How should we expand our participant pool?
Note: We’re not aiming to suggest new screeners for our existing participant pool (if you’d like to suggest a new screener, you can do so here). This is for you to suggest the new types of participants you’d like to see or expansions of demographics we already have.
Do you wish there were more of a particular group of participants on Prolific? Or new demographics we don’t currently offer?
This channel is dedicated to you proposing and voting for new demographics / expansions of existing ones. Below are some examples:
Expanding Existing Demographics:
- More right-wing people
- More vegans
- People from Non-OECD countries
- Small business owners
Simply create a new topic in this channel with your suggestions, and others will be able to vote on the ones they’d like to see! This will help us with our long-term future participant recruitment planning.
To manage expectations, we can’t promise that these demographics will be added to the platform. This is just to help us figure out where demand is highest.
My intuition is that the male:female imbalance in prolific is a manifestation of something non-representative in prolific recruitment/marketing, affecting where the whole sample is drawn from.
You might be able to research that: i.e., study to find what demographics predict not being in the prolific sample relative to a random sample.
The outcome would be good in terms of having a representative sample for researchers to sample from and might also satisfy a big part of the specific question in this topic: i.e., latent unmet-demographics desired by researchers.
It looks like lack of right wing people is mostly explained by the demographics. The participants seem to skew younger, more female and more educated. The “representative sample” (in the UK, at least) isn’t that far off in terms of party or Brexit support. So fixing those things would be help (and is worth doing anyway – the lack of older people is a big problem). I think this is much better than trying to recruit participants based directly on their views, which may not solve the underlying problem but would risk creating new ones.
(I haven’t tried this in the US, so there may be other factors at play there)