Shared dimensions between measured variables produce definitional, not causal, correlations.

Evidence

As concepts increase in dimensionality and breadth, one runs an increased risk of “conceptual tautology.” Conceptual tautology occurs when the independent variable has some of the same secondary-level dimensions as the dependent variable. For example, if the dependent variable has four dimensions and the independent variable has three and they have two dimensions in common, there are significant levels of conceptual tautology. The statistical analysis might find a statistically significant relationship; however, the relationship is not causal, but conceptual and definitional. — Gary Goertz, Social Science Concepts and Measurement, 541

Categorization

• DEFINE: Names and specifies ‘conceptual tautology’ as overlapping secondary-level dimensions between independent and dependent variables.
• XREF: Connects to causal-inference work on measurement validity and to the social capital literature’s insistence that indicators not be definitionally entangled with outcomes.
• SEED: An essay on how quantitative findings in the social sciences are routinely correlation artifacts of definitional overlap, with a checklist for diagnosing tautology before running regressions.