Abstract
This study investigated changes in the tripartite self-concept in a sample of pre-, early-, mid-, and late-adolescents. Participants provided self-descriptions under conditions that primed either a personal or social context. As expected, adolescents showed fluctuating use of the individual, relational, and collective levels of self. The collective and individual self became more prominent, particularly in later adolescence, while the relational self became less prominent. The individual self became more socially oriented through adolescence, and the relational self shifted focus from the family context in pre-adolescence to the peer context in early adolescence. Implications for theory on self-concept change and on adolescent identity are presented, and highlight the need for further research on the development of social identity in adolescence.
Notes
1. In the Australian education system, students begin school at 5 or 6 years of age, and they attend primary school to grade 6 and then secondary school for grades 7 to 12. Thus, the major social transitions are between grades 6 and 7 (aged 11 to 12 years) and after grade12 (age 17 to 18 years).
2. There were a number of TST responses that did not fit the coding scheme employed in this study and were classified “uncodable.” The distribution of uncodable responses was relatively even across conditions, and they were discarded from analyses.
3. Comparable analysis of the participants' importance ratings for self-descriptors, showed a pattern of results highly consistent with those for the proportion of descriptors generated. For example, a 4 (Group)×3 (Self-level Importance)×2 (Prime) mixed-design ANOVA on the importance ratings of individual, relational, and collective level self-descriptors showed a significant main effect of Self-level, F(1, 483) = 1787.18, p = .000, partial η2 = .79, and planned simple contrasts showed that participants rated individual self-descriptors as significantly more important than collective, F(1, 483) = 3497.87, p = .000, partial η2 = .88, and relational, F(1, 483) = 1598.41, p = .000, partial η2 = .77, descriptors. There was also a significant interaction between Group and Self-level, F(3, 483) = 9.26, p = .000, partial η2 = .05, and follow-up univariate tests revealed the same pattern of results as for proportion of descriptors generated. The importance of collective self-descriptors increased significantly through adolescence, F(3, 483) = 2.84, p = .04, partial η2 = .02, and was highest among late-adolescents, although Scheffe tests showed that the importance of collective self-descriptors reported by late-adolescents was not significantly higher than the other groups (ps > .05). The importance of individual self-descriptors also increased through the adolescence, F(3, 483) = 2.99, p = .03, partial η2 = .02, being highest among mid-adolescents. Post hoc tests showed mid-adolescents tended to rate greater importance to individual descriptors than pre-adolescents (p = .15) and early-adolescents (p = .06), and did not differ from late-adolescents. The importance of relational terms to describe the self also decreased through adolescence, F(3, 483) = 15.04, p = .000, partial η2 = .09, and Scheffe tests showed pre-adolescents and early-adolescents rated significantly greater importance to relational descriptors than mid-adolescents (p = .000 and .002, respectively) and late-adolescents (p = .000 and .02, respectively), who did not differ from each other. Also consistent with the results for proportion of descriptors, there were no significant effects due to the priming condition itself, or in interaction with Group, on the importance of individual, relational, and collective self-descriptors.