If I want to recruit a representative sample that is also balanced for another variable like highest education level, is that possible? If not, can a subsequent study be created to recruit additional participants in some categories, while excluding participants who were already included in the representative sample? Also, is there generally a balance of participants across education levels in Prolific?
Hi @Dani_Levine ,
- you could try to do something as suggested on this thread (Representative, big sample - #11 by paul) and create many studies with custom screeners corresponding to the representative sample. (You would need to look up which Prolific screeners correspond to the SES demographic of interest and the proportions of categories in the population of interest.
- No idea but you can mess around with the screeners and see how it affects the number of Prolific users that would be eligible
Although a Prolific “representative sample” doesn’t officially account for SES in its sampling, I ran a study last year (n =1000) which came out with a fairly similar income distribution to that which you’d expect given the national picture in the UK. The attached image shows the distribution in annual estimated household income (Prolific sample, 2020) and the UK picture for equivalised disposable household income (ONS, 2019). Obviously, the income measures are not exactly the same, but the shape of the distribution is broadly similar.
My point here is that, even if you don’t find you’re able to design filters that will give you a representative sample by SES, a large enough sample may well give you what you’re looking for, and you can always look at national data to see how well your distributions compare.
That’s super interesting @Gillian_Pepper - thanks for sharing! You’ve got the team thinking