Using Time-Series Network Analysis to Identify Diffusion Patterns: Literature Review and Proposed Methodology
To identify the causal mechanisms responsible for the diffusion of a given policy (e.g. coercion, economic competition, shared norms, and learning) scholars often use ‘process-tracing,’ or ‘spatial lags’ in regression models (Gilardi, 2012; Simmons & Elkins, 2004). Each of these methods involves different sets of trade-offs. Process-tracing is data intensive, and generated inferences are usually not generalizable beyond the cases observed. Conversely, using spatial lags in regression models can be limiting as not all causal mechanism are easily operationalized into a quantitative model (e.g. norm diffusion), and simultaneously controlling for alternative mechanisms is often unfeasible due to limited degrees of freedom. I posit that causal mechanisms generate distinct and observable spatiotemporal patterns, and present a new method of identifying these mechanisms using time-series network analysis.