Empirical generalization concerning word-of-mouth marketing using meta-analysis

Based on a literature review of word-of-mouth communication, a theoretical framework is proposed and tested, in which word-of-mouth (WOM) is considered the main construct, satisfaction and loyalty are the antecedents and WOM valence (i.e., positive, negative and mixed) is the moderator. This theoretical model is tested using data obtained from a meta-analysis. After reviewing 664 studies, a total of 90 are submitted to the analyses, since they have the relationships of interest and the required statistics (i.e., correlations between constructs or statistics for group differences). The analyzed subset of studies produced 119 independent samples and 36,673 respondents. The results showed a significant association of satisfaction and loyalty with WOM. More specifically, WOM valence introduced a significant moderating effect. This moderating effect is proposed in this paper as the following empirical generalization: considering satisfaction and loyalty as correlates of WOM, satisfaction is the variable that has a stronger relationship with positive WOM, while loyalty is the one that is more closely associated with negative WOM, which might be expressed as SAT→WOM(+), LEA→WOM(-).