in the average textbook price from the UCLA bookstore and from Amazon, conduct a hypothesis test. Identify the T-statistic of the price difference. Please round to 2 decimal places.",
    "answer": "2.20",
    "data": [
      "/data/qrdata/data/ucla_textbooks_f18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 7.18",
      "keywords": [
        "Statistics",
        "Numerical data",
        "T-statistic"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ucla_textbooks_f18.csv contains a sample of courses collected from UCLA from Fall 2018, and the corresponding textbook prices collected from the UCLA bookstore and also from Amazon.",
    "question": "Given the null hypothesis that there is no difference in the average textbook price from the UCLA bookstore and from Amazon, conduct a hypothesis test. Identify the p-value. Please round to 4 decimal places.",
    "answer": "0.0312",
    "data": [
      "/data/qrdata/data/ucla_textbooks_f18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Example 7.18",
      "keywords": [
        "Statistics",
        "Numerical data",
        "P-value"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ucla_textbooks_f18.csv contains a sample of courses collected from UCLA from Fall 2018, and the corresponding textbook prices collected from the UCLA bookstore and also from Amazon.",
    "question": "Create a 95% confidence interval for the average price difference between books at the UCLA bookstore and books on Amazon (price at the UCLA bookstore minus price on Amazon). Output the lower bound of the confidence interval. Please round to 2 decimal places.",
    "answer": "0.32",
    "data": [
      "/data/qrdata/data/ucla_textbooks_f18.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Guided Practice 7.19",
      "keywords": [
        "Statistics",
        "Numerical data",
        "Confidence interval"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file climate70.csv contains a random set of monitoring locations taken from NOAA data that had both years of interest (1948 and 2018) as well as data for both summary metrics of interest (dx70 and dx90, which are described below). It is a data frame with 197 observations on the following 7 variables.\nVariable Description\nstation: Station ID.\nlatitude: Latitude of the station.\nlongitude: Longitude of the station.\ndx70_1948: Number of days above 70 degrees in 1948.\ndx70_2018: Number of days above 70 degrees in 2018.\ndx90_1948: Number of days above 90 degrees in 1948.\ndx90_2018: Number of days above 90 degrees in 2018.",
    "question": "Please calculate the T-statistic of the hypothesis \"There is no difference in average number of days exceeding 90°F in 1948 and 2018 for NOAA stations\". Please round to the nearest hundredth.",
    "answer": "2.36",
    "data": [
      "/data/qrdata/data/climate70.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 7.19d",
      "keywords": [
        "Statistics",
        "Inference",
        "Hypothesis validation",
        "T-statistic"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file climate70.csv contains a random set of monitoring locations taken from NOAA data that had both years of interest (1948 and 2018) as well as data for both summary metrics of interest (dx70 and dx90, which are described below). It is a data frame with 197 observations on the following 7 variables.\nVariable Description\nstation: Station ID.\nlatitude: Latitude of the station.\nlongitude: Longitude of the station.\ndx70_1948: Number of days above 70 degrees in 1948.\ndx70_2018: Number of days above 70 degrees in 2018.\ndx90_1948: Number of days above 90 degrees in 1948.\ndx90_2018: Number of days above 90 degrees in 2018.",
    "question": "Please calculate the p-value of the hypothesis \"There is no difference in average number of days exceeding 90°F in 1948 and 2018 for NOAA stations\". Please round to the nearest thousandth.",
    "answer": "0.019",
    "data": [
      "/data/qrdata/data/climate70.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 7.19d",
      "keywords": [
        "Statistics",
        "Inference",
        "Hypothesis validation",
        "P-value"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file climate70.csv contains a random set of monitoring locations taken from NOAA data that had both years of interest (1948 and 2018) as well as data for both summary metrics of interest (dx70 and dx90, which are described below). It is a data frame with 197 observations on the following 7 variables.\nVariable Description\nstation: Station ID.\nlatitude: Latitude of the station.\nlongitude: Longitude of the station.\ndx70_1948: Number of days above 70 degrees in 1948.\ndx70_2018: Number of days above 70 degrees in 2018.\ndx90_1948: Number of days above 90 degrees in 1948.\ndx90_2018: Number of days above 90 degrees in 2018.",
    "question": "Calculate a 90% confidence interval for the average difference between the number of days exceeding 90°F between 1948 and 2018 (days in 2018 minus days in 1948). Output the upper bound of the confidence interval. Please round to 2 decimal places.",
    "answer": "4.93",
    "data": [
      "/data/qrdata/data/climate70.csv"
    ],
    "metadata": {
      "reference": "OpenIntro Statistics Ex 7.21a",
      "keywords": [
        "Statistics",
        "Inference",
        "Confidence interval"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_2.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L Central chest pain causes L shoulder impingement\nB. L shoulder impingement causes L Central chest pain\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_2.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_9.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. backache causes L leg pain\nB. L leg pain causes backache\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_9.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_10.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R thumb trouble causes L Central chest pain\nB. L Central chest pain causes R thumb trouble\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_10.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_16.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L ham problem causes R ham problem\nB. R ham problem causes L ham problem\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_16.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_19.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L knee pain causes L lateral foot pain\nB. L lateral foot pain causes L knee pain\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_19.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_22.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R leg pain causes R medical obesity\nB. R medical obesity causes R leg pain\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/Neuropathic_22.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_25.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R hip pain causes L L4 radiculopathy\nB. L L4 radiculopathy causes R hip pain\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_25.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_28.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L lateral arm discomfort causes R C6 radiculopathy\nB. R C6 radiculopathy causes L lateral arm discomfort\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_28.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_30.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. IBS causes R L1 radiculopathy\nB. R L1 radiculopathy causes IBS\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_30.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_33.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L C6 radiculopathy causes R hand problems\nB. R hand problems causes L C6 radiculopathy\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/Neuropathic_33.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_34.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L lateral obesity causes L L4 radiculopathy\nB. L L4 radiculopathy causes L lateral obesity\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_34.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_37.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R intracapular problems causes L C5 radiculopathy\nB. L C5 radiculopathy causes R intracapular problems\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_37.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_42.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R L5 radiculopathy causes L back headache\nB. L back headache causes R L5 radiculopathy\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/Neuropathic_42.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_43.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R chest disorders causes L T5 radiculopathy\nB. L T5 radiculopathy causes R chest disorders\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_43.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_46.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R Lumbago causes L L4 radiculopathy\nB. L L4 radiculopathy causes R Lumbago\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_46.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_pairwise_1.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. L obesity causes R S1 radiculopathy\nB. R S1 radiculopathy causes L obesity\nPlease answer with A or B.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_pairwise_1.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Pairwise causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_pairwise_3.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R L1 radiculopathy causes IBS\nB. IBS causes R L1 radiculopathy\nPlease answer with A or B.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/Neuropathic_pairwise_3.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Pairwise causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_pairwise_5.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. toracal dysfunction causes L T6 radiculopathy\nB. L T6 radiculopathy causes toracal dysfunction\nPlease answer with A or B.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_pairwise_5.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Pairwise causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_pairwise_8.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. R medical obesity causes R L4 radiculopathy\nB. R L4 radiculopathy causes R medical obesity\nPlease answer with A or B.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_pairwise_8.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Pairwise causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b"
      ]
    }
  },
  {
    "context": "The neuropathic pain diagnosis mainly consists of symptom diagnosis, pattern diagnosis, and pathophysiological diagnosis. The csv file Neuropathic_pairwise_28.csv contains neuropathic pain diagnosis records in the form of tables of which the row represents different patients and the column represents different diagnostic labels.",
    "question": "Which cause-and-effect relationship is more likely?\nA. toracal dysfunction causes R T4 radiculopathy\nB. R T4 radiculopathy causes toracal dysfunction\nPlease answer with A or B.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/Neuropathic_pairwise_28.csv"
    ],
    "metadata": {
      "reference": "Neuropathic pain dataset",
      "keywords": [
        "Causality",
        "Pairwise causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_tot_precip causes Residual_GH_mean\nB. Residual_GH_mean causes Residual_tot_precip\nC. The causal relation is double sided between Residual_tot_precip and Residual_GH_mean\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_u10m causes Residual_longwave\nB. Residual_longwave causes Residual_u10m\nC. The causal relation is double sided between Residual_u10m and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_cover causes Residual_longwave\nB. Residual_longwave causes Residual_cloud_cover\nC. The causal relation is double sided between Residual_cloud_cover and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_v10m causes Residual_tot_precip\nB. Residual_tot_precip causes Residual_v10m\nC. The causal relation is double sided between Residual_v10m and Residual_tot_precip\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_cover causes Residual_sea_ice\nB. Residual_sea_ice causes Residual_cloud_cover\nC. The causal relation is double sided between Residual_cloud_cover and Residual_sea_ice\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_tot_precip\nB. Residual_tot_precip causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_tot_precip\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_SLP\nB. Residual_SLP causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_SLP\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_u10m causes Residual_GH_mean\nB. Residual_GH_mean causes Residual_u10m\nC. The causal relation is double sided between Residual_u10m and Residual_GH_mean\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_heat_flux causes Residual_longwave\nB. Residual_longwave causes Residual_heat_flux\nC. The causal relation is double sided between Residual_heat_flux and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_sea_ice\nB. Residual_sea_ice causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_sea_ice\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_RH causes Residual_GH_mean\nB. Residual_GH_mean causes Residual_RH\nC. The causal relation is double sided between Residual_RH and Residual_GH_mean\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_SLP causes Residual_tot_precip\nB. Residual_tot_precip causes Residual_SLP\nC. The causal relation is double sided between Residual_SLP and Residual_tot_precip\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_heat_flux causes Residual_SLP\nB. Residual_SLP causes Residual_heat_flux\nC. The causal relation is double sided between Residual_heat_flux and Residual_SLP\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_shortwave causes Residual_GH_mean\nB. Residual_GH_mean causes Residual_shortwave\nC. The causal relation is double sided between Residual_shortwave and Residual_GH_mean\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_heat_flux causes Residual_v10m\nB. Residual_v10m causes Residual_heat_flux\nC. The causal relation is double sided between Residual_heat_flux and Residual_v10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_RH causes Residual_longwave\nB. Residual_longwave causes Residual_RH\nC. The causal relation is double sided between Residual_RH and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_sea_ice causes Residual_SLP\nB. Residual_SLP causes Residual_sea_ice\nC. The causal relation is double sided between Residual_sea_ice and Residual_SLP\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_GH_mean causes Residual_SLP\nB. Residual_SLP causes Residual_GH_mean\nC. The causal relation is double sided between Residual_GH_mean and Residual_SLP\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_RH causes Residual_cloud_cover\nB. Residual_cloud_cover causes Residual_RH\nC. The causal relation is double sided between Residual_RH and Residual_cloud_cover\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_u10m causes Residual_shortwave\nB. Residual_shortwave causes Residual_u10m\nC. The causal relation is double sided between Residual_u10m and Residual_shortwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_sea_ice causes Residual_u10m\nB. Residual_u10m causes Residual_sea_ice\nC. The causal relation is double sided between Residual_sea_ice and Residual_u10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_u10m\nB. Residual_u10m causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_u10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_u10m causes Residual_heat_flux\nB. Residual_heat_flux causes Residual_u10m\nC. The causal relation is double sided between Residual_u10m and Residual_heat_flux\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_RH causes Residual_cloud_water\nB. Residual_cloud_water causes Residual_RH\nC. The causal relation is double sided between Residual_RH and Residual_cloud_water\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_GH_mean causes Residual_sea_ice\nB. Residual_sea_ice causes Residual_GH_mean\nC. The causal relation is double sided between Residual_GH_mean and Residual_sea_ice\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_heat_flux causes Residual_tot_precip\nB. Residual_tot_precip causes Residual_heat_flux\nC. The causal relation is double sided between Residual_heat_flux and Residual_tot_precip\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_RH causes Residual_v10m\nB. Residual_v10m causes Residual_RH\nC. The causal relation is double sided between Residual_RH and Residual_v10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_v10m causes Residual_longwave\nB. Residual_longwave causes Residual_v10m\nC. The causal relation is double sided between Residual_v10m and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_GH_mean causes Residual_heat_flux\nB. Residual_heat_flux causes Residual_GH_mean\nC. The causal relation is double sided between Residual_GH_mean and Residual_heat_flux\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_heat_flux\nB. Residual_heat_flux causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_heat_flux\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_longwave causes Residual_cloud_water\nB. Residual_cloud_water causes Residual_longwave\nC. The causal relation is double sided between Residual_longwave and Residual_cloud_water\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_tot_precip causes Residual_RH\nB. Residual_RH causes Residual_tot_precip\nC. The causal relation is double sided between Residual_tot_precip and Residual_RH\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_cover causes Residual_u10m\nB. Residual_u10m causes Residual_cloud_cover\nC. The causal relation is double sided between Residual_cloud_cover and Residual_u10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_heat_flux causes Residual_shortwave\nB. Residual_shortwave causes Residual_heat_flux\nC. The causal relation is double sided between Residual_heat_flux and Residual_shortwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_cloud_water causes Residual_v10m\nB. Residual_v10m causes Residual_cloud_water\nC. The causal relation is double sided between Residual_cloud_water and Residual_v10m\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_SLP causes Residual_shortwave\nB. Residual_shortwave causes Residual_SLP\nC. The causal relation is double sided between Residual_SLP and Residual_shortwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_sea_ice causes Residual_RH\nB. Residual_RH causes Residual_sea_ice\nC. The causal relation is double sided between Residual_sea_ice and Residual_RH\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_sea_ice causes Residual_shortwave\nB. Residual_shortwave causes Residual_sea_ice\nC. The causal relation is double sided between Residual_sea_ice and Residual_shortwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_longwave causes Residual_sea_ice\nB. Residual_sea_ice causes Residual_longwave\nC. The causal relation is double sided between Residual_longwave and Residual_sea_ice\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The arctic dataset in arctic.csv is on the drivers of arctic sea ice thickness (or coverage): what causes the arctic sea coverage to increase or decrease? Variables in the dataset include total cloud water path, sea level pressure, geopotential height, meridional and zonal wind at 10m, net shortwave, longwave flux at the surface and so on.",
    "question": "Which cause-and-effect relationship is more likely?\nA. Residual_SLP causes Residual_longwave\nB. Residual_longwave causes Residual_SLP\nC. The causal relation is double sided between Residual_SLP and Residual_longwave\nD. No causal relationship exists\nPlease answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/arctic.csv"
    ],
    "metadata": {
      "reference": "Arctic sea ice dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. P38 causes plcg\nB. plcg causes P38\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pjnk causes plcg\nB. plcg causes pjnk\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKA causes PIP2\nB. PIP2 causes PKA\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PIP3 causes praf\nB. praf causes PIP3\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes PIP2\nB. PIP2 causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PIP2 causes pakts473\nB. pakts473 causes PIP2\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes p44/42\nB. p44/42 causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. praf causes plcg\nB. plcg causes praf\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. praf causes pjnk\nB. pjnk causes praf\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PIP3 causes PKA\nB. PKA causes PIP3\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. p44/42 causes PIP2\nB. PIP2 causes p44/42\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. praf causes PKA\nB. PKA causes praf\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes PIP3\nB. PIP3 causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes plcg\nB. plcg causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes PIP3\nB. PIP3 causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PIP3 causes P38\nB. P38 causes PIP3\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes PKC\nB. PKC causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pakts473 causes pmek\nB. pmek causes pakts473\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes pjnk\nB. pjnk causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes PKA\nB. PKA causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. PKC causes praf\nB. praf causes PKC\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "A",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes PIP2\nB. PIP2 causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. plcg causes pakts473\nB. pakts473 causes plcg\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pjnk causes PIP2\nB. PIP2 causes pjnk\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. P38 causes p44/42\nB. p44/42 causes P38\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes praf\nB. praf causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The data set in flow.csv offers continuous measurements of expression levels of multiple phosphorylated proteins and phospholipid components in human immune system cells. It contains 7466 cells (n = 7466) and flow cytometry measurements of 11 (p = 11) phosphorylated proteins and phospholipids.",
    "question": "Which cause-and-effect relationship is more likely?\nA. pmek causes plcg\nB. plcg causes pmek\nC. No causal relationship exists\nPlease answer with A, B, or C.",
    "answer": "C",
    "data": [
      "/data/qrdata/data/flow.csv"
    ],
    "metadata": {
      "reference": "Flow cytometry dataset",
      "keywords": [
        "Causality",
        "Full graph causal discovery",
        "Observational data"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "The CSV file ihdp_0.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.02",
    "data": [
      "/data/qrdata/data/ihdp_0.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_0.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the treated of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_0.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_0.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.02",
    "data": [
      "/data/qrdata/data/ihdp_0.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_1.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.05",
    "data": [
      "/data/qrdata/data/ihdp_1.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_1.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.06",
    "data": [
      "/data/qrdata/data/ihdp_1.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_2.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.10",
    "data": [
      "/data/qrdata/data/ihdp_2.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_2.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.12",
    "data": [
      "/data/qrdata/data/ihdp_2.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_3.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.27",
    "data": [
      "/data/qrdata/data/ihdp_3.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_3.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the treated of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_3.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_3.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.34",
    "data": [
      "/data/qrdata/data/ihdp_3.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_4.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.16",
    "data": [
      "/data/qrdata/data/ihdp_4.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_4.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the treated of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_4.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_4.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.20",
    "data": [
      "/data/qrdata/data/ihdp_4.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_5.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_5.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_5.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the treated of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_5.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_5.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_5.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_6.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "3.99",
    "data": [
      "/data/qrdata/data/ihdp_6.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_6.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "3.99",
    "data": [
      "/data/qrdata/data/ihdp_6.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_7.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "3.85",
    "data": [
      "/data/qrdata/data/ihdp_7.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_7.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "3.82",
    "data": [
      "/data/qrdata/data/ihdp_7.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_8.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "10.47",
    "data": [
      "/data/qrdata/data/ihdp_8.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_8.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the treated of the dataset? Please round to the nearest hundredth.",
    "answer": "4.00",
    "data": [
      "/data/qrdata/data/ihdp_8.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_8.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "11.94",
    "data": [
      "/data/qrdata/data/ihdp_8.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_9.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect of the dataset? Please round to the nearest hundredth.",
    "answer": "4.59",
    "data": [
      "/data/qrdata/data/ihdp_9.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file ihdp_9.csv contains data obtained from the Infant Health and Development Program (IHDP). The study is designed to evaluate the effect of home visit from specialist doctors on the cognitive test scores of premature infants. The confounders x (x1-x25) correspond to collected measurements of the children and their mothers, including measurements on the child (birth weight, head circumference, weeks born preterm, birth order, first born, neonatal health index, sex, twin status), as well as behaviors engaged in during the pregnancy (smoked cigarettes, drank alcohol, took drugs) and measurements on the mother at the time she gave birth (age, marital status, educational attainment, whether she worked during pregnancy, whether she received prenatal care) and the site (8 total) in which the family resided at the start of the intervention. There are 6 continuous covariates and 19 binary covariates.",
    "question": "What is the Average treatment effect on the control of the dataset? Please round to the nearest hundredth.",
    "answer": "4.72",
    "data": [
      "/data/qrdata/data/ihdp_9.csv"
    ],
    "metadata": {
      "reference": "IHDP dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the control",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_0.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.074",
    "data": [
      "/data/qrdata/data/jobs_0.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_1.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.025",
    "data": [
      "/data/qrdata/data/jobs_1.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_2.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.081",
    "data": [
      "/data/qrdata/data/jobs_2.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_3.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.110",
    "data": [
      "/data/qrdata/data/jobs_3.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_4.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.075",
    "data": [
      "/data/qrdata/data/jobs_4.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_5.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.097",
    "data": [
      "/data/qrdata/data/jobs_5.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_6.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.100",
    "data": [
      "/data/qrdata/data/jobs_6.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_7.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.075",
    "data": [
      "/data/qrdata/data/jobs_7.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_8.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.094",
    "data": [
      "/data/qrdata/data/jobs_8.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file jobs_9.csv contains data obtained from the National Supported Work program. The study is designed to evaluate the effect of job training (t) on the income and employment status after training (y). The confounders x (x0-x16) correspond to covariates such as age and education, as well as previous earnings.",
    "question": "Please estimate the Average treatment effect on the treated (ATT) of the dataset. Please round to the nearest thousandth.",
    "answer": "0.073",
    "data": [
      "/data/qrdata/data/jobs_9.csv"
    ],
    "metadata": {
      "reference": "Jobs dataset",
      "keywords": [
        "Causality",
        "Average treatment effect on the treated",
        "Observational data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the impacts of online class format on exam outcomes, the dataset online_classroom.csv was used to compare students in online classes versus face-to-face classes. In this analysis, the treatment effect is measured as the difference in exam scores between the two groups. Each row of the dataset contains a student's exam outcome (the variable falsexam), their classroom format (online, face-to-face, or blended), and other variables like gender and ethnicity.",
    "question": "What is the average treatment effect (ATE) of taking classes online on exam scores? Please use linear regression analysis considering only the students in face-to-face (format_blended = 0 and format_ol = 0) and online classes (format_ol = 1), do not consider other variables, and provide the ATE rounded to the nearest hundredth.",
    "answer": "-4.91",
    "data": [
      "/data/qrdata/data/online_classroom.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 5",
      "keywords": [
        "Causality",
        "Observational data",
        "Linear regression",
        "Average treatment effect"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the impact of an additional year of education on hourly wage, we look at a sample size representing individuals with varying levels of education and their hourly wages. The dataset wage.csv contains the necessary data for our analysis. The columns in this dataset include 'wage', representing the total income; 'hours', representing the total hours worked; and 'educ', representing the years of education; and other variables like the parents' education and the person's IQ score.",
    "question": "What percentage increase in hourly wage can be expected for each additional year of education, based on the data described? Please use linear regression and do not consider variables other than 'wage', 'hours', and 'educ'. Please provide your answer to the nearest hundredth of a percent.",
    "answer": "5.36%",
    "data": [
      "/data/qrdata/data/wage.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 5",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the impact of an additional year of education on hourly wage, we look at a sample size representing individuals with varying levels of education and their hourly wages. The dataset wage.csv contains the necessary data for our analysis. The columns in this dataset include 'wage', representing the total income; 'hours', representing the total hours worked; and 'educ', representing the years of education; and other variables like the parents' education and the person's IQ score.",
    "question": "What percentage increase in hourly wage can be expected for each additional year of education, based on the data described? Please use linear regression and take all provided variables into consideration. Please provide your answer to the nearest hundredth of a percent.",
    "answer": "4.11%",
    "data": [
      "/data/qrdata/data/wage.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 5",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the impact of education on hourly wage, we look at a sample size representing individuals with varying levels of education and their hourly wages. The dataset wage.csv contains the necessary data for our analysis. The columns in this dataset include 'wage', representing the total income; 'hours', representing the total hours worked; and 'educ', representing the years of education; and other variables like the parents' education and the person's IQ score.",
    "question": "Please estimate the effect of graduating 12th grade on hourly wage. Please use linear regression and do not consider variables other than 'wage', 'hours', and 'educ'. Please provide your answer to the nearest hundredth.",
    "answer": "4.90",
    "data": [
      "/data/qrdata/data/wage.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 6",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "In a study to determine the causal effect of sending an email reminder on the repayment of debts, a fintech company conducted a random test involving 5000 customers who were late on their payments. Each customer was randomly assigned to either receive an email about negotiating their debt or to be part of a control group that did not receive the email. Data was collected on the amounts paid by the late customers after this intervention. The dataset collections_email.csv contains variables including the amount paid (`payments`), whether the email was sent (`email`), whether the email was opened (`opened`), whether the customer contacted the collections department to negotiate their debt after having received the email (`agreement`), the customer's credit limit before being late (`credit_limit`), and the customer's risk score prior to the delivery of the email (`risk_score`).",
    "question": "What is the average treatment effect (ATE) on payments from sending the email reminder to late-paying customers? Please choose the variables to adjust for, conduct linear regression, and round your answer to the nearest hundredth.",
    "answer": "4.43",
    "data": [
      "/data/qrdata/data/collections_email.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 7",
      "keywords": [
        "Causality",
        "Randomized experiment",
        "Regression analysis",
        "Average treatment effect",
        "Confounder selection"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The dataset hospital_treatment.csv includes data from a randomized trial conducted by two hospitals. Both of them are conducting randomised trials on a new drug to treat a certain illness. The outcome of interest is days hospitalised. If the treatment is effective, it will lower the amount of days the patient stays in the hospital. For one of the hospitals, the policy regarding the random treatment is to give it to 90% of its patients while 10% get a placebo. The other hospital has a different policy: it gives the drug to a random 10% of its patients and 90% get a placebo. You are also told that the hospital that gives 90% of the true drug and 10% of placebo usually gets more severe cases of the illness to treat. The CSV file contains columns for `hospital` indicating the hospital a patient belongs to, `treatment` signifying if the patient received the new drug or placebo, `severity` reflecting the severity of the illness, and `days` representing the number of days the patient was hospitalized.",
    "question": "What is the average treatment effect (ATE) of the new drug on the amount of days the patient stays in the hospital? Please choose the variables to adjust for, conduct linear regression, and round your answer to the nearest hundredth. If the new drug reduces the amount of days the patient stays in the hospital, the answer should be negative.",
    "answer": "-7.59",
    "data": [
      "/data/qrdata/data/hospital_treatment.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 7",
      "keywords": [
        "Causality",
        "Randomized experiment",
        "Regression analysis",
        "Average treatment effect",
        "Confounder selection"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The dataset hospital_treatment.csv includes data from a randomized trial conducted by two hospitals. Both of them are conducting randomised trials on a new drug to treat a certain illness. The outcome of interest is days hospitalised. If the treatment is effective, it will lower the amount of days the patient stays in the hospital. For one of the hospitals, the policy regarding the random treatment is to give it to 90% of its patients while 10% get a placebo. The other hospital has a different policy: it gives the drug to a random 10% of its patients and 90% get a placebo. You are also told that the hospital that gives 90% of the true drug and 10% of placebo usually gets more severe cases of the illness to treat. The CSV file contains columns for `hospital` indicating the hospital a patient belongs to, `treatment` signifying if the patient received the new drug or placebo, `severity` reflecting the severity of the illness, and `days` representing the number of days the patient was hospitalized.",
    "question": "We are estimating the average treatment effect (ATE) of the new drug on the amount of days the patient stays in the hospital, and we already controlled for the severity. Should we also control for the 'hospital' variable? Please answer with 'yes' or 'no'.",
    "answer": "no",
    "data": [
      "/data/qrdata/data/hospital_treatment.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 7",
      "keywords": [
        "Causality",
        "Randomized experiment",
        "Regression analysis",
        "Average treatment effect",
        "Confounder selection"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "yes",
        "no"
      ]
    }
  },
  {
    "context": "The dataset provided, ak91.csv, contains information on individuals' log wages, years of schooling, year of birth, quarter of birth, and state of birth. The purpose of the analysis is to estimate the effect of education on wage, using linear regression and quarter of birth as an instrumental variable (IV). This idea takes advantage of US compulsory attendance law. Usually, they state that a kid must have turned 6 years by January 1 of the year they enter school. For this reason, kids that are born at the beginning of the year will enter school at an older age. Compulsory attendance law also requires students to be in school until they turn 16, at which point they are legally allowed to drop out. The result is that people born later in the year have, on average, more years of education than those born in the beginning of the year.",
    "question": "What is the average additional percentage wage increase associated with each additional year of education based on the instrumental variable of whether the individual is born in the last quarter q4? Please adjust for all the other variables, and round your answer to the nearest hundredth of a percent.",
    "answer": "8.53%",
    "data": [
      "/data/qrdata/data/ak91.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 8",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis",
        "Average treatment effect",
        "Instrumental variables"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A study is conducted to measure the effect of a marketing push on user engagement, specifically in-app purchases. Some customers who were assigned to receive the push are not receiving it, because they probably have an older phone that doesn’t support the kind of push the marketing team designed.\nThe dataset app_engagement_push.csv contains records for 10,000 random customers. Each record includes whether an in-app purchase was made (in_app_purchase), if a marketing push was assigned to the user (push_assigned), and if the marketing push was successfully delivered (push_delivered).",
    "question": "What is the Local Average Treatment Effect (LATE) of receiving the marketing push on in-app purchases, as estimated using linear regression and instrumental variable, rounded to two decimal places?",
    "answer": "3.29",
    "data": [
      "/data/qrdata/data/app_engagement_push.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 9",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis",
        "Local average treatment effect",
        "Instrumental variables"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "A study is conducted to measure the effect of a marketing push on user engagement, specifically in-app purchases. Some customers who were assigned to receive the push are not receiving it, because they probably have an older phone that doesn’t support the kind of push the marketing team designed.\nThe dataset app_engagement_push.csv contains records for 10,000 random customers. Each record includes whether an in-app purchase was made (in_app_purchase), if a marketing push was assigned to the user (push_assigned), and if the marketing push was successfully delivered (push_delivered).",
    "question": "Consider we estimate the Average Treatment Effect of receiving the marketing push on in-app purchases, by conducting linear regression with the formula \"in_app_purchase ~ 1 + push_assigned + push_delivered\". Will the true impact be A) higher than, B) lower than, C) the same with our estimated impact? Please answer with A, B, or C.",
    "answer": "B",
    "data": [
      "/data/qrdata/data/app_engagement_push.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 9",
      "keywords": [
        "Causality",
        "Observational data",
        "Regression analysis",
        "Average treatment effect",
        "Instrumental variables"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c"
      ]
    }
  },
  {
    "context": "To investigate the effect of a medication on the number of days it takes for a patient to recover from an illness, we have a dataset that includes several confounding variables like severity, sex, and age. The dataset medicine_impact_recovery.csv contains data on patients who have been prescribed medication and those who haven't. The variables include sex (0 or 1), age, the severity of the condition, whether the patient was on medication (0 or 1), and the number of days it took for each patient to recover.",
    "question": "What is the average treatment effect of the medication on the recovery time when controlling for severity, sex, and age using the K nearest neighbour matching? Please scale the features, use the euclidean norm as the matching measurement, and also adjust for the matching bias. Provide your answer rounded to two decimal places. The answer should be positive if the mediation makes the recovery time longer.",
    "answer": "-7.36",
    "data": [
      "/data/qrdata/data/medicine_impact_recovery.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 10",
      "keywords": [
        "Causality",
        "Observational data",
        "Matching",
        "Average treatment effect"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The National Study of Learning Mindsets is a randomised study conducted in U.S. public high schools which aims at finding the impact of a growth mindset. The way it works is that students receive from the school a seminar to instill in them a growth mindset. Then, they follow up the students in their college years to measure how well they’ve performed academically. This measurement was compiled into an achievement score and standardized. The CSV file learning_mindset.csv contains simulated data of this research.\nVariable Description\nintervention: the intervention of the growth mindset;\nachievement_score: the standardized academic achievement score;\nschoolid: identifier of the student’s school;\nsuccess_expect: self-reported expectations for success in the future, a proxy for prior achievement, measured prior to random assignment;\nethnicity: categorical variable for student race/ethnicity;\ngender: categorical variable for student identified gender;\nfrst_in_family: categorical variable for student first-generation status, i.e. first in family to go to college;\nschool_urbanicity: school-level categorical variable for urbanicity of the school, i.e. rural, suburban, etc;\nschool_mindset: school-level mean of students’ fixed mindsets, reported prior to random assignment, standardized;\nschool_achievement: school achievement level, as measured by test scores and college preparation for the previous 4 cohorts of students, standardized;\nschool_ethnic_minority: school racial/ethnic minority composition, i.e., percentage of student body that is Black, Latino, or Native American, standardized;\nschool_poverty: school poverty concentration, i.e., percentage of students who are from families whose incomes fall below the federal poverty line, standardized;\nschool_size: total number of students in all four grade levels in the school, standardized.",
    "question": "What is the average treatment effect of the growth mindset on the achievement score? Please use propensity score weighting in estimation, use logistic regression to estimate the propensity score, and round the final answer to the nearest hundredth.",
    "answer": "0.39",
    "data": [
      "/data/qrdata/data/learning_mindset.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 11",
      "keywords": [
        "Causality",
        "Observational data",
        "Matching",
        "Propensity score",
        "Average treatment effect"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The National Study of Learning Mindsets is a randomised study conducted in U.S. public high schools which aims at finding the impact of a growth mindset. The way it works is that students receive from the school a seminar to instill in them a growth mindset. Then, they follow up the students in their college years to measure how well they’ve performed academically. This measurement was compiled into an achievement score and standardized. The CSV file learning_mindset.csv contains simulated data of this research.\nVariable Description\nintervention: the intervention of the growth mindset;\nachievement_score: the standardized academic achievement score;\nschoolid: identifier of the student’s school;\nsuccess_expect: self-reported expectations for success in the future, a proxy for prior achievement, measured prior to random assignment;\nethnicity: categorical variable for student race/ethnicity;\ngender: categorical variable for student identified gender;\nfrst_in_family: categorical variable for student first-generation status, i.e. first in family to go to college;\nschool_urbanicity: school-level categorical variable for urbanicity of the school, i.e. rural, suburban, etc;\nschool_mindset: school-level mean of students’ fixed mindsets, reported prior to random assignment, standardized;\nschool_achievement: school achievement level, as measured by test scores and college preparation for the previous 4 cohorts of students, standardized;\nschool_ethnic_minority: school racial/ethnic minority composition, i.e., percentage of student body that is Black, Latino, or Native American, standardized;\nschool_poverty: school poverty concentration, i.e., percentage of students who are from families whose incomes fall below the federal poverty line, standardized;\nschool_size: total number of students in all four grade levels in the school, standardized.",
    "question": "What is the average treatment effect of the growth mindset on the achievement score? Please conduct doubly robust estimation, use logistic regression to estimate the propensity score, and round the final answer to the nearest hundredth.",
    "answer": "0.39",
    "data": [
      "/data/qrdata/data/learning_mindset.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 12",
      "keywords": [
        "Causality",
        "Observational data",
        "Matching",
        "Propensity score",
        "Average treatment effect"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The dataset billboard_impact.csv details information from a quasi-experiment assessing the influence of billboards on bank deposits in two cities: Porto Alegre (treatment group) and Florianopolis (control group). The csv file contains records with three variables: deposits (average bank deposits in Brazilian Reais), poa (A dummy indicator for the city of Porto Alegre. When it is zero, it means the samples are from Florianopolis.), and jul (A dummy for the month of July, or for the post intervention period. When it is zero it refers to samples from May, the pre-intervention period).",
    "question": "What was the average increase in bank deposits per customer in Porto Alegre after the billboard intervention, as estimated by the difference-in-differences approach? Please round the final answer to the nearest hundredth.",
    "answer": "6.52",
    "data": [
      "/data/qrdata/data/billboard_impact.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 13",
      "keywords": [
        "Causality",
        "Observational data",
        "Difference in differences"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the effect of cigarette taxation on its consumption, data from cigarette sales were collected and analyzed across 39 states in the United States from the years 1970 to 2000. Proposition 99, a Tobacco Tax and Health Protection Act passed in California in 1988, imposed a 25-cent per pack state excise tax on tobacco cigarettes and implemented additional restrictions, including the ban on cigarette vending machines in public areas accessible by juveniles and a ban on the individual sale of single cigarettes. Revenue generated was allocated for environmental and health care programs along with anti-tobacco advertising. We aim to determine if the imposition of this tax and the subsequent regulations led to a reduction in cigarette sales. The data is in the CSV file smoking2.csv.",
    "question": "By the year 2000, what was the estimated reduction in per-capita cigarette sales (in packs) in California as a result of Proposition 99, based on the synthetic control method? Please round the final answer to the nearest hundredth.",
    "answer": "24.83",
    "data": [
      "/data/qrdata/data/smoking2.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 15",
      "keywords": [
        "Causality",
        "Observational data",
        "Synthetic control"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the impacts of alcohol on death, we could use the fact that legal drinking age imposes a discontinuity on nature. In the US, those just under 21 years don’t drink (or drink much less) while those just older than 21 do drink. The csv file drinking.csv contains mortality data aggregated by age. Each row is the average age of a group of people and the average mortality by all causes (all), by moving vehicle accident (mva) and by suicide (suicide).",
    "question": "How much is the effect of alcohol consumption on death of all causes at 21 years? Please only consider data from people that are no older than 22 years and no younger than 20 years. Please answer with the magnitude of change in the number of deaths and round to the nearest hundredth.",
    "answer": "0.10",
    "data": [
      "/data/qrdata/data/drinking.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 16",
      "keywords": [
        "Causality",
        "Observational data",
        "Difference in differences"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To evaluate the effectiveness of email marketing on customer investment decisions, the dataset invest_email_rnd_train.csv is used for training, and invest_email_rnd_train.csv is used for testing. They contain customer data including id, age, income, insurance, invested amount, binary indicators for whether they received three different emails (em1, em2, em3), and a binary outcome variable indicating whether the customer converted (invested vs. didn't invest) after receiving an email. The goal is to personalize email marketing by sending email-1 only to customers predicted to have the highest increase in the probability of conversion.",
    "question": "Please train a boosted tree model to predict the conditional average treatment effect (CATE) of sending email-1 (em1) on the conversion rate, using demographic and financial attributes of the customers. What is the predicted increase in the probability that customer 6958 will purchase the investment product, if they are sent email-1, according to the CATE estimate provided by the model? Please provide the answer as a percentage, rounded to two decimal places.",
    "answer": "10.57%",
    "data": [
      "/data/qrdata/data/invest_email_rnd_train.csv",
      "/data/qrdata/data/invest_email_rnd_test.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 20",
      "keywords": [
        "Causality",
        "Observational data",
        "Conditional Average Treatment Effect",
        "Target transformation"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To find the effect of price on ice cream sales, the dataset ice_cream_sales.csv is used for training, and ice_cream_sales_rnd.csv is used for testing. Our test set has randomly assigned prices but our training data has only observational prices, which is potentially biased. Each unit is a day. For each day, we know which day of a week it is, what was the cost we had to make the ice cream (you can think of the cost as a proxy for quality) and the average temperature for that day. Then, we have our treatment, price, and our outcome, the number of ice cream sold.",
    "question": "Using the debiased machine learning approach after correcting for confounding biases related to temperature, cost, and weekday effects, what is the estimated average treatment effect (ATE) of the ice cream price on sales, rounded to two decimal places?",
    "answer": "-3.92",
    "data": [
      "/data/qrdata/data/ice_cream_sales.csv",
      "/data/qrdata/data/ice_cream_sales_rnd.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 22",
      "keywords": [
        "Causality",
        "Observational data",
        "Average treatment effect",
        "Debiased ML"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the effect of cigarette taxation on its consumption, data from cigarette sales were collected and analyzed across 39 states in the United States from the years 1970 to 2000. Proposition 99, a Tobacco Tax and Health Protection Act passed in California in 1988, imposed a 25-cent per pack state excise tax on tobacco cigarettes and implemented additional restrictions, including the ban on cigarette vending machines in public areas accessible by juveniles and a ban on the individual sale of single cigarettes. Revenue generated was allocated for environmental and health care programs along with anti-tobacco advertising. We aim to determine if the imposition of this tax and the subsequent regulations led to a reduction in cigarette sales. The data is in the CSV file smoking2.csv.",
    "question": "What is the Average Treatment Effect on the treated (ATT) of Proposition 99 on cigarette sales, as estimated by the difference-in-differences approach? Please round the final answer to the nearest hundredth. Do not consider variables other than cigsale, california, and after_treatment. The answer is positive if the proposition increases the sale.",
    "answer": "-27.35",
    "data": [
      "/data/qrdata/data/smoking2.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 25",
      "keywords": [
        "Causality",
        "Observational data",
        "Average treatment effect on the treated",
        "Difference in differences"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "To estimate the effect of cigarette taxation on its consumption, data from cigarette sales were collected and analyzed across 39 states in the United States from the years 1970 to 2000. Proposition 99, a Tobacco Tax and Health Protection Act passed in California in 1988, imposed a 25-cent per pack state excise tax on tobacco cigarettes and implemented additional restrictions, including the ban on cigarette vending machines in public areas accessible by juveniles and a ban on the individual sale of single cigarettes. Revenue generated was allocated for environmental and health care programs along with anti-tobacco advertising. We aim to determine if the imposition of this tax and the subsequent regulations led to a reduction in cigarette sales. The data is in the CSV file smoking2.csv.",
    "question": "What is the Average Treatment Effect on the treated (ATT) of Proposition 99 on cigarette sales, as estimated by the synthetic control approach? Please round the final answer to the nearest hundredth. Do not consider variables other than year, state, cigsale, california, and after_treatment. The answer is positive if the proposition increases the sale.",
    "answer": "-19.51",
    "data": [
      "/data/qrdata/data/smoking2.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 25",
      "keywords": [
        "Causality",
        "Observational data",
        "Average treatment effect on the treated",
        "Synthetic control"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "We are trying to estimate the effect of a trainee program on earnings. Data in the CSV file trainee_unique_on_age.csv contains trainee status, age, and earnings of each unit. Trainees are much younger than non trainees, which indicates that age is probably a confounder.",
    "question": "What is the Average Treatment Effect on the treated (ATT) of the trainee program on earnings, as estimated by a matching estimator? Please round the final answer to the nearest hundredth. The answer is positive if the trainee program increases the earnings.",
    "answer": "2457.89",
    "data": [
      "/data/qrdata/data/trainee_unique_on_age.csv"
    ],
    "metadata": {
      "reference": "Causal Inference for the Brave and True 10",
      "keywords": [
        "Causality",
        "Observational data",
        "Average treatment effect on the treated",
        "Matching"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Does racial discrimination exist in the labor market? Or, should racial disparities in the unemployment rate be attributed to other factors such as racial gaps in\neducational attainment? To answer this question, two social scientists conducted the following experiment.1 In response to newspaper ads, the researchers sent out\nrésumés of fictitious job candidates to potential employers. They varied only the names of job applicants, while leaving the other information in the résumés unchanged.\nFor some candidates, stereotypically African-American-sounding names such as Lakisha Washington or Jamal Jones were used, whereas other résumés contained\nstereotypically white-sounding names, such as Emily Walsh or Greg Baker. The researchers then compared the callback rates between these two groups and examined\nwhether applicants with stereotypically black names received fewer callbacks than those with stereotypically white names. The positions to which the applications were\nsent were either in sales, administrative support, clerical, or customer services.\n\nData is in the CSV data file resume.csv.\nVariable Description\nfirstname: first name of the fictitious job applicant\nsex: sex of applicant (female or male)\nrace: race of applicant (black or white)\ncall: whether a callback was made (1= yes, 0= no)",
    "question": "What is the racial gap in callback rate among white and black female applicants? Please round to the nearest thousandth.",
    "answer": "0.033",
    "data": [
      "/data/qrdata/data/resume.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.2.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Does racial discrimination exist in the labor market? Or, should racial disparities in the unemployment rate be attributed to other factors such as racial gaps in\neducational attainment? To answer this question, two social scientists conducted the following experiment.1 In response to newspaper ads, the researchers sent out\nrésumés of fictitious job candidates to potential employers. They varied only the names of job applicants, while leaving the other information in the résumés unchanged.\nFor some candidates, stereotypically African-American-sounding names such as Lakisha Washington or Jamal Jones were used, whereas other résumés contained\nstereotypically white-sounding names, such as Emily Walsh or Greg Baker. The researchers then compared the callback rates between these two groups and examined\nwhether applicants with stereotypically black names received fewer callbacks than those with stereotypically white names. The positions to which the applications were\nsent were either in sales, administrative support, clerical, or customer services.\n\nData is in the CSV data file resume.csv.\nVariable Description\nfirstname: first name of the fictitious job applicant\nsex: sex of applicant (female or male)\nrace: race of applicant (black or white)\ncall: whether a callback was made (1= yes, 0= no)",
    "question": "What is the racial gap in callback rate among white and black male applicants? Please round to the nearest thousandth.",
    "answer": "0.030",
    "data": [
      "/data/qrdata/data/resume.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.2.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Does racial discrimination exist in the labor market? Or, should racial disparities in the unemployment rate be attributed to other factors such as racial gaps in\neducational attainment? To answer this question, two social scientists conducted the following experiment.1 In response to newspaper ads, the researchers sent out\nrésumés of fictitious job candidates to potential employers. They varied only the names of job applicants, while leaving the other information in the résumés unchanged.\nFor some candidates, stereotypically African-American-sounding names such as Lakisha Washington or Jamal Jones were used, whereas other résumés contained\nstereotypically white-sounding names, such as Emily Walsh or Greg Baker. The researchers then compared the callback rates between these two groups and examined\nwhether applicants with stereotypically black names received fewer callbacks than those with stereotypically white names. The positions to which the applications were\nsent were either in sales, administrative support, clerical, or customer services.\n\nData is in the CSV data file resume.csv.\nVariable Description\nfirstname: first name of the fictitious job applicant\nsex: sex of applicant (female or male)\nrace: race of applicant (black or white)\ncall: whether a callback was made (1= yes, 0= no)",
    "question": "What is the callback rate of the first name with the lowest callback rate? Please round to the nearest thousandth.",
    "answer": "0.022",
    "data": [
      "/data/qrdata/data/resume.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.2.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Does racial discrimination exist in the labor market? Or, should racial disparities in the unemployment rate be attributed to other factors such as racial gaps in\neducational attainment? To answer this question, two social scientists conducted the following experiment.1 In response to newspaper ads, the researchers sent out\nrésumés of fictitious job candidates to potential employers. They varied only the names of job applicants, while leaving the other information in the résumés unchanged.\nFor some candidates, stereotypically African-American-sounding names such as Lakisha Washington or Jamal Jones were used, whereas other résumés contained\nstereotypically white-sounding names, such as Emily Walsh or Greg Baker. The researchers then compared the callback rates between these two groups and examined\nwhether applicants with stereotypically black names received fewer callbacks than those with stereotypically white names. The positions to which the applications were\nsent were either in sales, administrative support, clerical, or customer services.\n\nData is in the CSV data file resume.csv.\nVariable Description\nfirstname: first name of the fictitious job applicant\nsex: sex of applicant (female or male)\nrace: race of applicant (black or white)\ncall: whether a callback was made (1= yes, 0= no)",
    "question": "What is the callback rate of the first name with the highest callback rate? Please round to the nearest thousandth.",
    "answer": "0.159",
    "data": [
      "/data/qrdata/data/resume.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.2.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Two social science researchers examined the impact of raising the minimum wage on employment in the fast-food industry. In 1992, the state of New Jersey (NJ) in the United States raised the minimum wage from $4.25 to $5.05 per hour. Did such an increase in the minimum wage reduce employment as economic theory predicts? As discussed above, answering this question requires inference about the NJ employment rate in the absence of such a raise in the minimum wage. Since this counterfactual outcome is not observable, we must somehow estimate it using observed data.\nOne possible strategy is to look at another state in which the minimum wage did not increase. For example, the researchers of this study chose the neighboring state, Pennsylvania (PA), on the grounds that NJ’s economy resembles that of Pennsylvania, and hence the fast-food restaurants in the two states are comparable. Under this cross-section comparison design, therefore, the fast-food restaurants in NJ serve as the treatment group receiving the treatment (i.e., the increase in the minimum wage), whereas those in PA represent the control group, which did not receive such a treatment. To collect pretreatment and outcome measures, the researchers surveyed the fast-food restaurants before and after the minimum wage increase. Specifically, they gathered information about the number of full-time employees, the number of part-time employees, and their hourly wages, for each restaurant.\n\nThe CSV file minwage.csv contains this data set.\nVariable Description\nchain: name of the fast-food restaurant chain\nlocation: location of the restaurants (centralNJ, northNJ, PA, shoreNJ, southNJ)\nwageBefore: wage before the minimum-wage increase\nwageAfter: wage after the minimum-wage increase\nfullBefore: number of full-time employees before the minimum-wage increase\nfullAfter: number of full-time employees after the minimum-wage increase\npartBefore: number of part-time employees before the minimum-wage increase\npartAfter: number of part-time employees after the minimum-wage increase",
    "question": "What is the difference in the proportion of full-time employees in NJ and PA? Please round to the nearest thousandth.",
    "answer": "0.048",
    "data": [
      "/data/qrdata/data/minwage.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.5.1",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Two social science researchers examined the impact of raising the minimum wage on employment in the fast-food industry. In 1992, the state of New Jersey (NJ) in the United States raised the minimum wage from $4.25 to $5.05 per hour. Did such an increase in the minimum wage reduce employment as economic theory predicts? As discussed above, answering this question requires inference about the NJ employment rate in the absence of such a raise in the minimum wage. Since this counterfactual outcome is not observable, we must somehow estimate it using observed data.\nOne possible strategy is to look at another state in which the minimum wage did not increase. For example, the researchers of this study chose the neighboring state, Pennsylvania (PA), on the grounds that NJ’s economy resembles that of Pennsylvania, and hence the fast-food restaurants in the two states are comparable. Under this cross-section comparison design, therefore, the fast-food restaurants in NJ serve as the treatment group receiving the treatment (i.e., the increase in the minimum wage), whereas those in PA represent the control group, which did not receive such a treatment. To collect pretreatment and outcome measures, the researchers surveyed the fast-food restaurants before and after the minimum wage increase. Specifically, they gathered information about the number of full-time employees, the number of part-time employees, and their hourly wages, for each restaurant.\n\nThe CSV file minwage.csv contains this data set.\nVariable Description\nchain: name of the fast-food restaurant chain\nlocation: location of the restaurants (centralNJ, northNJ, PA, shoreNJ, southNJ)\nwageBefore: wage before the minimum-wage increase\nwageAfter: wage after the minimum-wage increase\nfullBefore: number of full-time employees before the minimum-wage increase\nfullAfter: number of full-time employees after the minimum-wage increase\npartBefore: number of part-time employees before the minimum-wage increase\npartAfter: number of part-time employees after the minimum-wage increase",
    "question": "We would like to reduce the confounding bias due to different fast-food chains and the location of restaurants. We only consider Burger King restaurants, and focus on the Burger King restaurants located in northern and southern NJ that are near PA, while excluding those in the Jersey shore and central New Jersey. What is the difference in the proportion of full-time Burger King employees in NJ (northern and southern only) and PA? Please round to the nearest thousandth.",
    "answer": "0.032",
    "data": [
      "/data/qrdata/data/minwage.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.5.2",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Confounding bias"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Two social science researchers examined the impact of raising the minimum wage on employment in the fast-food industry. In 1992, the state of New Jersey (NJ) in the United States raised the minimum wage from $4.25 to $5.05 per hour. Did such an increase in the minimum wage reduce employment as economic theory predicts? As discussed above, answering this question requires inference about the NJ employment rate in the absence of such a raise in the minimum wage. Since this counterfactual outcome is not observable, we must somehow estimate it using observed data.\nOne possible strategy is to look at another state in which the minimum wage did not increase. For example, the researchers of this study chose the neighboring state, Pennsylvania (PA), on the grounds that NJ’s economy resembles that of Pennsylvania, and hence the fast-food restaurants in the two states are comparable. Under this cross-section comparison design, therefore, the fast-food restaurants in NJ serve as the treatment group receiving the treatment (i.e., the increase in the minimum wage), whereas those in PA represent the control group, which did not receive such a treatment. To collect pretreatment and outcome measures, the researchers surveyed the fast-food restaurants before and after the minimum wage increase. Specifically, they gathered information about the number of full-time employees, the number of part-time employees, and their hourly wages, for each restaurant.\n\nThe CSV file minwage.csv contains this data set.\nVariable Description\nchain: name of the fast-food restaurant chain\nlocation: location of the restaurants (centralNJ, northNJ, PA, shoreNJ, southNJ)\nwageBefore: wage before the minimum-wage increase\nwageAfter: wage after the minimum-wage increase\nfullBefore: number of full-time employees before the minimum-wage increase\nfullAfter: number of full-time employees after the minimum-wage increase\npartBefore: number of part-time employees before the minimum-wage increase\npartAfter: number of part-time employees after the minimum-wage increase",
    "question": "Estimate the average treatment effect of raising the mininum wage on the proportion of full-time employees using difference-in-differences (DiD). Please round to the nearest thousandth. The answer is positive if raising the mininum wage increases the proportion of full-time employees.",
    "answer": "-0.062",
    "data": [
      "/data/qrdata/data/minwage.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 2.5.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Difference in differences"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "After the September 11 attacks, the United States and its allies invaded Afghanistan with the goal of dismantling al-Qaeda, which had been operating there under the protection of the Taliban government. In 2003, the North Atlantic Treaty Organization (NATO) became involved in the conflict, sending in a coalition of international troops organized under the name of the International Security Assistance Force (ISAF). To wage this war against the Taliban insurgency, the ISAF engaged in a “hearts and minds” campaign, combining economic assistance, service delivery, and protection in order to win the support of civilians. To evaluate the success of such a campaign, it is essential to measure and understand civilians’ experiences and sentiments during the war. However, measuring the experiences and opinions of civilians during wartime is a challenging task because of harsh security conditions, posing potential threats to interviewers and respondents. This means that respondents may inaccurately answer survey questions in order to avoid giving socially undesirable responses.\nA group of social scientists conducted a public opinion survey in southern Afghanistan, the heartland of the insurgency. The survey was administered to a sample of 2754 respondents between January and February 2011. The researchers note that the participation rate was 89%. That is, they originally contacted 3097 males and 343 of them refused to take the survey. Because local culture prohibited interviewers from talking to female citizens, the respondents were all males.\n\nThe CSV file afghan.csv contains this data set.\nVariable Description\nprovince: province where the respondent lives\ndistrict: district where the respondent lives\nvillage.id: ID of the village where the respondent lives\nage: age of the respondent\neduc.years: years of education of the respondent\nemployed: whether the respondent is employed\nincome: monthly income of the respondent (five levels)\nviolent.exp.ISAF: whether the respondent experienced violence by ISAF\nviolent.exp.taliban: whether the respondent experienced violence by the Taliban\nlist.group: randomly assigned group for the list experiment (control, ISAF, taliban)\nlist.response: response to the list experiment question (0–4)",
    "question": "What proportion of respondents did not report the monthly income? Please round to the nearest thousandth.",
    "answer": "0.056",
    "data": [
      "/data/qrdata/data/afghan.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 3.2",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Missing value"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "After the September 11 attacks, the United States and its allies invaded Afghanistan with the goal of dismantling al-Qaeda, which had been operating there under the protection of the Taliban government. In 2003, the North Atlantic Treaty Organization (NATO) became involved in the conflict, sending in a coalition of international troops organized under the name of the International Security Assistance Force (ISAF). To wage this war against the Taliban insurgency, the ISAF engaged in a “hearts and minds” campaign, combining economic assistance, service delivery, and protection in order to win the support of civilians. To evaluate the success of such a campaign, it is essential to measure and understand civilians’ experiences and sentiments during the war. However, measuring the experiences and opinions of civilians during wartime is a challenging task because of harsh security conditions, posing potential threats to interviewers and respondents. This means that respondents may inaccurately answer survey questions in order to avoid giving socially undesirable responses.\nA group of social scientists conducted a public opinion survey in southern Afghanistan, the heartland of the insurgency. The survey was administered to a sample of 2754 respondents between January and February 2011. The researchers note that the participation rate was 89%. That is, they originally contacted 3097 males and 343 of them refused to take the survey. Because local culture prohibited interviewers from talking to female citizens, the respondents were all males.\n\nThe CSV file afghan.csv contains this data set.\nVariable Description\nprovince: province where the respondent lives\ndistrict: district where the respondent lives\nvillage.id: ID of the village where the respondent lives\nage: age of the respondent\neduc.years: years of education of the respondent\nemployed: whether the respondent is employed\nincome: monthly income of the respondent (five levels)\nviolent.exp.ISAF: whether the respondent experienced violence by ISAF\nviolent.exp.taliban: whether the respondent experienced violence by the Taliban\nlist.group: randomly assigned group for the list experiment (control, ISAF, taliban)\nlist.response: response to the list experiment question (0–4)\n\nIn the Afghanistan survey, within each of the five provinces of interest, the researchers sampled districts and then villages within each selected district. Within each sampled village, interviewers selected a household in an approximately random manner based on their location within the village, and finally administered a survey to a male respondent aged 16 years or older, who was sampled using the Kish grid method. While the probability of selecting each individual in the population is known only approximately, the method in theory should provide a roughly representative sample of the target population.\nWe examine the representativeness of the randomly sampled villages in the Afghanistan data. The data file afghan-village.csv contains the altitude and population of each village. For the population variable, it is customary to take the logarithmic transformation so that the distribution does not look too skewed with a small number of extremely large or small values.\n\nVariable Description\nvillage.surveyed: whether a village is sampled for survey\naltitude: altitude of the village\npopulation: population of the village",
    "question": "Does the survey sample appear to be representative of the population in altitude and population? A. Representative in neither aspect, B. Representative in altitude but not representative in population, C. Representative in population but not representative in altitude, D. Representative in both aspects. Please answer with A, B, C, or D.",
    "answer": "D",
    "data": [
      "/data/qrdata/data/afghan.csv",
      "/data/qrdata/data/afghan-village.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 3.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Distribition"
      ],
      "question_type": "multiple_choice",
      "multiple_choices": [
        "a",
        "b",
        "c",
        "d"
      ]
    }
  },
  {
    "context": "It has been pointed out that rising income inequality may be responsible for the widening partisan gap. To measure income inequality, we use the Gini coefficient (Gini index), which is best understood graphically. Figure 3.4 illustrates the idea. The horizontal axis represents the cumulative share of people.\n\nThe CSV data file, USGini.csv, contains the Gini coefficient from 1947 to 2013. We notice that both political polarization and income inequality have been steadily increasing in the United States.\n\nVariable Description\nyear: year\ngini: US Gini coefficient\npolarization: measure of political polarization",
    "question": "What is the correlation between the Gini coefficient and the measure of political polarization? Since each US congressional session lasts two years, we take the Gini coefficient for the second year of each session. Please round to the nearest thousandth.",
    "answer": "0.942",
    "data": [
      "/data/qrdata/data/USGini.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 3.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Correlation"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The United States’s unique Electoral College system makes predicting election outcomes challenging. A candidate is elected to office by winning an absolute majority of electoral votes. Each of the 538 electors casts a single electoral vote. As of 2016, 535 of these votes are allocated among 50 states, corresponding to the 435 members of the House of Representatives and the 100 members of the Senate. The remaining 3 votes are given to the District of Columbia. In most cases, the electors vote for the candidate who won the plurality of votes in the state they represent, leading to a “winner-takeall” system in these states. In fact, some states have criminal penalties for voting for the candidate who did not win the plurality of votes. A winning presidential candidate must obtain at least 270 electoral votes.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election. The election date was 11/4/2008.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file polls08.csv, which contains many polls within each state leading up to the election.\nVariable Description\nstate: abbreviated name of the state in which the poll was conducted\nObama: predicted support for Obama (percentage)\nMcCain: predicted support for McCain (percentage)\nPollster: name of the organization conducting the poll\nmiddate: middate of the period when the poll was conducted",
    "question": "What is the average prediction error of the margin between the percentage share of Obama and McCain across states? We only consider the latest polls within each state. If there are several latest polls, we consider the mean of them. Please round to the nearest thousandth.",
    "answer": "1.062",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/polls08.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.1.3",
      "keywords": [
        "Social science",
        "Statistics"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The United States’s unique Electoral College system makes predicting election outcomes challenging. A candidate is elected to office by winning an absolute majority of electoral votes. Each of the 538 electors casts a single electoral vote. As of 2016, 535 of these votes are allocated among 50 states, corresponding to the 435 members of the House of Representatives and the 100 members of the Senate. The remaining 3 votes are given to the District of Columbia. In most cases, the electors vote for the candidate who won the plurality of votes in the state they represent, leading to a “winner-takeall” system in these states. In fact, some states have criminal penalties for voting for the candidate who did not win the plurality of votes. A winning presidential candidate must obtain at least 270 electoral votes.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election. The election date was 11/4/2008.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file polls08.csv, which contains many polls within each state leading up to the election.\nVariable Description\nstate: abbreviated name of the state in which the poll was conducted\nObama: predicted support for Obama (percentage)\nMcCain: predicted support for McCain (percentage)\nPollster: name of the organization conducting the poll\nmiddate: middate of the period when the poll was conducted",
    "question": "What is the root mean squared prediction error of the margin between the percentage share of Obama and McCain across states? We only consider the latest polls within each state. If there are several latest polls, we consider the mean of them. Please round to the nearest thousandth.",
    "answer": "5.909",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/polls08.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.1.3",
      "keywords": [
        "Social science",
        "Statistics"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The United States’s unique Electoral College system makes predicting election outcomes challenging. A candidate is elected to office by winning an absolute majority of electoral votes. Each of the 538 electors casts a single electoral vote. As of 2016, 535 of these votes are allocated among 50 states, corresponding to the 435 members of the House of Representatives and the 100 members of the Senate. The remaining 3 votes are given to the District of Columbia. In most cases, the electors vote for the candidate who won the plurality of votes in the state they represent, leading to a “winner-takeall” system in these states. In fact, some states have criminal penalties for voting for the candidate who did not win the plurality of votes. A winning presidential candidate must obtain at least 270 electoral votes.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election. The election date was 11/4/2008.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file polls08.csv, which contains many polls within each state leading up to the election.\nVariable Description\nstate: abbreviated name of the state in which the poll was conducted\nObama: predicted support for Obama (percentage)\nMcCain: predicted support for McCain (percentage)\nPollster: name of the organization conducting the poll\nmiddate: middate of the period when the poll was conducted",
    "question": "How many states got poll results that were wrong? We only consider the latest polls within each state. If there are several latest polls, we consider the mean of them. Please answer with an integer.",
    "answer": "3",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/polls08.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.1.3",
      "keywords": [
        "Social science",
        "Statistics"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The United States’s unique Electoral College system makes predicting election outcomes challenging. A candidate is elected to office by winning an absolute majority of electoral votes. Each of the 538 electors casts a single electoral vote. As of 2016, 535 of these votes are allocated among 50 states, corresponding to the 435 members of the House of Representatives and the 100 members of the Senate. The remaining 3 votes are given to the District of Columbia. In most cases, the electors vote for the candidate who won the plurality of votes in the state they represent, leading to a “winner-takeall” system in these states. In fact, some states have criminal penalties for voting for the candidate who did not win the plurality of votes. A winning presidential candidate must obtain at least 270 electoral votes.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election. The election date was 11/4/2008.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file polls08.csv, which contains many polls within each state leading up to the election.\nVariable Description\nstate: abbreviated name of the state in which the poll was conducted\nObama: predicted support for Obama (percentage)\nMcCain: predicted support for McCain (percentage)\nPollster: name of the organization conducting the poll\nmiddate: middate of the period when the poll was conducted",
    "question": "What is the gap between the actual and predicted total number of electoral votes won by Obama? We only consider the latest polls within each state. If there are several latest polls, we consider the mean of them. Please answer with an integer.",
    "answer": "15",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/polls08.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.1.3",
      "keywords": [
        "Social science",
        "Statistics"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Several psychologists have reported the intriguing result of an experiment showing that facial appearance predicts election outcomes better than chance. In their experiment, the researchers briefly showed student subjects the black-and-white head shots of two candidates from a US congressional election (winner and runner-up). The exposure of subjects to facial pictures lasted less than a second, and the subjects were then asked to evaluate the two candidates in terms of their perceived competence.\nThe researchers used these competence measures to predict election outcomes. The key hypothesis is whether or not a within-a-second evaluation of facial appearance can predict election outcomes. The CSV data set, face.csv, contains the data from the experiment. Note that we include data only from subjects who did not know the candidates’ political parties, their policies, or even which candidate was the incumbent or challenger. They were simply making snap judgments about which candidate appeared more competent based on their facial expression alone.\n\nVariable Description\ncongress: session of Congress\nyear: year of the election\nstate: state of the election\nwinner: name of the winner\nloser: name of the runner-up\nw.party: party of the winner\nl.party: party of the loser\nd.votes: number of votes for the Democratic candidate\nr.votes: number of votes for the Republican candidate\nd.comp: competence measure for the Democratic candidate\nr.comp: competence measure for the Republican candidate",
    "question": "What is the correlation between the perceived competence of the Democratic candidate and the vote share differential of the Democratic candidate minus the Republican candidate? Please round to the nearest thousandth.",
    "answer": "0.433",
    "data": [
      "/data/qrdata/data/face.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.2",
      "keywords": [
        "Social science",
        "Statistics",
        "Correlation"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Several psychologists have reported the intriguing result of an experiment showing that facial appearance predicts election outcomes better than chance. In their experiment, the researchers briefly showed student subjects the black-and-white head shots of two candidates from a US congressional election (winner and runner-up). The exposure of subjects to facial pictures lasted less than a second, and the subjects were then asked to evaluate the two candidates in terms of their perceived competence.\nThe researchers used these competence measures to predict election outcomes. The key hypothesis is whether or not a within-a-second evaluation of facial appearance can predict election outcomes. The CSV data set, face.csv, contains the data from the experiment. Note that we include data only from subjects who did not know the candidates’ political parties, their policies, or even which candidate was the incumbent or challenger. They were simply making snap judgments about which candidate appeared more competent based on their facial expression alone.\n\nVariable Description\ncongress: session of Congress\nyear: year of the election\nstate: state of the election\nwinner: name of the winner\nloser: name of the runner-up\nw.party: party of the winner\nl.party: party of the loser\nd.votes: number of votes for the Democratic candidate\nr.votes: number of votes for the Republican candidate\nd.comp: competence measure for the Democratic candidate\nr.comp: competence measure for the Republican candidate",
    "question": "Fit a linear regression model using the Democratic margin in the two-party vote share as the response variable and the perceived competence for Democratic candidates as the predictor. What is the estimated slope of the model? Please round to the nearest thousandth.",
    "answer": "0.660",
    "data": [
      "/data/qrdata/data/face.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Correlation",
        "Linear regression"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Several psychologists have reported the intriguing result of an experiment showing that facial appearance predicts election outcomes better than chance. In their experiment, the researchers briefly showed student subjects the black-and-white head shots of two candidates from a US congressional election (winner and runner-up). The exposure of subjects to facial pictures lasted less than a second, and the subjects were then asked to evaluate the two candidates in terms of their perceived competence.\nThe researchers used these competence measures to predict election outcomes. The key hypothesis is whether or not a within-a-second evaluation of facial appearance can predict election outcomes. The CSV data set, face.csv, contains the data from the experiment. Note that we include data only from subjects who did not know the candidates’ political parties, their policies, or even which candidate was the incumbent or challenger. They were simply making snap judgments about which candidate appeared more competent based on their facial expression alone.\n\nVariable Description\ncongress: session of Congress\nyear: year of the election\nstate: state of the election\nwinner: name of the winner\nloser: name of the runner-up\nw.party: party of the winner\nl.party: party of the loser\nd.votes: number of votes for the Democratic candidate\nr.votes: number of votes for the Republican candidate\nd.comp: competence measure for the Democratic candidate\nr.comp: competence measure for the Republican candidate",
    "question": "Fit a linear regression model using the Democratic margin in the two-party vote share as the response variable and the perceived competence for Democratic candidates as the predictor. What is the root-mean-squared error (RMSE) of the model estimation? Please round to the nearest thousandth.",
    "answer": "0.264",
    "data": [
      "/data/qrdata/data/face.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Correlation",
        "Linear regression",
        "Root-mean-squared error"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Regression towards the mean represents an empirical phenomenon where an observation with a value of the predictor further away from the distribution’s mean tends to have a value of an outcome variable closer to that mean. This tendency can be explained by chance alone.\n\nWe will examine whether or not the US presidential election data exhibit the regression towards the mean phenomenon. To do this, we use Obama’s vote share in the 2008 election to predict his vote share in his 2012 reelection.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file pres12.csv, which contains the election results by state of the 2012 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nObama: Obama’s vote share (percentage)\nRomney: Romney’s vote share (percentage)\nEV: number of Electoral College votes for the state",
    "question": "We standardize vote shares across elections by computing their z-scores so that we can measure Obama’s electoral performance in each state relative to his average performance of that year. That is, we subtract the mean from Obama’s vote share in each election and then divide it by the standard deviation. We regress Obama’s 2012 standardized vote share on his 2008 standardized vote share. What is the estimated slope of the regression model? Please round to the nearest thousandth.",
    "answer": "0.983",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/pres12.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Correlation",
        "Linear regression"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Regression towards the mean represents an empirical phenomenon where an observation with a value of the predictor further away from the distribution’s mean tends to have a value of an outcome variable closer to that mean. This tendency can be explained by chance alone.\n\nWe will examine whether or not the US presidential election data exhibit the regression towards the mean phenomenon. To do this, we use Obama’s vote share in the 2008 election to predict his vote share in his 2012 reelection.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file pres12.csv, which contains the election results by state of the 2012 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nObama: Obama’s vote share (percentage)\nRomney: Romney’s vote share (percentage)\nEV: number of Electoral College votes for the state",
    "question": "We standardize vote shares across elections by computing their z-scores so that we can measure Obama’s electoral performance in each state relative to his average performance of that year. That is, we subtract the mean from Obama’s vote share in each election and then divide it by the standard deviation.\nWhat is the proportion of states where Obama received a greater share of standardized votes in 2012 than he did in 2008, in the bottom quartile of Obama’s 2008 (standardized) vote share? Please round to the nearest thousandth.",
    "answer": "0.571",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/pres12.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Quartile"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Regression towards the mean represents an empirical phenomenon where an observation with a value of the predictor further away from the distribution’s mean tends to have a value of an outcome variable closer to that mean. This tendency can be explained by chance alone.\n\nWe will examine whether or not the US presidential election data exhibit the regression towards the mean phenomenon. To do this, we use Obama’s vote share in the 2008 election to predict his vote share in his 2012 reelection.\n\nThe CSV data file pres08.csv contains the election results by state of the 2008 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nstate.name: unabbreviated name of the state\nObama: Obama’s vote share (percentage)\nMcCain: McCain’s vote share (percentage)\nEV: number of Electoral College votes for the state\n\nIn addition, we have the CSV file pres12.csv, which contains the election results by state of the 2012 US presidential election.\nVariable Description\nstate: abbreviated name of the state\nObama: Obama’s vote share (percentage)\nRomney: Romney’s vote share (percentage)\nEV: number of Electoral College votes for the state",
    "question": "We standardize vote shares across elections by computing their z-scores so that we can measure Obama’s electoral performance in each state relative to his average performance of that year. That is, we subtract the mean from Obama’s vote share in each election and then divide it by the standard deviation.\nWhat is the proportion of states where Obama received a greater share of standardized votes in 2012 than he did in 2008, in the top quartile of Obama’s 2008 (standardized) vote share? Please round to the nearest thousandth.",
    "answer": "0.461",
    "data": [
      "/data/qrdata/data/pres08.csv",
      "/data/qrdata/data/pres12.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.5",
      "keywords": [
        "Social science",
        "Statistics",
        "Quartile"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Consider the problem of predicting the 2000 US election results in Florida using the 1996 US election results from the same state at the county level. In Florida, there are 68 counties, and the CSV file florida.csv contains the number of votes cast for each candidate in those two elections.\n\nVariable Description\ncounty: county name\nClinton96: Clinton’s votes in 1996\nDole96: Dole’s votes in 1996\nPerot96: Perot’s votes in 1996\nBush00: Bush’s votes in 2000\nGore00: Gore’s votes in 2000\nBuchanan00: Buchanan’s votes in 2000",
    "question": "We focus on libertarian candidates Ross Perot in 1996 and Pat Buchanan in 2000, using the votes for the former to predict the votes for the latter with a linear regression model. What is the Coefficient of determination of the model? Please round to the nearest thousandth.\nThe coefficient of determination is a measure of model fit and represents the proportion of variation in the outcome variable explained by the predictor. It is defined as one minus the ratio of the sum of squared residuals (SSR) to the total sum of squares (TSS).",
    "answer": "0.513",
    "data": [
      "/data/qrdata/data/florida.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.6",
      "keywords": [
        "Social science",
        "Statistics",
        "Linear regression",
        "Coefficient of determination"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Consider the problem of predicting the 2000 US election results in Florida using the 1996 US election results from the same state at the county level. In Florida, there are 68 counties, and the CSV file florida.csv contains the number of votes cast for each candidate in those two elections.\n\nVariable Description\ncounty: county name\nClinton96: Clinton’s votes in 1996\nDole96: Dole’s votes in 1996\nPerot96: Perot’s votes in 1996\nBush00: Bush’s votes in 2000\nGore00: Gore’s votes in 2000\nBuchanan00: Buchanan’s votes in 2000",
    "question": "We focus on libertarian candidates Ross Perot in 1996 and Pat Buchanan in 2000, using the votes for the former to predict the votes for the latter with a linear regression model. As the Palm Beach county is an outlier, we fit the model without Palm Beach county. What is the Coefficient of determination of the model? Please round to the nearest thousandth.\nThe coefficient of determination is a measure of model fit and represents the proportion of variation in the outcome variable explained by the predictor. It is defined as one minus the ratio of the sum of squared residuals (SSR) to the total sum of squares (TSS).",
    "answer": "0.851",
    "data": [
      "/data/qrdata/data/florida.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.2.6",
      "keywords": [
        "Social science",
        "Statistics",
        "Linear regression",
        "Coefficient of determination",
        "Outlier"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Researchers conducted randomized policy experiment in India where, since the mid-1990s, one-third of village council heads have been randomly reserved for female politicians. The CSV data set women.csv contains a subset of this data from West Bengal. The policy was implemented at the level of government called Gram Panchayat or GP. Each GP contains many villages. For this study, two villages were selected at random within each GP for detailed data collection. Each observation in the data set represents a village and there are two villages associated with each GP.\n\nVariable Description\nGP: identifier for the Gram Panchayat (GP)\nvillage: identifier for each village\nreserved: binary variable indicating whether the GP was reserved for women leaders or not\nfemale: binary variable indicating whether the GP had a female leader or not\nirrigation: variable measuring the number of new or repaired irrigation facilities in the village since the reserve policy started\nwater: variable measuring the number of new or repaired drinking water facilities in the village since the reservation policy started",
    "question": "To explore if female politicians are more likely to support policies about drinking water, please estimate the average causal effects of the reservation policy on the number of new or repaired drinking water facilities since the reserve policy started. Please round to the nearest hundredth.",
    "answer": "9.25",
    "data": [
      "/data/qrdata/data/women.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.3.1",
      "keywords": [
        "Social science",
        "Causality",
        "Average treatment effect",
        "Interventional data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Three social scientists conducted an RCT in which they investigated whether social pressure within neighborhoods increases participation. Specifically, during a primary election in the state of Michigan, they randomly assigned registered voters to receive different get-out-the-vote (GOTV) messages and examined whether sending postcards with these messages increased turnout. The researchers exploited the fact that the turnout of individual voters is public information in the United States. The GOTV message of particular interest was designed to induce social pressure by telling voters that after the election their neighbors would be informed about whether they voted in the election or not. The researchers hypothesized that such a namingand-shaming GOTV strategy would increase participation.\nThere are three treatment groups: voters who receive either the social pressure message, the civic duty message, or the Hawthorne effect message. The Hawthorne effect refers to the phenomenon where study subjects behave differently because they know they are being observed by researchers. The experiment also has a control group which consists of those voters receiving no message. The researchers randomly assigned each voter to one of the four groups and examined whether the voter turnout was different across the groups.\n\nThe data is in the file social.csv.\nVariable Description\nhhsize: household size of the voter\nmessages: GOTV messages the voter received (Civic Duty, Control, Neighbors, Hawthorne)\nsex: sex of the voter (female or male)\nyearofbirth: year of birth of the voter\nprimary2004: whether the voter voted in the 2004 primary election (1=voted, 0=abstained)\nprimary2006: whether the voter turned out in the 2006 primary election (1=voted, 0=abstained)",
    "question": "What is the difference in the average causal effect of the Neighbors message on whether the voter voted in the 2006 primary election between those who voted in the 2004 primary election and those who did not? Please round the result to the nearest thousandth.",
    "answer": "0.027",
    "data": [
      "/data/qrdata/data/social.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.3.3",
      "keywords": [
        "Social science",
        "Causality",
        "Heterogenous treatment effects",
        "Interventional data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Three social scientists conducted an RCT in which they investigated whether social pressure within neighborhoods increases participation. Specifically, during a primary election in the state of Michigan, they randomly assigned registered voters to receive different get-out-the-vote (GOTV) messages and examined whether sending postcards with these messages increased turnout. The researchers exploited the fact that the turnout of individual voters is public information in the United States. The GOTV message of particular interest was designed to induce social pressure by telling voters that after the election their neighbors would be informed about whether they voted in the election or not. The researchers hypothesized that such a namingand-shaming GOTV strategy would increase participation.\nThere are three treatment groups: voters who receive either the social pressure message, the civic duty message, or the Hawthorne effect message. The Hawthorne effect refers to the phenomenon where study subjects behave differently because they know they are being observed by researchers. The experiment also has a control group which consists of those voters receiving no message. The researchers randomly assigned each voter to one of the four groups and examined whether the voter turnout was different across the groups.\n\nThe data is in the file social.csv.\nVariable Description\nhhsize: household size of the voter\nmessages: GOTV messages the voter received (Civic Duty, Control, Neighbors, Hawthorne)\nsex: sex of the voter (female or male)\nyearofbirth: year of birth of the voter\nprimary2004: whether the voter voted in the 2004 primary election (1=voted, 0=abstained)\nprimary2006: whether the voter turned out in the 2006 primary election (1=voted, 0=abstained)",
    "question": "What is the average causal effect of the Neighbors message on whether the voter voted in the 2006 primary election if the voter's age was 25 in 2006? Please estimate a linear regression model with both message and age as predictors, and round the result to the nearest thousandth.",
    "answer": "0.064",
    "data": [
      "/data/qrdata/data/social.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.3.3",
      "keywords": [
        "Social science",
        "Causality",
        "Heterogenous treatment effects",
        "Interventional data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Three social scientists conducted an RCT in which they investigated whether social pressure within neighborhoods increases participation. Specifically, during a primary election in the state of Michigan, they randomly assigned registered voters to receive different get-out-the-vote (GOTV) messages and examined whether sending postcards with these messages increased turnout. The researchers exploited the fact that the turnout of individual voters is public information in the United States. The GOTV message of particular interest was designed to induce social pressure by telling voters that after the election their neighbors would be informed about whether they voted in the election or not. The researchers hypothesized that such a namingand-shaming GOTV strategy would increase participation.\nThere are three treatment groups: voters who receive either the social pressure message, the civic duty message, or the Hawthorne effect message. The Hawthorne effect refers to the phenomenon where study subjects behave differently because they know they are being observed by researchers. The experiment also has a control group which consists of those voters receiving no message. The researchers randomly assigned each voter to one of the four groups and examined whether the voter turnout was different across the groups.\n\nThe data is in the file social.csv.\nVariable Description\nhhsize: household size of the voter\nmessages: GOTV messages the voter received (Civic Duty, Control, Neighbors, Hawthorne)\nsex: sex of the voter (female or male)\nyearofbirth: year of birth of the voter\nprimary2004: whether the voter voted in the 2004 primary election (1=voted, 0=abstained)\nprimary2006: whether the voter turned out in the 2006 primary election (1=voted, 0=abstained)",
    "question": "What is the average causal effect of the Neighbors message on whether the voter voted in the 2006 primary election if the voter's age was 65 in 2006? Please estimate a linear regression model with both message and age as predictors, and round the result to the nearest thousandth.",
    "answer": "0.089",
    "data": [
      "/data/qrdata/data/social.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.3.3",
      "keywords": [
        "Social science",
        "Causality",
        "Heterogenous treatment effects",
        "Interventional data"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "We consider how much politicians can increase their personal wealth due to holding office. Scholars investigated this question by analyzing members of Parliament (MPs) in the United Kingdom.6 The authors of the original study collected information about personal wealth at the time of death for several hundred competitive candidates who ran for office in general elections between 1950 and 1970.\n\nThe data are contained in the CSV file MPs.csv.\nVariable Description\nsurname: surname of the candidate\nfirstname: first name of the candidate\nparty: party of the candidate (labour or tory)\nln.gross: log gross wealth at the time of death\nln.net: log net wealth at the time of death\nyob: year of birth of the candidate\nyod: year of death of the candidate\nmargin.pre: margin of the candidate’s party in the previous election\nregion: electoral region\nmargin: margin of victory (vote share)",
    "question": "What is the average causal effect of becoming members of Parliament on the log net wealth for Tory candidates? The difference in predicted values at the point of discontinuity, i.e., a zero margin of victory, between the two regressions represents the average causal effect on personal wealth of serving as an MP. Please round the result to the nearest integer.",
    "answer": "255051",
    "data": [
      "/data/qrdata/data/MPs.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 4.3.4",
      "keywords": [
        "Social science",
        "Causality",
        "Average treatment effect",
        "Discontinuity"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file, florentine.csv, contains an adjacency matrix whose entries represent the existence of relationships between two units (one unit represented by the row and the\nother represented by the column). Specifically, there are 16 elite Florentine families in the data, leading to a 16×16 adjacency matrix. If the (i, j) entry of this adjacency matrix is 1, then it implies that the ith and jth Florentine families had a marriage relationship. In contrast, a value of 0 indicates the absence of a marriage.",
    "question": "The closeness of a node is a measure of centrality in a network, calculated as the reciprocal of the sum of the length of the shortest paths between the node and all other nodes in the graph. What is the closeness of the GUADAGNI family? Please round the result to the nearest thousandth.",
    "answer": "0.022",
    "data": [
      "/data/qrdata/data/florentine.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 5.2.2",
      "keywords": [
        "Social science",
        "Statistics",
        "Network data",
        "Centrality",
        "Closeness"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "The CSV file, florentine.csv, contains an adjacency matrix whose entries represent the existence of relationships between two units (one unit represented by the row and the\nother represented by the column). Specifically, there are 16 elite Florentine families in the data, leading to a 16×16 adjacency matrix. If the (i, j) entry of this adjacency matrix is 1, then it implies that the ith and jth Florentine families had a marriage relationship. In contrast, a value of 0 indicates the absence of a marriage.",
    "question": "The betweenness of a node is a measure of centrality in a network. According to this measure, a node is considered to be central if it is responsible for connecting other nodes. In particular, if we assume that communication between a pair of nodes occurs through the shortest path between them, a node that lies on many such shortest paths may possess special leverage within a network. For a given node v, we calculate betweenness in three steps. First, compute the proportion of the shortest paths between two other nodes, t and u, that contain v. For example, two shortest paths occur between the Albizzi family and Tornabuon family, but we want only the proportion that contain v. Second, calculate this proportion for every unique pair of nodes t and u in the graph, excluding v. Third, sum all proportions.\nWhat is the betweenness of the BISCHERI family? Please round the result to the nearest thousandth.",
    "answer": "9.500",
    "data": [
      "/data/qrdata/data/florentine.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 5.2.2",
      "keywords": [
        "Social science",
        "Statistics",
        "Network data",
        "Centrality",
        "Betweenness"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "We analyze Twitter-following data among US senators as directed network data. In this data set, an edge represents an instance of a senator following another senator’s Twitter account. The data consist of two files, one listing pairs of the Twitter screen names of the following and followed politicians, twitter-following.csv, and the other containing information about each politician, twitter-senator.csv.\n\nVariable Description\nTwitter-following data\nfollowing: Twitter screen name of the following senator\nfollowed: Twitter screen name of the followed senator\nTwitter senator data\nscreen_name: Twitter screen name\nname: name of senator\nparty: party (D = Democrat, R = Republican, I = Independent)\nstate: state abbreviation",
    "question": "What is the indegree of the senator with the greatest value of indegree? Please output an integer.",
    "answer": "63",
    "data": [
      "/data/qrdata/data/twitter-following.csv",
      "/data/qrdata/data/twitter-senator.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 5.2.3",
      "keywords": [
        "Social science",
        "Statistics",
        "Network data",
        "Centrality",
        "Degree"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Here is a random sample of 10,000 registered voters from Florida contained in the CSV file FLVoters.csv.\nVariable Description\nsurname: surname\ncounty: county ID of the voter’s residence\nVTD: voting district ID of the voter’s residence\nage: age\ngender: gender m = male and f = female\nrace: self-reported race\n\nThe US Census Bureau releases a list of common surnames with their frequency. For example, the most common surname was “Smith” with 2,376,206 occurrences, followed by “Johnson” and “Williams” with 1,857,160 and 1,534,042, respectively. This data set is quite comprehensive, including a total of more than 150,000 surnames that occurred at least 100 times. In addition, the census provides the relative frequencies of individual race within each surname, using a six-category self-reported race measure: non-Hispanic white, non-Hispanic black, non-Hispanic Asian and Pacific Islander, Hispanic origin, non-Hispanic American Indian and Alaskan Native, and non-Hispanic of two or more races. This census name list is contained in the CSV data file names.csv.\nVariable Description\nsurname: surname\ncount: number of individuals with a specific surname\npctwhite: percentage of non-Hispanic whites among those who have a specific surname\npctblack: percentage of non-Hispanic blacks among those who have a specific surname\npctapi: percentage of non-Hispanic Asians and Pacific Islanders among those who have a specific surname\npcthispanic: percentage of Hispanic origin among those who have a specific surname\npctothers: percentage of the other racial groups among those who have a specific surname",
    "question": "What proportion of voters whose races are black are correctly classified in their racial category? Race is considered correctly classified if the racial category with the greatest conditional probability is identical to the self-reported race. Please remove those voters who contain at least a missing value. Please round to the nearest hundredth.",
    "answer": "0.16",
    "data": [
      "/data/qrdata/data/FLVoters.csv",
      "/data/qrdata/data/names.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 6.2.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Here is a random sample of 10,000 registered voters from Florida contained in the CSV file FLVoters.csv.\nVariable Description\nsurname: surname\ncounty: county ID of the voter’s residence\nVTD: voting district ID of the voter’s residence\nage: age\ngender: gender m = male and f = female\nrace: self-reported race\n\nThe US Census Bureau releases a list of common surnames with their frequency. For example, the most common surname was “Smith” with 2,376,206 occurrences, followed by “Johnson” and “Williams” with 1,857,160 and 1,534,042, respectively. This data set is quite comprehensive, including a total of more than 150,000 surnames that occurred at least 100 times. In addition, the census provides the relative frequencies of individual race within each surname, using a six-category self-reported race measure: non-Hispanic white, non-Hispanic black, non-Hispanic Asian and Pacific Islander, Hispanic origin, non-Hispanic American Indian and Alaskan Native, and non-Hispanic of two or more races. This census name list is contained in the CSV data file names.csv.\nVariable Description\nsurname: surname\ncount: number of individuals with a specific surname\npctwhite: percentage of non-Hispanic whites among those who have a specific surname\npctblack: percentage of non-Hispanic blacks among those who have a specific surname\npctapi: percentage of non-Hispanic Asians and Pacific Islanders among those who have a specific surname\npcthispanic: percentage of Hispanic origin among those who have a specific surname\npctothers: percentage of the other racial groups among those who have a specific surname",
    "question": "What proportion of voters whose races are non-Hispanic Asians and Pacific Islanders are correctly classified in their racial category? Race is considered correctly classified if the racial category with the greatest conditional probability is identical to the self-reported race. Please remove those voters who contain at least a missing value. Please round to the nearest hundredth.",
    "answer": "0.56",
    "data": [
      "/data/qrdata/data/FLVoters.csv",
      "/data/qrdata/data/names.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 6.2.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Here is a random sample of 10,000 registered voters from Florida contained in the CSV file FLVoters.csv.\nVariable Description\nsurname: surname\ncounty: county ID of the voter’s residence\nVTD: voting district ID of the voter’s residence\nage: age\ngender: gender m = male and f = female\nrace: self-reported race\n\nThe US Census Bureau releases a list of common surnames with their frequency. For example, the most common surname was “Smith” with 2,376,206 occurrences, followed by “Johnson” and “Williams” with 1,857,160 and 1,534,042, respectively. This data set is quite comprehensive, including a total of more than 150,000 surnames that occurred at least 100 times. In addition, the census provides the relative frequencies of individual race within each surname, using a six-category self-reported race measure: non-Hispanic white, non-Hispanic black, non-Hispanic Asian and Pacific Islander, Hispanic origin, non-Hispanic American Indian and Alaskan Native, and non-Hispanic of two or more races. This census name list is contained in the CSV data file names.csv.\nVariable Description\nsurname: surname\ncount: number of individuals with a specific surname\npctwhite: percentage of non-Hispanic whites among those who have a specific surname\npctblack: percentage of non-Hispanic blacks among those who have a specific surname\npctapi: percentage of non-Hispanic Asians and Pacific Islanders among those who have a specific surname\npcthispanic: percentage of Hispanic origin among those who have a specific surname\npctothers: percentage of the other racial groups among those who have a specific surname",
    "question": "What is the false discovery rate of the white race? A voter is classified into the racial category with the greatest conditional probability P (race | surname). The false discovery rate represents the proportion of voters who are not white among those classified as white. Please remove those voters who contain at least a missing value. Please round to the nearest hundredth.",
    "answer": "0.20",
    "data": [
      "/data/qrdata/data/FLVoters.csv",
      "/data/qrdata/data/names.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 6.2.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Here is a random sample of 10,000 registered voters from Florida contained in the CSV file FLVoters.csv.\nVariable Description\nsurname: surname\ncounty: county ID of the voter’s residence\nVTD: voting district ID of the voter’s residence\nage: age\ngender: gender m = male and f = female\nrace: self-reported race\n\nThe US Census Bureau releases a list of common surnames with their frequency. For example, the most common surname was “Smith” with 2,376,206 occurrences, followed by “Johnson” and “Williams” with 1,857,160 and 1,534,042, respectively. This data set is quite comprehensive, including a total of more than 150,000 surnames that occurred at least 100 times. In addition, the census provides the relative frequencies of individual race within each surname, using a six-category self-reported race measure: non-Hispanic white, non-Hispanic black, non-Hispanic Asian and Pacific Islander, Hispanic origin, non-Hispanic American Indian and Alaskan Native, and non-Hispanic of two or more races. This census name list is contained in the CSV data file names.csv.\nVariable Description\nsurname: surname\ncount: number of individuals with a specific surname\npctwhite: percentage of non-Hispanic whites among those who have a specific surname\npctblack: percentage of non-Hispanic blacks among those who have a specific surname\npctapi: percentage of non-Hispanic Asians and Pacific Islanders among those who have a specific surname\npcthispanic: percentage of Hispanic origin among those who have a specific surname\npctothers: percentage of the other racial groups among those who have a specific surname\n\nIn the United States, voter files contain voters’ addresses. Using this information, our data set also provides the voting district where each voter lives. In addition, we will utilize the Florida census data, which contains the racial composition of each voting district. The census data set is in FLCensusVTD.csv.\nVariable Description\ncounty: county census ID of the voting district\nVTD: voting district census ID (only unique within the county)\ntotal.pop: total population of the voting district\nwhite: proportion of non-Hispanic whites in the voting district\nblack: proportion of non-Hispanic blacks in the voting district\napi: proportion of non-Hispanic Asians and Pacific Islanders in the voting district\nhispanic: proportion of voters of Hispanic origin in the voting district\nothers: proportion of the other racial groups in the voting district",
    "question": "What proportion of voters whose races are black are correctly classified in their racial category? Race is considered correctly classified if the racial category with the greatest conditional probability P(race | surname, residence) is identical to the self-reported race. Please remove those voters who contain at least a missing value. Please round to the nearest hundredth.",
    "answer": "0.63",
    "data": [
      "/data/qrdata/data/FLVoters.csv",
      "/data/qrdata/data/names.csv",
      "/data/qrdata/data/FLCensusVTD.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 6.2.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability",
        "Bayesian rule"
      ],
      "question_type": "numerical"
    }
  },
  {
    "context": "Here is a random sample of 10,000 registered voters from Florida contained in the CSV file FLVoters.csv.\nVariable Description\nsurname: surname\ncounty: county ID of the voter’s residence\nVTD: voting district ID of the voter’s residence\nage: age\ngender: gender m = male and f = female\nrace: self-reported race\n\nThe US Census Bureau releases a list of common surnames with their frequency. For example, the most common surname was “Smith” with 2,376,206 occurrences, followed by “Johnson” and “Williams” with 1,857,160 and 1,534,042, respectively. This data set is quite comprehensive, including a total of more than 150,000 surnames that occurred at least 100 times. In addition, the census provides the relative frequencies of individual race within each surname, using a six-category self-reported race measure: non-Hispanic white, non-Hispanic black, non-Hispanic Asian and Pacific Islander, Hispanic origin, non-Hispanic American Indian and Alaskan Native, and non-Hispanic of two or more races. This census name list is contained in the CSV data file names.csv.\nVariable Description\nsurname: surname\ncount: number of individuals with a specific surname\npctwhite: percentage of non-Hispanic whites among those who have a specific surname\npctblack: percentage of non-Hispanic blacks among those who have a specific surname\npctapi: percentage of non-Hispanic Asians and Pacific Islanders among those who have a specific surname\npcthispanic: percentage of Hispanic origin among those who have a specific surname\npctothers: percentage of the other racial groups among those who have a specific surname\n\nIn the United States, voter files contain voters’ addresses. Using this information, our data set also provides the voting district where each voter lives. In addition, we will utilize the Florida census data, which contains the racial composition of each voting district. The census data set is in FLCensusVTD.csv.\nVariable Description\ncounty: county census ID of the voting district\nVTD: voting district census ID (only unique within the county)\ntotal.pop: total population of the voting district\nwhite: proportion of non-Hispanic whites in the voting district\nblack: proportion of non-Hispanic blacks in the voting district\napi: proportion of non-Hispanic Asians and Pacific Islanders in the voting district\nhispanic: proportion of voters of Hispanic origin in the voting district\nothers: proportion of the other racial groups in the voting district",
    "question": "What proportion of voters whose races are non-Hispanic Asians and Pacific Islanders are correctly classified in their racial category? Race is considered correctly classified if the racial category with the greatest conditional probability P(race | surname, residence) is identical to the self-reported race. Please remove those voters who contain at least a missing value. Please round to the nearest hundredth.",
    "answer": "0.61",
    "data": [
      "/data/qrdata/data/FLVoters.csv",
      "/data/qrdata/data/names.csv",
      "/data/qrdata/data/FLCensusVTD.csv"
    ],
    "metadata": {
      "reference": "Quantitative Social Science 6.2.4",
      "keywords": [
        "Social science",
        "Statistics",
        "Probability",
        "Conditional probability",
        "Bayesian rule"
      ],
      "question_type": "numerical"
    }
  }
]