Read on Twitter
As promised. Here are some good references if you want to demonstrate the socio-economic biases that operate in recruitment.
A thread:
A thread:
1st up. Very nice paper using a birth cohort (ALSPAC) to test for systematic biases when you try and recruit cohort members subsequently. The population representative nature of the baseline cohort provides a perfect test bed for these kinds of biases. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2802508/
2nd up. Incredibly useful paper from the Journal of Survey Stats and Methods. Authors make a direct comparison of opt-in vs opt-out recruitment. It is about data linkage - the direct comparison of opt-in/out is v rare, but v useful. https://academic.oup.com/jssam/article/4/3/382/2399768
3rd up. No empirical data on biases, but a nice overview of why it matters that we have more representative samples. Theories, interventions and policy decisions are all based on generalisations... so who we sample matters. https://www.nature.com/articles/466029a
4th up. UK Biobank. When recruiting 500k participants, researchers tested for systematic differences between responders and non-responders. Guess what... those who opted-in were significantly more likely to live in affluent areas. https://academic.oup.com/aje/article/186/9/1026/3883629
Do add in more if you know of them!
