Effects of Classroom Likeability Composition on Adolescent Loneliness: A Brief Introduction to the Group Actor-Partner Interdependence Model (GAPIM) (SRA)

Abstract

Ample evidence shows that loneliness is related to low peer acceptance (Kingery et al., 2011; Vanhalst et al., 2014). This is in line with social needs theory which assumes that loneliness occurs when important social needs are not met, such as being accepted by peers (Weiss, 1973). This implies that low peer acceptance by itself may cause loneliness. However, this effect also may depend on the group in which low acceptance occurs. If an adolescent is disliked by other peers, feelings of social exclusion and loneliness may be stronger when these others are generally liked than when they are generally disliked. We examined this frog pond effect (Davis, 1966) of likability on loneliness. We also tested for optimal distinctiveness (Brewer, 1991) which posits that it matters how similar an adolescent is to others relative to how similar they are to each other. Higher heterogeneity among others’ likeability may lead to less loneliness. We hypothesized that similarity between individual and group acceptance affects loneliness over and beyond the effect of low acceptance on loneliness. Participants were 945 adolescents (Mage = 14.27; SDage = 1.25; 49% boys) from 41 classrooms (grades 7 to 10) who completed peer ratings of likeability (-3 = very disliked; +3 = very well-liked) and self-reports of loneliness (LLCA; Marcoen, Goossens, & Caes, 1987). Likeability was the average rating received from classmates (M = 1.19; SD = .64; range -1.83 to +2.47). Loneliness scores ranged from 1 to 3.83 on a 4-point scale with higher scores indicating more loneliness (M = 1.39; SD = .46). The Group Actor-Partner Interdependence Model (GAPIM; Garcia, Meagher, & Kenny, 2014; Kenny & Garcia, 2012) was used to investigate whether participants’ loneliness was affected by: their own likeability (actor effect), others’ average likeability (others’ effect), how similar participants were to all other group members (actor similarity effect), and the similarity among all others (others’ similarity effect). Six nested multilevel models were run. The full contrast model was the most parsimonious and best fitting (Table 1). To interpret the effects, predicted means were plotted for prototypical groups varying in likeability composition (Figure 1). The negative actor effect and positive others’ effect confirmed the hypothesized frog pond effect; adolescents compare themselves to others in terms of likeability and these comparisons related to their felt loneliness. As shown in Figure 1, loneliness was most pronounced for disliked adolescents in groups of generally liked others (Group B). The negative actor similarity effect and positive others’ similarity effect were in line with optimal distinctiveness; loneliness was lowest when (a) adolescents were similar to the average likeability of their group, and (b) there was heterogeneity among the other group members (Group A and D). This study demonstrated the GAPIM in research on person-environment fit and confirmed the importance of the peer group context to understand adolescent loneliness. It also showed that the GAPIM is a promising new approach to examine influence processes in groups and to test theories that may explain these effects.