A quantification of the reliability of self-reports following a simulated stressful event

Abstract

Interviews and surveys are the most commonly used data-gathering and data-generating techniques when investigating human behaviour in emergencies. However, these approaches suffer from several limitations, including potential errors in memory accuracy, a lack of quantitative reliability. This study focuses on a survey performed on participants who had taken part in a stressful experiment. The survey was carried out three months afterwards, asking them to recall their experience. Analysis of this data quantitatively assesses their recall, across multiple different domains. This study observed several differences between experimental and control group participants, as well as differences between participants in VR and Physical experimental groups. However, it observes no increase in confabulation as a result of increased stress. The outcome of this study is to provide insight into the quantitative reliability of interviews and surveys of people involved in emergencies.