Fit indices for confirmatory factor analyses.

Associations between childhood poverty and cognitive outcomes have been examined from multiple perspectives. However, most evidence is based on cross-sectional data or longitudinal data covering only segments of the developmental process. Moreover, previous longitudinal research has mostly relied on data from Western nations, limiting insights of poverty dynamics in low- and middle-income countries. Here, we use data from the Mauritius Child Health Project, a large-scale prospective longitudinal study conducted in a then low-income country, to examine long-term associations between poverty in early childhood and cognitive performance across childhood and adolescence. Poverty-related factors were assessed at age 3 years and comprised indicators of psychosocial adversity and malnutrition. Cognitive functioning was assessed at ages 3 and 11 years by using standardized intelligence measures and at age 17 years by means of a computerized test battery. Using multiple hierarchical regression models, we found that chronic malnutrition and parental characteristics showed similar-sized, independent associations with initial cognitive functioning at age 3 as well as at age 11 years. For age 17 years, however, associations with early childhood risk factors vanished and instead, cognitive functioning was predicted by performance on prior cognitive assessments. Sex was also found to be a powerful predictor of cognitive trajectories, with boys improving and girls worsening over time, regardless of the level of their initial exposure to risk. The current findings indicate that, to prevent cognitive impairment, interventions tackling poverty and malnutrition should focus on the infancy period and be designed in a gender-sensitive way.