Thomas Behaviour assessment and bias
In order to determine whether the Thomas Behaviour assessment was free from bias, a study was carried out with a UK-based recruitment company. In this study, 265 applicants that had applied for roles over a 12-month period, completed the Thomas Behaviour assessment as part of the recruitment process together with a demographic questionnaire.
The behaviour assessment factor scores were found to be statistically significantly different (p=0.2) between women and men or between BAME (Black, Asian, and Minority Ethnic) people and white people. This supports that the Thomas Behaviour assessment is free from bias. By not finding any biases in a specific sample from one controlled occupational context, this suggests that gender differences found in the study above are likely to be artefacts of differing job demands of roles that are more frequently taken up by one gender than another.
Gender, Age, Ethnicity and Education
The following group of studies were carried out in October 2020 and was used to test for differences between groups where none was expected. These groups include demographic characteristics of gender, age, ethnicity and education. This analysis has been used to check that the behaviour scales are not biased towards or against members of certain groups.
Gender - T tests were used to assess differences in behaviour factor scores between males and females. Data was extracted from 61,076 individuals. The analysis showed negligible effect sizes between male and female factor scores for both Influence and Compliance. Small effect sizes were discovered for both Dominance and Steadiness. Females were more likely to have lower Dominance scores and higher Steadiness scores compared to Males.
Age - Pearson’s Product Moment Correlational analysis was used to test for a relationship between respondent age and behaviour factor scores. Data was extracted from 61,076 individuals. The analysis showed negligible correlations across all four factors. This demonstrates that respondent age does not impact the score on each behaviour factor scale.
Ethnicity - A one-way ANOVA was used to test for differences in behaviour factor scores between different ethnic groups. Data was extracted from 59,702 individuals who had both completed the Thomas behaviour assessment and provided their ethnic information. Analysis revealed significant relationships between Ethnicity and the four behaviour factor scales, however, all effect sizes were negligible according to Cohen’s conventions (Cohen, 1988). This demonstrates that respondent ethnicity does not impact the resulting score on each behaviour factor scale.
Education - T tests were used to look for differences in behaviour factor scores between those with different educational backgrounds. Data was extracted from 61,076 individuals who had completed the Thomas behaviour assessment for selection and development purposes. The study specifically looked at those with a degree and those without. Effect sizes demonstrate that respondent education level, namely those who have a degree compared to those without, does not impact the result score on each behaviour factor scale.
We trust you found this article valuable. Please get in touch if there are specific topics you would like to read more about.