Content area
Full text
* Corresponding author. Email: [email protected]
Multi-method research approaches have become increasingly popular in recent years as tools to make more robust causal inferences in the social sciences (Beach and Rohlfing 2018; Goertz 2017; Humphreys and Jacobs 2015; Lieberman 2005; Schneider and Rohlfing 2013, 2016; Seawright 2016). 1 The most common combination involves cross-case comparative analysis (e.g. statistically assessing mean causal effects of a large number of cases) and in-depth within-case analysis (e.g. process-tracing case studies).
The promise of multi-method research in comparative politics is that different methodological tools can compensate for each other’s relative weaknesses, enabling more robust causal inferences to be made. Yet while much progress has been made, there is still considerable confusion about the underlying assumptions and ontological/epistemological underpinnings of different methods for causal inference. The result is that scholars interested in using multi-method designs in the study of comparative politics will receive very different guidance in different accounts, making it into almost an ‘everything goes’ situation.
This contribution intends to clear up some of the confusion by identifying the key points of contention underlying the current debates about multi-method research. Drawing on recent developments in the broader philosophy of science literature (Clarke et al. 2014; Russo and Williamson 2011), and within social science methodology (Beach and Pedersen 2016; Goertz and Mahoney 2012; Ragin 2000), I put forward that there is a larger methodological divide than commonly understood, making true multi-method research very difficult. The divide is between what can be termed a ‘bottom-up’ case-based approach that focuses on tracing how causal mechanisms play out in individual cases, and a ‘top-down’ variance-based approach that assesses the mean causal effect of variables within a population (or sample thereof).
This review article starts by introducing the ontological and epistemological underpinnings of different methods by differentiating approaches into a bottom-up, case-based, and the top-down, variance-based, approach, focusing in particular on their relative strengths and weaknesses in making causal inferences. The key strength of case-based studies is that we learn how a causal process actually works in a given case (or small set of cases); termed ‘how actually’ explanations in the literature. However, the downside is that we are left in the dark regarding how it works within a larger, more diverse population. In essence,...