WEIRD has become a commonly used concept to describe a blind spot that many social sciences have been said to have because so much research tends to be done on convenience samples such as undergraduate students in the universities of the researchers, so it is something of a stretch to universalize their findings when there are so many variables that people around the world might have, culturally, economically and linguistically.
Some of this research can be interesting, but in my experience reading some of the literature, there does tend to be a bit of overegging of the pudding. How are the biases often detected? In some cases, they may be reaction times and we find that some people whose first language does not include certain features, press a button milliseconds faster than someone whose language does contain that feature.
Similarly, eye-tracking or EEG or proportions of respondents choosing one thing or another tends to be the measurement. But to what extent does it have real world significance, it’s often hard to say. In fact, in some cases, interesting findings often don’t replicate. For example, I recently finished reading a book called the Power of Language by Viorica Marian. It had some interesting findings on the way that grammatical gender influences speakers of, say, Spanish, where a tomato fruit is feminine and a plant is masculine. Apparently when asked if a tomato should have a girl’s name or a boy’s name, some respondents answered a girl’s name for tomato. Similarly, apparently a key in one language is described as pretty or solid depending on whether the language classes it as masculine or feminine. Yet it turns out that the original work by Boroditsky et. al. does not replicate. Same with claims that Chinese people think of time going up instead of along, the evidence is … tenuous.
Another favourite was the claim that in some languages people save more money if they have no strong future markers in their language. The theory is that the future is considered more immediate and therefore more important to the speakers. Apparently this is the case for Finnish and Estonian and some others. However, other linguists and economists are, naturally, skeptical and suggested that the reason why there might be a spurious correlation is down to the fact that languages are not independent of each other and therefore a shared behaviour across languages may only be accidentally related to similarly shared language backgrounds what with Estonian and Finnish being related etc…
Anyway, I haven’t looked into this particular paper, but making some assumptions….
