Methods: This study used a longitudinal design with data collected from adolescent girls (N = 4,693) who participated in the Adolescent Girls Empowerment Program in Zambia. We analyzed data collected at Rounds 2 and 3. Nutritional status, self-efficacy, and gender attitudes were measured at Round 2. Cognitive and academic skills were measured at round 3, one year later. Confirmatory Factor Analysis (CFA) was used to determine whether our hypothesized latent variables adequately represented the relationship that exists in the data before estimating the structural model. After assessing the adequacy of the measurement model, a structural equation model (SEM) was specified to test direct and indirect associations.
Results: Results indicated adequate fit of our hypothesized structural equation model (χ2 [163, N = 4,345] = 824.668, p < .001, RMSEA = .031 [90% CI, .029, .033], CFI = .946, TLI = .927, SRMR = .042). Underweight adolescent girls reported significantly lower cognitive and academic skills than adolescent girls who were not underweight (β = -0.436, p = .008). Being underweight was also associated with lower self-efficacy (β = -0.139, p = .020). Self-efficacy and gender attitudes were significantly associated with cognitive and academic skills. Higher levels of self-efficacy were directly associated with higher cognitive (β = 0.283, p < .001), literacy (β = 0.471, p < .001), numeracy (β = 0.169, p = .001) and arithmetic (β = 0.137, p = .002) skills. Self-efficacy mediated the association of underweight with cognitive and academic skills. For every one-standardized unit increase in gender-biased attitudes at Round 2, cognitive, literacy, numeracy and arithmetic skills at Round 3 decreased by 0.394, 0.557, 0.288, and 0.319 points, respectively (p < .001). Our results found no significant association of normal weight with cognitive and academic abilities.
Conclusions and Implications: The consistent association of undernutrition with adverse outcomes illustrates long-term vulnerabilities that adolescent girls may experience due to the interplay of undernutrition, low self-efficacy, and weak cognitive and academic abilities during their adolescence. School meal programs could combine nutrition support with age-appropriate and gender-responsive socioeconomic interventions to increase self-efficacy and to improve cognitive and academic skills in adolescent girls.