Update README.md

README.md

@@ -19,7 +19,7 @@ tags:
 
 <!-- Provide a quick summary of the dataset. -->
 
-ConstructCIE is a dataset for extracting causal information from construction accident narratives. Each accident report is annotated with a hierarchy of causal factors (managerial, behavioral, equipment, and working-condition factors), circumstances, and consequences, supporting both extraction and classification tasks.
+ConstructCIE is a dataset for extracting causal information from construction accident narratives. Each accident report is annotated with a hierarchy of causal factors.
 
 ## Dataset Details
 
@@ -27,7 +27,7 @@ ConstructCIE is a dataset for extracting causal information from construction ac
 
 <!-- Provide a longer summary of what this dataset is. -->
 
-The dataset contains 530 English construction accident narratives. Each narrative is annotated with a tree of causal information: **extraction** nodes carry supporting text spans and keywords, and **classification** nodes carry categorical labels (accident type, construction trade, severity). It supports two settings: End-to-End Joint Hierarchical Extraction (**JHE**) and End-to-End Individual Hierarchical Extraction (**IHE**).
+The dataset contains 530 English construction accident narratives drawn from OSHA accident investigation summaries published between 2011 and 2023. Each narrative is annotated with a tree of causal information: **extraction** nodes carry supporting text spans and keywords, and **classification** nodes carry categorical labels (accident type, construction trade, severity).
 
 - **Language:** English
 - **License:** Apache 2.0
@@ -66,17 +66,29 @@ Each record is one JSON line with three top-level fields:
 - `accident_report` — the narrative text with its hierarchical annotation tree
 - `accident_type` — classification: `caught-in/between`, `electrocution`, `fall`, or `struck-by`
 
-Nodes in the `accident_report` tree are either **extraction** nodes (`{text, keywords, children}`) or **classification** nodes (`{value}`), organized as:
+Nodes in the `accident_report` tree are either **extraction** nodes (`{text, keywords, children}`) or **classification** nodes (`{value}`). Factors and their subfactors (marked ↳) are:
+
+| Factor / Subfactor | Definition |
+|--------------------|------------|
+| `working_circumstances` | Physical and operational contexts present at the time of the accident that characterize the immediate work situation. |
+| ↳ `construction_trade` | The primary construction activity division being performed at the time of the accident. |
+| `managerial_factors` | Deficiencies at the organizational or supervisory level that allowed unsafe conditions or behaviors to exist. |
+| ↳ `failure_of_hazard_management` | Failure to identify or include foreseeable hazards during pre-task planning or risk assessment. |
+| ↳ `deficiency_in_safety_training` | Failure to provide adequate task-specific training enabling workers to perform tasks safely and respond to foreseeable hazards. |
+| `working_condition_factors` | Physical conditions of the work environment, such as natural conditions and workspace characteristics, that contributed to the occurrence of the accident. |
+| ↳ `weather_condition` | Weather-related conditions that influenced the work environment or contributed to the accident sequence. |
+| ↳ `workspace_condition` | Physical conditions related to the workspace, or structural elements, including spatial arrangement, support conditions, integrity, surface characteristics, or surrounding physical layout. |
+| `equipment_factors` | Conditions related to personal or collective protective equipment or work equipment whose state contributed to the accident sequence. |
+| ↳ `protective_equipment_condition` | The state, functionality, availability, configuration, or appropriateness of personal or collective protective equipment. |
+| ↳ `work_equipment_condition` | The functional state, physical integrity, configuration, or availability of tools, machinery, or powered/non-powered work equipment. |
+| `behavioral_factors` | Task-level human factors involving cognitive lapses or procedural deviations that directly contributed to the occurrence of the accident. |
+| ↳ `inattentive_behavior` | Behavior resulting from cognitive lapses such as reduced vigilance, distraction, or loss of situational awareness. |
+| ↳ `noncompliant_behavior` | Behavior involving the disregard of required safety rules, safe work practices, or protective measures. |
+| `consequences` | The adverse outcomes of the incident, including the seriousness of injury and the part(s) of the body affected. |
+| ↳ `severity` | The degree and seriousness of injury: `fatality`, `hospitalized injury`, or `non hospitalized injury`. |
+| ↳ `affected_body_part` | The part(s) of the body that sustained harm due to the incident. |
+| `object_involved` | The equipment, structure, material, or object that physically interacted with or contributed to the accident mechanism. |
 
-- `working_circumstances` → `construction_trade` (17 CSI-division classes)
-- `managerial_factors` → `failure_of_hazard_management`, `deficiency_in_safety_training`
-- `working_condition_factors` → `weather_condition`, `workspace_condition`
-- `equipment_factors` → `protective_equipment_condition`, `work_equipment_condition`
-- `behavioral_factors` → `inattentive_behavior`, `noncompliant_behavior`
-- `consequences` → `severity` (`fatality`, `hospitalized injury`, `non hospitalized injury`), `affected_body_part`
-- `object_involved`
-
-The dataset ships with five alternate train/val/test partitions (`split1`–`split5`) of the same 530 records, 424/53/53 each, for multi-seed evaluation.
 
 ## Dataset Creation
 
@@ -84,7 +96,7 @@ The dataset ships with five alternate train/val/test partitions (`split1`–`spl
 
 <!-- Motivation for the creation of this dataset. -->
 
-Construction accident reports contain rich causal information locked in free text. Structuring it as a hierarchy of causal factors enables systematic safety analysis and provides a benchmark for hierarchical information extraction methods.
+Construction accident reports contain rich causal information locked in free text. Unlike trigger-centered event extraction, accident causality is often implicit, long-span, and distributed across multiple sentences. Structuring it as a hierarchy of causal factors enables systematic safety analysis and provides a benchmark for hierarchical causal information extraction.
 
 ### Source Data
 
@@ -94,29 +106,13 @@ Construction accident reports contain rich causal information locked in free tex
 
 <!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
 
-[More Information Needed]
+OSHA accident investigation summaries published between 2011 and 2023 were screened in two stages: record-level screening removed duplicates and reports unrelated to construction activities, and content-level screening retained only narratives that explicitly describe both accident circumstances and causes, yielding 530 reports. All date information was removed using regular-expression matching to prevent models from learning false temporal patterns.
 
 #### Who are the source data producers?
 
 <!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->
 
-[More Information Needed]
-
-### Annotations [optional]
-
-<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
-
-#### Annotation process
-
-<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->
-
-[More Information Needed]
-
-#### Who are the annotators?
-
-<!-- This section describes the people or systems who created the annotations. -->
-
-[More Information Needed]
+The Occupational Safety and Health Administration (OSHA), which publishes accident investigation summaries of workplace incidents in the United States.
 
 #### Personal and Sensitive Information
 
@@ -124,18 +120,6 @@ Construction accident reports contain rich causal information locked in free tex
 
 Narratives describe real workplace injuries and fatalities. Workers are referred to anonymously (e.g., "Employee #1") and no names or direct identifiers appear in the text.
 
-## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
-
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-[More Information Needed]
-
 ## Citation
 
 <!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->