Using Disaggregated Process Theories to Supplement Regression-Based Analysis

In this contribution, we argue that analysts need to go beyond simple causal graphs in order to reap fully the analytical benefits of using within-case methods such as process tracing as a supplemental tool that can strengthen confidence in the causal nature of findings from quantitative regression-based analyses (e.g., Seawright 2016). Instead, we contend that analysts need to unpac the causal structure of a causal process in more granular detail in order to trace the process empirically, as well as explore the contextual conditions underlying a correlation found using regression-based cross-case analysis. We argue that using simple causal graphs that draw arrows between independent, intervening and dependent variables does not enable the researcher to engage in the type of careful, withn-case detective work that is required to critically assess whether a correlation is actually causal and how the linkages in-between actually work.