Publication type
Understanding Society Working Paper Series
Series Number
2025-03
Series
Understanding Society Working Paper Series
Authors
Publication date
May 20, 2025
Abstract:
Survey non-response biases cause invalid inference. Minimising biases requires information on their causes, but those in ethnic minority and long term respondent subgroups are comparatively unstudied. We investigate bias risks and their causes (estimating biases is difficult as population values do not exist) in these subgroups in Understanding Society – the UK Household Longitudinal Study (UKHLS). We find that both exhibit bias risks, and that, unlike in overall survey datasets, several ethnicities are under-represented in minority samples, and those with below degree levels of education in long term respondent samples. We then discuss the implications of our findings for survey design.
Paper download#588626