A Meta-Analysis of Empathy Training Programs for Client Populations

Purpose: 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. Empathy is frequently a target of social work interventions that target child abuse, interpersonal violence, bullying, and sexual offending. 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 research questions for this study are as follows: 1. Do empathy training programs increase empathy in client populations? 2. Is there variation in empathy training program outcomes? 3. What moderating variables contribute to empathy training programs and what is the relative strength of these variables?

Method: A systematic, multistep sampling procedure was used to identify studies for inclusion in the analysis. To identify potential articles, a literature search was conducted of major electronic databases using the search terms empathy and (train* or interven* or treat* or educ* or program*). The data bases were searched using online portals: EBSCO was used to search CINAHL, ERIC, MEDLINE, and PSYCHINFO; CSA Illumina for Social Service Abstracts, Sociological Abstracts, Criminal Justice Abstracts; and OVID for Social Work Abstracts. The date range of the studies included in the analysis ranged from 1976 to 2008. The statistical analysis used Hedges' g as the index of effect and a mixed-effect model.

Results: The findings suggest that empathy training programs are effective in increasing empathy with an overall large effect size of g = 0.911, although significant variation is present in the study Qw (23, k = 24) = 358.407, p = 0.001 and I2 = 93.583. Due to this heterogeneity of effect, moderator analyses were conducted on study design variables and study quality, program characteristics, and client variables. The analysis suggests that empathy training is effective across a variety of client populations and training modalities. One significant moderator effect relates 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, studies were partitioned by measurement type and moderator analyses were conducted within these groups of data. Once the studies were divided by outcome type, none of the moderators continued to be significant across both measurement types. Although several moderators, like the use of psycho education, gender of the participant, and voluntary status show significant interaction effects that lead to recommendations of practice and further research.

Implications: 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.