It's not necessarily a bad thing that 67% of studies are done on psychology undergrads. 90% or so of medical studies are done on mice. If you find something big, you should go outside your cheapest testing group (be it mice or psychology undergrads), but if something fails to produce interesting results even on them, you just saved yourself a lot of money and effort - and failures will be much more common than interesting finds.
Most of the research on cognitive biases and other psychological phenomena that we draw on here is based on samples of students at US universities. To what extent are we uncovering human universals, and to what extent facts about these WEIRD (Western, Educated, Industrialized, Rich, and Democratic) sample sources? A paper in press in Behavioural and Brain Sciences the evidence from studies that reach outside this group and highlights the many instances in which US students are outliers for many crucial studies in behavioural economics.
Epiphenom: How normal is WEIRD?
Henrich, J., Heine, S. J., & Norenzayan, A. (in press). The Weirdest people in the world? (PDF) Behavioral and Brain Sciences.
Broad claims about human psychology and behavior based on narrow samples from Western societies are regularly published in leading journals. Are such species-generalizing claims justified? This review suggests not only that substantial variability in experimental results emerges across populations in basic domains, but that standard subjects are in fact rather unusual compared with the rest of the species - frequent outliers. The domains reviewed include visual perception, fairness, categorization, spatial cognition, memory, moral reasoning and self‐concepts. This review (1) indicates caution in addressing questions of human nature based on this thin slice of humanity, and (2) suggests that understanding human psychology will require tapping broader subject pools. We close by proposing ways to address these challenges.