Statistical analyses of a questionnaire about diversity and inclusion were conducted using three stages of statistical methods. First, we validated the questionnaire with both a PCA analysis and a Cronbach α-test to check the internal consistency of the responses. Second, we used a k-means clustering algorithm on the (varimax-rotated) component scores from the PCA analysis. In the third stage, we derived five different linear regressions – four based on different groups of questions, and one based on of the PCA-components.
A significance level of 5% was used in testing the significance of groups (e.g. Male and Female) under each cluster, and the significance of effects in the regression analyses.