Coded HFACS dataset: Runway excursion investigations in Indonesia, 2007–2024
Description
Dataset supporting the study "Human Factors in Runway Excursions: An HFACS Analysis Across Aviation Operation Types in Indonesia." Contains the population of 52 KNKT runway excursion investigation reports, independent coding by two raters, reconciled coding, inter-coder reliability calculations (Krippendorff's alpha), HFACS frequency analysis, Fisher's exact test results, and operational-category stratification.
Files
Steps to reproduce
Runway excursion investigation reports (N=52 of 53 identified cases, 2007–2024) were retrieved from KNKT's public database (https://knkt.go.id). Two independent coders classified contributing factors in each report into the 18 HFACS causal categories using a structured coding protocol (see "Coding_Coder1" and "Coding_Coder2" sheets). Codes were then reconciled through discussion to produce a final consensus dataset ("Reconciliation_Final" sheet). Inter-coder reliability was computed using Krippendorff's alpha, calculated separately for each of the four HFACS levels on the subset of cases both coders judged codeable (N=42; see "Reliability_Alpha" sheet). Frequency distributions of HFACS categories were tabulated from the reconciled data ("HFACS_Frequency_Analysis" sheet). Association between HFACS categories and occurrence severity (Accident vs. non-Accident) was tested using Fisher's exact test ("Fisher_Exact_Analysis" sheet). Cases were further stratified by operational category (scheduled commercial jet, STOL feeder/utility, regional turboprop, flight training) to compare failure patterns across operation types ("Operational_Stratification" sheet). All statistical analyses were performed in a standard spreadsheet environment; formulas underlying each derived statistic are retained in the corresponding worksheet cells.