How much should we trust coalition identification in policy networks? A new method for the Advocacy Coalition Framework
Résumé
Identify political coalitions is crucial to understand precisely a policy process. Focusing on the coalition phenomena, the Advocacy Coalition Framework (ACF) is one of the most prominent approaches offering recently a fertile articulation with the policy network analysis. These studies apply frequently Block Modeling and Community Detection (BMCD) strategies to define homogeneous political groups. However, the BMCD literature is growing quickly showing a large variety of algo-rithms and interesting selection methods much more diverse than the ones used in the ACF.
Thus, identify the best option can be difficult and few ACF studies give an explicit justification. On the other hand, few BMCD works offer a systematic comparison on real social networks and never applied to policy network datasets. The paper offers a new relevant 4-Steps selection method to reconcile the ACF and BMCD advances.
Using an application on original African policy network data collected in Madagascar and Niger, we provide a useful set of practical recommendations for future ACF works using network analysis: (i) the density and the size of the policy network affect the identification process, (ii) the ”best algorithm” can be rigorously determined by maximizing a novel indicator based on the convergence and the homogeneity between algorithm results, (iii) researchers need to be careful with missing data, they affect the results and the imputation does not solve the problem