How to fail at breaking down a study population into groups

1. Inform the public that you’re doing a study on comparing rates of schizophrenia and other psychoses in immigrants of different ethnic groups in Sweden with the majority population in Sweden (i.e. Swedish-born people with Swedish-born parents).

2. Get funding and obtain a large sample size (N = 1.47 million).

3. Break down the sample size into the following, seemingly random, categories based on country of origin:

A. Sweden
B. Finland
C. Western countries (Norway, Denmark, Iceland, Germany, Great Britain, USA, Canada, Australia, Other Western)
D. Eastern and Southern Europe (Poland, Hungary, Other Eastern Europe, Yugoslavia, Greece, Italy, Other Southern Europe)
E. Non-European (Middle East: Turkey, Iran, Iraq, Other Middle East; Asia: Far East, South Asia; Latin America: Chile, Other Latin America; Africa)

4. Claim that your break-down of the population into these groups is due to the diversity of ethnic backgrounds of immigrants in Sweden, which left you no other option but to create categories based on “similarities in lifestyle and living conditions.”

(4b. Because surely the lifestyle and living conditions of immigrants from Finland is hugely different to that of immigrants from Norway, as opposed to immigrants from Japan, Chile, Iran and The Country of Africa, which are practically like part of the same big happy colourful nation.)

5. Congratulations, you’ve arrived at your destination: FAIL.

Hjern, A., Wicks, S., & Dalman, C. (2004). Social adversity contributes to high morbidity in psychoses in immigrants – A national cohort study in two generations of Swedish residents. Psychological Medicine, 34(6), 1025-1033.