[ { "id": 1, "name": "IUIPC_mean_absolute_CohensD", "source_paper": "Internet Users' Information Privacy Concerns (IUIPC): The Construct, the Scale, and a Causal Model", "description": "After calculating IUIPC subscales (collection, control, awareness) and trust, risk, intention scores from simulated IUIPC survey, compare the simulated scores’ distributions to published distributions.", "scenario": "survey_IUIPC.jsonl", "action_space": "survey_IUIPC.jsonl", "quantification": "Compare the simulated scores’ distributions with published distributions by calculating each score’s absolute Cohen’s d, and then take the average of Cohen’s d of collection, control, awareness, trust, risk, and intention. Cohen’s d = |M2 - M1| / SDpooled where M1 is mean of group1, M2 is mean of group2, and SDpooled = sqrt(((n1-1)*SD1 + (n2-1)*SD2)) / (n1 + n2 - 2)) where n1 is sample size of group1, SD1 is standard deviation of group1, n2 is sample size of group2, and SD2 is standard deviation of group2.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 2, "name": "IUIPC_model_coeff_MAE", "source_paper": "Internet Users' Information Privacy Concerns (IUIPC): The Construct, the Scale, and a Causal Model", "description": "After calculating IUIPC subscales (collection, control, awareness) and trust, risk, intention scores from simulated IUIPC survey, replicate the published structural model by running SEM. Then, compare the simulated correlation coefficients to the published coefficients.", "scenario": "survey_IUIPC.jsonl", "action_space": "survey_IUIPC.jsonl", "quantification": "Compare the simulated correlation coefficients to the published coefficients by calculating the absolute difference in each correlation, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 3, "name": "SeBIS_mean_absolute_CohensD", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey, compare the simulated items’ distributions to published distributions.", "scenario": "survey_SeBIS.jsonl", "action_space": "survey_SeBIS.jsonl", "quantification": "Compare the simulated item score distributions with published distributions by calculating each item’s absolute Cohen’s d, and then take the average of Cohen’s d of 16 items.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 41, "name": "SeBIS_MAE_DoSpeRT", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey and Domain-Specific Risk-Taking Scale (DoSpeRT).", "scenario": "surveys_SeBIS.jsonl", "action_space": "surveys_SeBIS.jsonl", "quantification": "Performing Pearson correlations between SeBIS subscales and psychometrics, compare the coefficients to the published coefficients by calculating absolute difference, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 42, "name": "SeBIS_MAE_GDMS", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey and General Decision-Making Style (GDMS).", "scenario": "surveys_SeBIS.jsonl", "action_space": "surveys_SeBIS.jsonl", "quantification": "Performing Pearson correlations between SeBIS subscales and psychometrics, compare the coefficients to the published coefficients by calculating absolute difference, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 43, "name": "SeBIS_MAE_NFC", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey and Need for Cognition (NFC).", "scenario": "surveys_SeBIS.jsonl", "action_space": "surveys_SeBIS.jsonl", "quantification": "Performing Pearson correlations between SeBIS subscales and psychometrics, compare the coefficients to the published coefficients by calculating absolute difference, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 44, "name": "SeBIS_MAE_BIS", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey and Barratt Impulsiveness Scale (BIS).", "scenario": "surveys_SeBIS.jsonl", "action_space": "surveys_SeBIS.jsonl", "quantification": "Performing Pearson correlations between SeBIS subscales and psychometrics, compare the coefficients to the published coefficients by calculating absolute difference, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 45, "name": "SeBIS_MAE_CFC", "source_paper": "Scaling the Security Wall Developing a Security Behavior Intentions Scale (SeBIS)", "description": "After simulating SeBIS’s 16-item survey and Consideration for Future Consequences (CFC).", "scenario": "surveys_SeBIS.jsonl", "action_space": "surveys_SeBIS.jsonl", "quantification": "Performing Pearson correlations between SeBIS subscales and psychometrics, compare the coefficients to the published coefficients by calculating absolute difference, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 5, "name": "SA6_distribution_abs_CohensD", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Compare the simulated SA-6 overall score distribution to the paper’s overall distribution.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Cohen’s d between the simulation’s SA-6 mean and the paper’s mean using the pooled SD, then take the absolute value: |d| = |(M_sim − M_paper) / SD_pooled|.", "expectation": "Closer to 0 indicates a better match to the paper’s SA-6 distribution.", "target": 0, "direction": "~0", "compromise": "" }, { "id": 6, "name": "SA6_SeBIS_spearman_rho", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Quantify convergent validity by correlating SA-6 with SeBIS.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Spearman’s ρ between SA-6 score (mean of 6 items) and SeBIS score (mean of 16 items after specified reverse-coding). Report the correlation value.", "expectation": "Closer to 0.540 (paper) indicates better alignment.", "target": 0.540, "direction": "~0.540", "compromise": "" }, { "id": 7, "name": "SA6_BIS_spearman_rho", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Association of SA-6 with Barratt Impulsiveness (perceived behavioral control indicator).", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Spearman’s ρ between SA-6 score and BIS score (mean across BIS items after instrument-appropriate reverse-coding, if any). Report the correlation value.", "expectation": "Closer to −0.180 (paper) indicates better alignment; sign should be negative.", "target": -0.180, "direction": "~-0.180", "compromise": "" }, { "id": 8, "name": "SA6_GSE_spearman_rho", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Association of SA-6 with General Self-Efficacy (perceived behavioral control indicator).", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Spearman’s ρ between SA-6 score and GSE score (mean of 10 items on a 1–4 scale). Report the correlation value.", "expectation": "Closer to 0.208 (paper) indicates better alignment.", "target": 0.208, "direction": "~0.208", "compromise": "" }, { "id": 9, "name": "SA6_SSE_spearman_rho", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Association of SA-6 with Social Self-Efficacy (perceived behavioral control indicator).", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Spearman’s ρ between SA-6 score and SSE score (mean of 6 items after reverse-coding items 1, 3, 5 on a 1–5 scale). Report the correlation value.", "expectation": "Closer to 0.363 (paper) indicates better alignment.", "target": 0.363, "direction": "~0.363", "compromise": "" }, { "id": 10, "name": "SA6_t_age_18_39_minus_40_plus", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Age difference in SA-6 scores using the paper’s contrast direction.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Welch’s t for SA-6 with contrast A–B defined as (18–39) − (40+). Report the t statistic.", "expectation": "Closer to −2.172 (paper) indicates better alignment; negative sign means younger < older.", "target": -2.172, "direction": "~-2.172", "compromise": "" }, { "id": 11, "name": "SA6_t_gender_male_minus_female", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Gender difference in SA-6 scores using the paper’s contrast direction.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Welch’s t for SA-6 with contrast A–B defined as Male − Female. Report the t statistic.", "expectation": "Closer to +2.19 (paper) indicates better alignment; positive sign means male > female.", "target": 2.19, "direction": "~2.19", "compromise": "" }, { "id": 12, "name": "SA6_t_edu_no_college_minus_attend_college", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Education difference in SA-6 scores using the paper’s contrast direction.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Welch’s t for SA-6 with contrast A–B defined as No college − Attend college. Report the t statistic.", "expectation": "Closer to −2.76 (paper) indicates better alignment; negative sign means no-college < college.", "target": -2.76, "direction": "~-2.76", "compromise": "" }, { "id": 13, "name": "SA6_t_income_below25k_minus_above25k", "source_paper": "Security Attitudes (SA-6), SOUPS 2019 (Faklaris et al.)", "description": "Income difference in SA-6 scores using the paper’s contrast direction.", "scenario": "survey_SA_6.jsonl", "action_space": "survey_SA_6.jsonl", "quantification": "Compute Welch’s t for SA-6 with contrast A–B defined as Below $25K − Above $25K. Report the t statistic.", "expectation": "Closer to −3.42 (paper) indicates better alignment; negative sign means below-income < above-income.", "target": -3.42, "direction": "~-3.42", "compromise": "" }, { "id": 14, "name": "OPC_PUI_mean_absolute_CohensD", "source_paper": "Privacy, Trust, and Self-Disclosure Online", "description": "After calculating scores (Privacy Concern, General Caution, and Technical Protection) from simulated OPC_PUI survey, compare the simulated scores’ distributions to published distributions.", "scenario": "survey_OPC_PUI.jsonl", "action_space": "survey_OPC_PUI.jsonl", "quantification": "Compare the simulated scores’ distributions with published distributions by calculating each score’s absolute Cohen’s d, and then take the average of Cohen’s d of collection, control, awareness, trust, risk, and intention. Cohen’s d = |M2 - M1| / SDpooled where M1 is mean of group1, M2 is mean of group2, and SDpooled = sqrt(((n1-1)*SD1 + (n2-1)*SD2)) / (n1 + n2 - 2)) where n1 is sample size of group1, SD1 is standard deviation of group1, n2 is sample size of group2, and SD2 is standard deviation of group2.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "Respondents’ scores distribution is not mentioned in paper ‘Development of measures of online privacy concern and protection for use on the Internet’, and we choose to simulate a follow-up paper ‘Privacy, Trust, and Self-Disclosure Online’ which replicates the study and contains scores’ distribution." }, { "id": 15, "name": "OPC_PUI_corr_coeff_MAE", "source_paper": "Privacy, Trust, and Self-Disclosure Online", "description": "After calculating scores (Privacy Concern, General Caution, and Technical Protection) from simulated OPC_PUI survey, compare the simulated correlation coefficients of the three pairs of scores to the published coefficients.", "scenario": "survey_OPC_PUI.jsonl", "action_space": "survey_OPC_PUI.jsonl", "quantification": "Compare the simulated correlation coefficients to the published coefficients by calculating the absolute difference in each correlation, and take the average.", "expectation": "Closer to 0 indicates a better performance.", "target": 0, "direction": "~0", "compromise": "We choose to simulate the paper ‘Privacy, Trust, and Self-Disclosure Online’ which replicates the study and contains scores’ distribution." }, { "id": 16, "name": "premium_spending_for_privacy", "source_paper": "The Effect of Online Privacy Information on Purchasing Behavior: An Experimental Study", "description": "Users choose between merchants with or without clear privacy–protection information. Without visible privacy cues, users buy from the lowest-price seller; when privacy protections are made salient, users pay ~5% premium to purchase from more privacy-protecting merchants.", "scenario": { "Without Privacy Cues": "You have entered 'Duracell AA batteries 8-pack' into the Finder search bar and hit Search. The result page is shown in the image. Determine which action you will take.", "With Privacy Cues": "You have entered 'Duracell AA batteries 8-pack' into the Finder search bar and hit Search. The result page is shown in the image. Each listing includes a row of Privacy Report squares next to the title (more green box means higher privacy rating). Determine which action you will take." }, "scenario_img": ["5_1.png", "5_2.png"], "action_space": { "Without Privacy Cues": [ "Buy from the first listing", "Buy from the second listing", "Buy from the third listing", "Buy from the fourth listing", "Spend some time to click through view each merchant's privacy policy" ], "With Privacy Cues": [ "Buy from the first listing", "Buy from the second listing", "Buy from the third listing", "Buy from the fourth listing" ] }, "quantification": "average premium spending for privacy.", "expectation": "Positive indicates expected result, close to 0.59 indicates close to published result.", "target": 0.59, "direction": ">0", "compromise": "Requires visual input." }, { "id": 17, "name": "attendance_rate_without_vs_with_assurance", "source_paper": "Confidentiality assurances in surveys: reassurance or threat?", "description": "Participants receive invitations to a survey with or without assurances of confidentiality. Assurances of confidentiality reduces attendance rate, because it leads to participants expecting more sensitive questions.", "scenario": { "Without assurances": "You receive an invitation to participate in a ‘Citizens’ Survey. It includes a general survey description suggesting a neutral, non threatening survey topic. Decide if you are willing to participate in the survey.", "With assurances": "You receive an invitation to participate in a ‘Citizens’ Survey. It includes a general survey description suggesting a neutral, non threatening survey topic. It also includes an elaborate confidentiality assurance and a one-page description of how the confidentiality of the data would be safeguarded. Decide if you are willing to participate in the survey." }, "action_space": { "Without assurances": [ "Willing to attend", "Unwilling to attend" ], "With assurances": [ "Willing to attend", "Unwilling to attend" ] }, "quantification": "attendance rate without assurance divided by the rate with assurance", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "The source paper contains the confidentiality assurance in the invitation letter but doesn't contain the survey description. The one-page confidentiality assurance is too long to be included in the scenario description." }, { "id": 18, "name": "ratio_keep_switch_anonomous_card", "source_paper": "What Is Privacy Worth?", "description": "Shoppers given either a $10 anonymous gift card or a $12 trackable gift card, then offered the chance to switch. Participants holding the anonymous card tend to keep it, while those holding the trackable card tend not to switch. Those endowed with anonymity value it more than those without it.", "scenario": { "Having Anonymous Card": "After completing a short survey at a mall kiosk, you are handed a $10 VISA gift card labeled 'Anonymous – purchases cannot be linked to your name.' The researcher now offers: 'You can keep this $10 anonymous card or switch to a $12 card that will link purchases to your name.' What do you do?", "Having Trackable Card": "After completing a short survey at a mall kiosk, you are handed a $12 VISA gift card labeled 'Trackable – purchases will be linked to your name.' The researcher now offers: 'You can keep this $12 trackable card or switch to a $10 anonymous card that will not be linked to your name.' What do you do?" }, "action_space": { "Having Anonymous Card": [ "Keep the $10 anonymous card", "Switch to the $12 trackable card" ], "Having Trackable Card": [ "Keep the $12 trackable card", "Switch to the $10 anonymous card" ] }, "quantification": "rate of keeping anonymous card when offered trackable card divided by rate of switching to anonymous card from trackable card.", "expectation": "Larger than 1 indicates expected result, close to 5 indicates close to published result.", "target": 5, "direction": ">1", "compromise": "Only comparing the keep-switch ratio, ignoring the magnitude of each percentage." }, { "id": 19, "name": "disclosure_rate_casual_vs_professional_interface", "source_paper": "Strangers on a Plane: Context-Dependent Willingness to Divulge Sensitive Information", "description": "Participants take a survey containing sensitive information on a casual or professional interface. Individuals disclose more sensitive information on an unprofessional, casual site than on a formal interface even when the formal interface is judged safer.", "scenario": { "Professional Interface": "Based on the web interface in the image, decide how you will respond. Assume you did commit everything asked.", "Casual Interface": "Based on the web interface in the image, decide how you will respond. Assume you did commit everything asked." }, "scenario_img": ["8_1.png", "8_2.png"], "action_space": { "Professional Interface": [ "Answer questions honestly", "Skip the question", "Provide false answers" ], "Casual Interface": [ "Answer questions honestly", "Skip the question", "Provide false answers" ] }, "quantification": "disclosure rate under casual interface divided by disclosure rate under professional interface.", "expectation": "Larger than 1 indicates expected result, close to 1.7 indicates close to published result.", "target": 1.7, "direction": ">1", "compromise": "Requires visual input. Scenario hardcoded such that all personas committed activities in the survey." }, { "id": 20, "name": "disclosure_rate_high_vs_low_norm", "source_paper": "The Impact of Relative Standards on the Propensity to Disclose", "description": "Participants answer sensitive questions with low or high admission norms. Disclosure rates are higher when participants see that a majority of others have admitted the behavior.", "scenario": { "Without Normative Information": "You are completing an online questionnaire titled 'Study of Personal Behaviors.' The page displays one question per screen: 'Have you ever shoplifted an item under $20?' with three radio buttons: 'Yes,' 'No,' and 'Prefer not to answer.' You did shoplift an item. Decide how you will respond.", "With Low-Admission Norm": "You are completing an online questionnaire titled 'Study of Personal Behaviors.' The page displays one question per screen: 'Have you ever shoplifted an item under $20?' with three radio buttons: 'Yes,' 'No,' and 'Prefer not to answer.' A grey banner appears beneath your response reading '25% of participants in an earlier survey admitted they have done this.' You did shoplift an item. Decide how you will respond." }, "action_space": { "Without Normative Information": [ "Select 'Yes'", "Skip the question", "Select 'No'" ], "With Low-Admission Norm": [ "Select 'Yes'", "Skip the question", "Select 'No'" ] }, "quantification": "disclosure rate under high norm exposure divided by disclosure rate under low norm exposure.", "expectation": "Larger than 1 indicates expected result, close to 1.27 indicates close to published result.", "target": 1.27, "direction": ">1", "compromise": "In the source paper, respondents get the previous question's norm as they move on to the next question, and the last question’s disclosure rates are compared. We simulated one question with or without normative information displayed for simplicity." }, { "id": 21, "name": "disclosure_rate_high_vs_missing_norm", "source_paper": "The Impact of Relative Standards on the Propensity to Disclose", "description": "Participants answer sensitive questions with high or without admission norms. Disclosure rates are higher when participants see that a majority of others have admitted the behavior.", "scenario": { "Without Normative Information": "You are completing an online questionnaire titled 'Study of Personal Behaviors.' The page displays one question per screen: 'Have you ever shoplifted an item under $20?' with three radio buttons: 'Yes,' 'No,' and 'Prefer not to answer.' You did shoplift an item. Decide how you will respond.", "With High-Admission Norm": "You are completing an online questionnaire titled 'Study of Personal Behaviors.' The page displays one question per screen: 'Have you ever shoplifted an item under $20?' with three radio buttons: 'Yes,' 'No,' and 'Prefer not to answer.' A grey banner appears beneath your response reading '75% of participants in an earlier survey admitted they have done this.' You did shoplift an item. Decide how you will respond." }, "action_space": { "Without Normative Information": [ "Select 'Yes'", "Skip the question", "Select 'No'" ], "With High-Admission Norm": [ "Select 'Yes'", "Skip the question", "Select 'No'" ] }, "quantification": "disclosure rate under high norm exposure divided by disclosure rate without norm exposure.", "expectation": "Larger than 1 indicates expected result, close to 1.21 indicates close to published result.", "target": 1.21, "direction": ">1", "compromise": "In the source paper, respondents get the previous question's norm as they move on to the next question, and the last question’s disclosure rates are compared. We simulated one question with or without normative information displayed for simplicity." }, { "id": 22, "name": "switch_visibility_with_vs_without_unintended_sharing", "source_paper": "Friends Only: Examining a Privacy-Enhancing Behavior in Facebook", "description": "Users' action on personal profile privacy settings with or without exposure to an unintended audience. Users whose personal profile is exposed to an unintended audience are much more likely to restrict their profile visibility to friends-only.", "scenario": { "Without Unintended Sharing": "You log in to ConnectSpace and open your Privacy Settings. Your profile visibility is currently set to 'Public,' meaning anyone on the platform can see your posts, photos, and personal info. There have been no recent notifications or incidents of unintended sharing. Decide what you want your profile visibility to be.", "With Unintended Sharing": "You log in to ConnectSpace and see a notification: 'Alex viewed your profile.' You didn't intend to disclose your personal profile to casual acquaintances like Alex. Your Privacy Settings show 'Public.' Decide what you want your profile visibility to be." }, "action_space": { "Without Unintended Sharing": [ "Keep profile visibility as Public", "Change profile visibility to Friends only" ], "With Unintended Sharing": [ "Keep profile visibility as Public", "Change profile visibility to Friends only" ] }, "quantification": "rate of changing profile visibility to Friends only with unintended disclosure divided by the rate without unintended disclosure.", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "Subject to hyper-focus problem: agents tend to switch to friends-only regardless of unintended disclosure. Close to 3.31 indicates close to published results." }, { "id": 23, "name": "setting_visit_rate_with_generic_vs_personalized_ads", "source_paper": "Online Display Advertising: Targeting and Obtrusiveness", "description": "Consumers' action on privacy settings when ads remain generic or become overly fine-tuned and reference personal data. Users who receive overly fine-tuned ads tend to visit privacy settings than those who receive generic ads.", "scenario": { "Generic Ads": "You navigate to NewsHub's desktop homepage. At the top is a static banner reading 'Spring Car Sale – Save on New 2025 Sedans,' with a photo of a red sedan and no reference to your browsing history or personal data. On the right sidebar, a second ad invites you to 'Join the Summer Reading List' with a generic stack of books. There are no messages about personalization or targeting. Decide how you respond to these advertisements.", "Personalized Ads": "You navigate to NewsHub's desktop homepage. At the top is a static banner reading 'Your Dream Sedan Awaits!' and shows a photo of the exact red sedan you researched yesterday. It says 'Based on your searches for \u20182025 sedans,' get 15% off at AutoMax.' The sidebar ad reads 'Brooklyn JazzFest Tickets, Alex – Loved that jazz club you visited? 20% off ends soon.' The small text below both ads says 'Ads personalized using your browsing data.' Decide how you respond." }, "action_space": { "Generic Ads": [ "Ignore the ads and continue reading or click on one of the ads", "Open browser's ad settings to review personalization options" ], "Personalized Ads": [ "Ignore the ads and continue reading or click on one of the ads", "Open browser's ad settings to opt out of personalization" ] }, "quantification": "rate of viewing settings seeing personalized ads divided by this rate seeing generic ads.", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "Source paper focuses on click rate of advertisements controlling high/low ads visibility/obtrusiveness and with/without contextual targeting, while we only focus on the latter. The exact personalized ads’ texts are not shown in the paper. We hardcoded in the scenario so that the personalized advertisement refers to personas’ search history." }, { "id": 24, "name": "disclosure_rate_with_vs_without_privacy_policy", "source_paper": "Alan Westin's Privacy Homo Economicus", "description": "On a sign-up page with or without a privacy policy link (which does not guarantee data won't be shared), users choose whether to share accurate personal information. Presence of a visible privacy policy link increases users' likelihood of submitting real personal details compared to when no policy is shown.", "scenario": { "Without Privacy Policy": "You arrive at a website's signup page (shown in image). Decide how you will respond.", "With Privacy Policy": "You arrive at a website's signup page (shown in image). Decide how you will respond." }, "scenario_img": ["13_1.png", "13_2.png"], "action_space": { "Without Privacy Policy": [ "Provide real information, click 'Sign Up'", "Provide fake or partial information, click 'Sign Up'" ], "With Privacy Policy": [ "Provide real information, click 'Sign Up'", "Provide fake or partial information, click 'Sign Up'" ] }, "quantification": "rate of disclosing real information with a policy link divided by disclosure rate without a policy link.", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "Requires visual input. The source paper focuses on people’s misbelief about privacy policy rather than real privacy behaviors. (62% falsely believe “If a website has a privacy policy, it means that the site cannot share information about you with other companies, unless you give the website your permission”). We manually created two sign-up interfaces for simulation." }, { "id": 25, "name": "LBS_usage_with_vs_without_gov_regulation", "source_paper": "The Role of Push-Pull Technology in Privacy Calculus: The Case of Location-Based Services", "description": "Users decide whether to sign up for a push-based, location-based mobile coupon service, with one scenario mentioning a government privacy act and the other not. The mention of government regulation reduces perceived privacy risks, leading to a higher likelihood of users signing up for the service.", "scenario": { "Without Government Regulation": "You are considering signing up for a new mobile coupon service. To join, you register your mobile phone number and select categories of merchants you are interested in (e.g., restaurants, clothing stores). When you are near one of these merchants, the service will automatically send promotional information and coupons to your phone via text message. Your location will be used to target you with relevant advertisements. Would you sign up for this service?", "With Government Regulation": "You are considering signing up for a new mobile coupon service. To join, you register your mobile phone number and select categories of merchants you are interested in (e.g., restaurants, clothing stores). When you are near one of these merchants, the service will automatically send promotional information and coupons to your phone via text message. Your location will be used to target you with relevant advertisements. The service provider notes that your privacy is protected under the 'Privacy and Wireless Communications Protection Act.' Would you sign up for this service?" }, "action_space": { "Without Government Regulation": [ "Sign up for the service", "Decline the service" ], "With Government Regulation": [ "Sign up for the service", "Decline the service" ] }, "quantification": "rate of using LBS with government regulation divided by use rate without government regulation.", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "Source paper includes self-reported survey examining respondents’ knowledge of Privacy and Wireless Communications Protection Act and their perceived risks of location-based services (LBS). Government regulation is found to have a negative effect on perceived privacy risks in push-based LBS. We craft the push-based LBS scenarios with or without a government regulation." }, { "id": 26, "name": "disclosure_rate_high_vs_low_user_control", "source_paper": "Misplaced Confidences: Privacy and the Control Paradox", "description": "A user is asked sensitive questions for a research report. In one condition, they are given explicit control over which of their answers can be published. In the other, they are not. Providing users with explicit control over their data leads them to reveal more sensitive information than they would have otherwise.", "scenario": { "Low Control": "You are participating in an anonymous online survey for a research report on modern behaviors. Your answers will be used to compile the report. Please answer the following question: 'Have you ever fantasized about something you were ashamed of?'", "High Control": "You are participating in an anonymous online survey for a research report on modern behaviors. Your answers will be used to compile the report. Importantly, after you have answered all the questions, you will be able to review your individual answers and decide which, if any, will be included in the report. You have the final say.\n\nPlease answer the following question: 'Have you ever fantasized about something you were ashamed of?'" }, "action_space": { "Low Control": [ "Answer the question honestly and in detail.", "Decline to answer the question." ], "High Control": [ "Answer the question honestly and in detail.", "Decline to answer the question." ] }, "quantification": "rate of disclosing sensitive information with low control divided by disclosure rate with high control.", "expectation": "Larger than 1 indicates expected result.", "target": null, "direction": ">1", "compromise": "In the source paper, participants answer sensitive yes/no questions with or without sharing control over each answer. The exact survey questions are not described in the source paper; only topic examples are mentioned." } ]