Barrier 4: Fallacy of Averages

In the drawing below on the right you see two equally sized groups of ones and three. Their average is then a collection of twos but each two is just a calculated construct. We have only ones and threes.

Imagine that you are participating in an opinion poll and you have to select one of two statements, A or B. You choose either A or B, but you may read in the results that people favored A by 74%. However, no respondent answered with 74% A and 26% B; rather 100% A or 100% B.

If the groups of ones stand for women and the group of threes stand for men, what is then the average?

When researchers study culture, they often take representative samples of people, submit questions to them and calculate the average answers. However, a representative sample is always an approximation of the composition of people in a given group. And the questions may not be properly grounded in theory or not formulated carefully or … The biggest problem, however is the calculation of averages because the average does not show the variation (standard deviation) and hence, not the richness of culture. Bringing down an aspect of culture to a specific number can for that reason not be more than an indication of where you might be finding answers to your quest of understanding the other culture.