Data quality, scaling assumptions and reliability of crowd behaviour scale

Hajj is an annual religious activity that involves a large gathering of people performing the same rituals at the same time within a confined space. Human behaviour in this mass gathering is an important issue of serious concern with life-threatening consequences due to extreme congestion.A new scale was developed as part of a study aims to understand observable crowd behaviour during this religious activity. Subscales of observable crowd behaviour included aggressive behaviour, coping behaviour, defensive behaviour, avoidance behaviour, protective behaviour, tolerant behaviour and hazardous behaviour.This article reports on the underlying assumptions of data quality and scaling assumptions under the traditional Likert scaling criteria, and the reliability of the scale.The subjects of the survey included 203 respondents during pre-Hajj training program at three different locations in the country.Methods of testing include item-level descriptive statistics, multitrait scaling/multi-item correlation matrix and scale reliability.The results showed that one item each within subscales of aggressive, coping and tolerant behaviour do not meet the Likert scaling criteria. However, all subscales have acceptable level of reliability.The findings from this study can be used to further improve the scale.