Engagement and Mentor Support as Drivers of Social Development in the Project K Youth Development Program

No difference in gender representation was found between participants excluded due to missing data (n = 38) and those retained (n = 327). Comparing participants identifying as Māori, Pacific, or European ethnicity, chi square analyses revealed that European youth were more likely, χ ² (1) = 17.20, p < .001, while Pacific youth were less likely to be retained, χ ² (1) = 15.59, p < .001. Due to low response frequencies (<5 for excluded data) it was not possible to compare Asian, other, or combined ethnicities. Finally, independent t-tests with a Bonferroni correction were employed to assess baseline social resources between groups. Those excluded due to missing data reported significantly higher levels of social self-efficacy at baseline (M = 4.15, SD = 0.91) than those retained in the analyses (M = 3.84, SD = 0.89), p = .01. The implications of these systematic group differences are addressed in the discussion.

Sample sizes for gender allowed for a robust assessment of the difference in program effects across groups (144 females to 183 males; both n > 100). Other groupings were not balanced and consisted of smaller sample sizes (for example, only the sample of European ethnicity had n > 100) or contained too small a range to make comparisons meaningful (e.g., age). Therefore, in accordance with conventions of invariance testing (Thompson and Green 2006), other group comparisons were not carried out.