Poster #123 - Empathy and Prosocial Behavior from Childhood through Young Adulthood: An Application of Integrative Data Techniques

• In Event: 3-053 - Poster Session 10
  In Poster Session: PS 10 Section - Moral Development

Integrative Statement

Empathy is a multidimensional construct with both affective and cognitive components that results from the comprehension or apprehension of another person’s situation or emotions (Davis, 1994; Eisenberg, Spinrad, & Knafo-Noam, 2015). Processes such as empathy are believed to motivate prosocial behavior (Einolf, 2008) by prompting children to express concern for others and act on that concern (Eisenberg et al., 2014). Empathy and prosocial behavior develop across the life span, but most studies focus on a limited age period. For example, there has been relatively little longitudinal research from adolescence into young adulthood (Eisenberg et al., 2015).
To address these issues, we capitalized on our access to five longitudinal data sets with three measurements of empathy and prosocial behavior, and we pooled them together using a set of analytic tools termed integrative data analysis (IDA). IDA allows researchers to combine data from existing studies by addressing heterogeneity due to sampling, study, and other factors (e.g., Curran & Hussong, 2009). Researchers then fit one model, or set of models, to the aggregated data (e.g., Johnson et al., 2018). Using this technique, we investigated the development of empathy and prosocial behavior from childhood through young adulthood. Participants included 1,781 adolescents. At the first data collection, the mean age was 11.59 (SD = 3.72, range 8 – 22); 52.1% identified as girls/women, and no racial identification was a majority. The empathy items included: I feel sorry for another child who is hurt or upset, when I see someone being teased or picked on I feel sorry for them, and when I see someone being picked on I want to help them. Prosocial behavior was measured using two items regarding the frequency of helping people that the participant knew (friends, family members, neighbors); the word of the questions varied slightly by age to be developmentally appropriate. All items had a five-point Likert-type response scale. We fit a parallel process growth curve model to model trajectories of empathy and prosocial behavior simultaneously (Grimm & Ram, 2012). We also included age as a predictor of all intercepts and slopes. This model fit the data reasonably well, RMSEA = 0.047 (90% CI 0.034, 0.061), CFI = 0.978. The average intercepts were 4.02 (p < .001) for empathy and 4.86 for prosocial behavior, with significant variation (ps < .001) for both. The average slope for empathy (0.071) was positive but not significantly different from zero (p = .181), but it did vary across participants (p = .005). The average slope for prosocial behavior was -0.120 (p = .022), also with significant variation between participants (p = .007). Higher initial scores on empathy were associated with higher initial scores on prosocial behavior, and the slopes were also associated. However, higher initial scores on empathy and prosocial behavior were associated with weaker slopes (potentially due at least in part to a ceiling effect). Age at first time point was negatively associated with the initial value of prosocial behavior. Ongoing analyses include growth mixture modeling to identify qualitatively different types of trajectories.

Authors

• Sara K Johnson, Tufts University
  Non-Presenting Author

• Michelle B Weiner, Tufts University
  Presenting Author

• Mary H Buckingham, Tufts University
  Non-Presenting Author