Measuring the Success of Context-Aware Security Behaviour Surveys

University College London researchers developed a methodology for scenario-based security behaviour surveys deployed in two large partner organizations. The dataset includes 516 and 195 free-text responses coded by annotators, plus 5196 and 1824 appropriateness and severity scores. The work aims to validate participant clustering and improve survey design by grounding questions in specific organizational tensions.

Use Cases

• Validating participant clustering based on free-text responses.
• Assessing the appropriateness and severity of security survey scenarios for specific organizations.
• Improving survey design methodology by comparing inter-annotator agreement metrics.
• Analyzing the relationship between scenario recognizability and survey quality.

Strengths

• Includes 516 and 195 free-text responses from two organizations, providing qualitative depth.
• Analyzes 5196 and 1824 appropriateness and severity scores for methodological validation.
• Reports strong inter-annotator agreement with Krippendorff's α > 0.7.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the primary survey data is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the two specific partner organizations studied.

Provenance

Source

University College London

Collection Method

Survey methodology developed and deployed in two large partner organizations, with free-text responses coded by annotators.