A more scientific approach to applied economics: Reconstructing statistical, analytical significance, and correlation analysis

Abstract

There is a deep and well-regarded tradition in applied economic research and in other social sciences to assign causality to correlation analysis and statistical significance. I critique the application of correlation analysis, unsubstantiated with empirical backing of prior assumptions, as the core analytical measure for causation. Moreover, this paper critiques the past and current focus on statistical significance as the core indicator of substantive or analytical significance, especially when paired with correlation analysis. I attempt to explain the persistence of the misplaced use of these statistical techniques in the applied literature. I also propose a positive analytical frame wherein correlation analysis and tests of statistical significance, as part of a large analytical toolbox, can make a positive contribution to the analytical literature.