More on Stress & Depression: Next Steps for the Mentoring Study

Hi everyone!!

It’s hard to believe that the summer is coming to a close. For those of you who have been reading since the very beginning, thank you!! In this last blog post I want to return to the findings I covered two posts ago (Stress and Depression Matter: Analysis of Preliminary Results).

In that post I discussed some preliminary findings–particularly the positive correlation between mentor early life stress and youth relationship satisfaction and the negative correlation between mentor depression and mentor relationship satisfaction. While these correlations are significant in and of themselves, my advisor and I also wanted to see if these findings remained constant across schools.

Thus far, I had analyzed my variables at one level: match level. This means that I’d looked at the effects of mentor stress and depression on each relationship as an individual. However, because our mentoring program partner has chapters at different schools across the country, my variables could also vary at a school level. Perhaps mentors at one school have elevated rates of depression in comparison to mentors at another school? Perhaps youth at one school are more excited about their time with mentors than youth at another school? These school level variances could be impacting my data and obscuring the findings. This is called a nesting effect.

In order to account for the potential nesting effect, I reran my my analyses using a hierarchical linear model (HLM). The HLM allowed me to see how much variance in the data was caused by differences across schools. It also allowed me to hold school differences constant and more accurately observe effects of mentor stress and depression.

HLM yielded results consistent with my preliminary findings: the correlations I observed through single level regression maintained their statistical significance in an HLM. This means the the correlations between mentor stress and youth relationship satisfaction and mentor depression and mentor relationship satisfaction maintained across schools. School level variance only had a statistically significant impact on mentor relationship satisfaction. Mentors at different schools reported statistically significant different levels of relationship satisfaction. However, even with this effect, mentor depression maintained a negative correlation with mentor relationship satisfaction. On all other outcomes (youth relationship satisfaction, mentor avoidance), there was no school level variance.

These findings demonstrate the strength of mentor stress and depression as predictors of mentoring relationship satisfaction. I am hopeful that my work this summer will contribute to and perpetuate the mentoring research as well as inform educators and administrators working to implement mentoring programs in their schools.

If these blogs drew you in or sparked your interest, please reach out in the comments or continue to follow this research! As my work for the summer wraps up, I’ll focus on preparing these results for presentation. I am presenting in September at the Charles Center Summer Research symposium here on campus, and in November at the annual conference of the Association for Cognitive and Behavioral Therapies in Washington, D.C.. I am also writing up my research this summer for submission to a peer reviewed journal.

Until September!




  1. Adam Kearney says:

    Wow! So cool!

  2. Charles P says:

    Important research. Thanks for your strong efforts here and the overall work of the Charles Center.