Last week we announced our first round of Junior Researcher Grants. In addition to giving budding researchers a boost, we hope these grants will promote Open Science, and the increase in robustness and reproducibility of findings that it brings. This week, we’re tackling a slightly different problem: Generalisability. So, we’re announcing an upcoming feature that will give researchers the power to recruit nationally representative samples with the flick of a switch.
Have you ever wondered whether your research findings would be the same if you could survey everyone in the country?
In most types of social science and health research, scientists collect and analyse data from a subset of the population, known as a sample. This group of participants is used as a proxy for the whole population and we assume, with certain caveats, that our conclusions drawn from this group will ‘generalise’ to the wider population. We run into problems, however, when the sample is not a good model of the population. We know, for example, that the US population is roughly 49% male and 51% female, and we know that 28% of the US adult population is over the age of 60. So is it reasonable to expect that findings based on an 80% young, female, undergraduate sample will be reflected at a population level? Perhaps not. What’s needed is a sample that’s demographically similar to the population…
Using Prolific, researchers have easy access to a wide range of participants, but collecting demographically diverse samples can still be fiddly. Over the last few months we’ve been hard at work to make it simple to collect nationally representative samples on our platform, and today, we’re announcing that this feature is officially “coming soon”.
If you want to hear more about the representative samples feature, then click here, leave us your details, and we’ll keep you in the loop. If you just can’t wait, and want to run a representative sample on Prolific as soon as possible, then you’re in luck! We’re looking for alpha testers to help us swat bugs and iron out niggles, and we plan to start running test-representative samples within a few weeks. So come aboard, it’s time to generalise some findings!