A comparison of hierarchical linear and structural equation modeling approaches to evaluating youth mentoring program effectiveness
Description
This study compares hierarchical linear and structural equation modeling approaches to evaluating the effectiveness of a youth mentoring program Under a variety of conditions such as growth curve models assuming large sample size, the two approaches yield essential identical solutions. However, for smaller samples, it is not clear which approach will perform better. The two analytical methods are compared in terms of the estimate of regression coefficient, their standard errors and overall significance The assessment of the effectiveness of the youth mentoring program is approached in two ways. In a cross-sectional design, the study evaluates youth and mentor characteristics that are strongly associated with youth psychological outcomes. Particularly, the effect of gender and the quality of mentoring on youth psychological outcomes is examined The mentoring effect is also assessed using longitudinal data and thus estimating the rate of change in youth psychological wellbeing For both clustered cross-sectional data and longitudinal data, HLM and SEM provided essential similar coefficients and standards errors. However, SEM analysis provided an additional analysis of various psychosocial outcomes Using the depression scale as the psychological measure, the findings suggest that youth gender, number of meals per day, whether one or both parents were killed in genocide, marginalization from their surrounding communities, and grief levels were associated with youth depression symptoms. The negative attitude toward mentor was significantly associated with youth psychological outcomes, but there was no evidence of the effect of the similarity or dissimilarity in youth and mentor gender There was no variation in youth psychological wellbeing among youth-mentor dyads, but mentor psychosocial wellbeing factors have an important effect on youth psychological wellbeing. In addition, there was a modest improvement in youth psychological wellbeing, as reported by the mentors, during the first 9 months of the program. However, the levels plateaued during the last 3 months Overall, this dissertation enriches the understanding of the mentor effect on youth psychological wellbeing. Analytical and programmatic implications are also discussed