I’ll first note that I haven’t (yet) read the study this article references myself. However, as represented in this article from Reuters, this is some icky pseudoscience which apparently fails to take into account that the pink = femininity association is historically a very recent development, arising sometime in the past 100 years. Prior to that, in Victorian England, blue was the color for girls. In other parts of the world at other points in history, there are other associations. (No citations for my own arguments, alas, because I don’t have them handy and I’m too tired to search some out.)
In their experimental group, they seem to have found a gender divide between the colors preferred by women versus men. Then they proceed to “speculate that this sex difference arose from sex-specific functional specialization in the evolutionary division of labor.” Rather than a much simpler explanation: our color preferences are heavily influenced by our social environment. Consider how quickly a particular shade can pass from neutral to favored to distinctly dated in the fashion world. Color preferences are not static, and while individuals clearly have native preferences, it’s also clear that these preferences are shaped by our social groups.
Insisting that a pattern found in a small group (which was most likely taken from a pool of people who are all members of more-or-less the same cultural group!) must not only apply to all people around the world, but that this pattern necessarily has a biological and evolutionary basis is an overinterpretation of the data and is piss-poor science.
If any of my readers have a source through which I can read the original paper, I’ll be most appreciative. If they’ve addressed any of these points in the paper, I’d hate to be complaining unnecessarily. I suspect, though, that getting to read the original source will only allow me to skewer it more effectively.
And remember, boys and girls, when you’re doing science, look for the simplest explanation that accounts for all known data! Sometimes a correlation is just a correlation.