Which dimension drives a country’s composite score determines the correct causal explanation.

Evidence

This has important implications for the dependent variable side. The causal mechanism and hypotheses must focus on, say, explaining why a country scored low on a given dimension. For example, India usually scores quite low, around 133rd out of 188 in 2016, on the Gender Inequality Index. On two of the three dimensions India scores reasonably well (adolescent birth rate and share of seats in parliament), but on the third (maternal mortality) it does very badly. Thus, explaining why India scores low on GII comes down to explaining why it does poorly on that dimension. Aggregation and concept structure analysis (see chapter 6) are critical because a different concept structure—best shot—could lead to a different explanatory strategy. — Gary Goertz, Social Science Concepts and Measurement, 527

Categorization

• XREF: Decompose-before-explaining logic parallels epidemiological work on which cause drives a composite mortality rate.
• QUESTION: How often do published cross-national regressions explain a composite outcome without first checking which sub-dimension actually moves the variation?