Strengthening Causal Inference through Qualitative Analysis of Regression Residuals: Explaining Forest Governance in the Indian Himalaya

Abstract

This paper contributes to fertile debates in environmental social sciences on the uses of and potential synergies between qualitative and quantitative analytical approaches for theory development and validation. Relying on extensive fieldwork on local forest governance in India, and using a dataset on 205 forest commons, we propose a methodological innovation for combining qualitative and quantitative analyses to improve causal inference. Specifically, we demonstrate that qualitative knowledge of cases that are the least well predicted by quantitative modeling can strengthen causal inference by helping check for possible omitted variables, measurement errors, nonlinearities in posited relationships, and possible interaction effects, and thereby lead to analytical improvements in the quantitative analysis. In the process, the paper also presents a contextually informed and theoretically engaged empirical analysis of forest governance in north India, showing in particular the importance of institutional and historical factors in influencing commons outcomes.