Study 1
Using a quasi-experimental design, Brown and colleagues (2009) assessed the impact of the adult mentoring program on the psychosocial well-being of youth-headed households in the Gikongoro province of Rwanda. Youth-headed households in two of the four districts in the province received the mentoring program during the study period, and youth-headed households in the other two districts served as the comparison group. Following the study period, youth in the comparison districts received the mentoring program. Data were collected at baseline and at follow up 18 months later. Interviews were conducted with youth who were 24 years old and younger, at baseline; and 27 years old and younger, at follow up. The questionnaire was administered in the predominant language of the study area, Kinyarwanda.
The baseline sample included 692 youth (intervention group = 347, comparison group = 345); the follow-up sample at 18 months included 593 youth. At baseline, the average age was 20.6 years for the intervention group and 19.7 for the comparison group. Less than half of the overall sample at baseline was female (45.5 percent of the intervention group and 47.2 percent of the comparison group). More than 60 percent reported having served as the head of their household for 4 or more years, and 70 percent reported that both parents were dead (91 percent reported that their mother was dead, and 75 percent reported that their father was dead). There were no statistically significant equivalences between the intervention and comparison groups at baseline with respect to age (the intervention group was older) and education (the intervention group had higher levels of education). Furthermore, the groups had statistically significant differences on outcome variables; the intervention group reported having less adult support, greater marginalization, and higher levels of grief and depression. These differences were controlled for in the outcome analyses regardless of statistical significance.
Of youth in the intervention group, 49 percent reported being visited by their mentors at least weekly, 35 percent reported being visited once or twice a month, and 16 percent reported that their mentors visited them less than once a month.
Outcomes of interest included adult support, marginalization, grief, and depression. Adult support was measured through a 4-item scale assessing the presence of a trusted adult who offered advice and guidance, assisted in going to authorities for help, provided comfort when the respondent was sick or sad, and who the respondent could always depend on. Marginalization was measured through a 6-item scale to characterize the level experienced by the youth. The scale included such items as “people in this community would rather hurt you than help you”, “you feel isolated from others in the community”, “people speak badly about you or your family”, and “the community rejects orphans”. Grief was measured through a 7-item scale that included the following variables: 1) you think about the death of your loved one(s) almost all the time; 2) you feel angry when you think about the death(s); 3) you still can’t believe your loved one(s) is/are really dead (or gone); 4) your faith in God is shaken since the death of your loved one(s); 5) since the death of your loved one, you have lost confidence in people; and 6) since the death of your loved one, life is meaningless. The scales for adult support, marginalization, and grief were scored on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree). Higher scores indicated greater adult support, marginalization, and grief. Finally, depression was measured through the 20-item Centers for Epidemiologic Studies–Depression Scale (Radloff 1977). Possible scores ranged from 0 to 60, with a higher score indicating greater depressive symptoms. This instrument was adapted for use in a non-Western setting by using youth focus groups and other qualitative methods to identify local terms for distress. The final translations were also reviewed by a committee of local youth and professionals.
Program effects on the four outcomes were assessed using linear regression analyses to test for a group-by-time interaction, controlling for age, gender, educational achievement, whether the youth heads of household lived alone or with other children, and whether they had lost a parent in the genocide. Analyses also controlled for self-reported health status (excellent/good, fair, and poor), whether youth reported eating more than one meal per day, number of basic services youth had received from the basic program, and asset ownership (counting items such as shoes, a blanket, and other basic essentials). Analyses were conducted on those individuals who completed both baseline and follow-up data collection. The study authors did not conduct subgroup analyses.
