Differential attainment in specialty training recruitment in the UK: an observational analysis of the impact of psychometric testing assessment in Public Health postgraduate selection

Objectives To determine how current psychometric testing approaches used in selection of postgraduate training in UK Public Health are associated with socioeconomic and sociocultural background of applicants (including ethnicity).Design Observational study using contemporaneous data collected during recruitment and psychometric test scores.Setting Assessment centre of UK national Public Health recruitment for postgraduate Public Health training. The assessment centre element of selection comprises three psychometric assessments: Rust Advanced Numerical Reasoning, Watson-Glaser Critical Thinking Assessment II and Public Health situational judgement test.Participants 629 applicants completed the assessment centre in 2021. 219 (34.8%) were UK medical graduates, 73 (116%) were international medical graduates and 337 (53.6%) were from backgrounds other than medicine.Main outcome measure Multivariable-adjusted progression statistics in the form of adjusted OR (aOR), accounting for age, sex, ethnicity, professional background and surrogate measures of familial socioeconomic and sociocultural status.Results 357 (56.8%) candidates passed all three psychometric tests. Candidate characteristics negatively associated with progression were black ethnicity (aOR 0.19, 0.08 to 0.44), Asian ethnicity (aOR 0.35, 0.16 to 0.71) and coming from a non-UK medical graduate background (aOR 0.05, 0.03 to 0.12); similar differential attainment was observed in each of the psychometric tests. Even within the UK-trained medical cohort, candidates from white British backgrounds were more likely to progress than those from ethnic minorities (89.2% vs 75.0%, p=0.003).Conclusion Although perceived to mitigate the risks of conscious and unconscious bias in selection to medical postgraduate training, these psychometric tests demonstrate unexplained variation that suggests differential attainment. Other specialties should enhance their data collection to evaluate the impact of differential attainment on current selection processes and take forward opportunities to mitigate differential attainment where possible.