Behavioral science has a coverage problem: too much of it studies the same wealthy, Western few. A common reply is that going global is simply impractical. Is it?
Much of behavioral science draws on samples that are Western, Educated, Industrialized, Rich, and Democratic — WEIRD — and then generalizes to humanity. A frequent rejoinder is that doing better is infeasible, because online survey platforms cover too few places to reach the rest of the world.
That infeasibility claim, the paper argues, is overstated. Existing commercial survey infrastructure already provides access to nationally representative samples in a strikingly large number of countries.
That coverage is far wider than the platforms most researchers default to. Where MTurk and Prolific concentrate on a handful of wealthy countries, the broader marketplace spans over a hundred — enough to make a study substantially less WEIRD than the critique assumes.
the paper reports coverage of over 110 countries via Lucid — far broader than MTurk or Prolific; platform details are in the paper
If the infrastructure to study the world already exists, then a WEIRD literature is increasingly a choice, not a constraint. The paper hands researchers both the argument and the tools to make their work more representative of the species it claims to describe.
Crabtree, Charles. “Yes, You Can Do Global, Cross-Cultural Behavioral Science Research Using Existing Survey Firms.” Proceedings of the National Academy of Sciences, 2025.
read the paper →the “110+” figure and platform comparison are drawn from the paper's abstract; full details and caveats are in the paper