Surveys fail to capture big five personality traits in non-WEIRD populations

Questions commonly used to explore the "Big Five" personality traits--Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism--generally fail to measure the intended personality traits in developing countries, according to a new study. This is because measurement of these traits relies on surveys typically applied to White Educated Industrialized Rich and Democratic (WEIRD) populations. The Big Five personality trait model is used to inform issues related to crime, employment, and wages in high-income countries. Increasingly, it has been applied to explore related issues in middle- and low-income countries, too. To better understand if this approach is valid in these countries, Richard Laajaj and his team analyzed data including 29 face-to-face surveys conducted with over 90,000 respondents, often in their local language, across 23 low and middle-income countries. Their findings included that the relationship between Conscientiousness--a trait particularly predictive of earnings in the U.S.--was not a significant income predictor in 10 of the 14 developing countries. And also in contrast with most findings in WEIRD populations, Openness appears to be a strong predictor of income in developing countries, the authors say. However, data the team collected for the same set of countries via the internet more closely reflected the Big Five traits as observed in WEIRD populations--suggesting cultural differences are not the main driver of the surveys' low validity. Rather, say the authors, language used to administer the survey may be a key influence, as surveyors can explain things differently in person. The researchers suggest several factors may explain why the survey questions were less valid in the surveyed countries, including stronger bias in these places towards providing agreeable responses during face-to-face interviews. "It really gives us a warning to be careful before we start expanding the use of these surveys too much," says Laajaj. "We need to adapt them more to this context and perhaps combine them with other ways to capture these abilities and these personality traits. That can have a lot of impact on future policies." Lajaaj and the team suggest reducing survey biases by randomly assigning surveyors, improving surveyor training, and shifting to self-administrated surveys.

Credit: 
American Association for the Advancement of Science