Binary logistic regression. We performed a logistic regression to calculate the propensity score of participation in the CRP and then ran the matching procedure. Generallypropensity score matching is applicable only if the exogenous variables of receiving the treatment (i.e., participation in a CRP) have been established theoretically or in previous empirical studies (Caliendo and Kopeinig 2008). Many studies outline the antecedents of articulating word of mouth (for an overview, see East, Hammond, and Wright 2007), so we consider this requirement fulfilled.