||Empathy is a core component of human relationships and a cornerstone of effective interpersonal skills. Low levels of empathy are associated with negative outcomes while high empathy contributes to prosocial behaviors. A large body of literature exists on empathy but as yet there is no comprehensive review or meta-analysis of empathy training programs with client populations. This study is a meta-analysis that includes 24 studies of empathy training programs for client populations. The findings suggest that empathy training programs are effective in increasing empathy with an overall large effect size of g = 0.911. Moderator analyses were conducted on study design variables, program characteristics, and client variables. Moderator analyses suggest that empathy training is effective across a variety of client populations and training modalities. One significant moderator effect related to the way empathy is measured; studies using self-report outcomes show moderate gains in empathy (g = 0.386) while studies using observer-report outcomes show large gains (g = 1.488). Because of this large difference, the studies were divided by measurement type. Once the studies were divided by outcome type, none of the moderators continued to be significant across both measurement types. So while it appears that empathy training has a moderate to large effect on increasing empathy, the way empathy is measured is important and moderates the strength of the effect. These findings lead to a discussion on the measurement of empathy, the components of empathy, and the very construct of empathy. The study concludes with recommendations for further research on empathy training programs.