This dissertation explores issues regarding the legitimacy of the Flynn effect by using the National-Longitudinal-Survey-of-Youth Children (NLSYC) data, in which there are scores from a mathematics subscale of an achievement test, the Peabody Individual Achievement Test (the PIAT). Study I explores the Flynn effect within population subgroups by demographic characteristics: gender, race/ethnicity, maternal education, household income, and urbanization within the PIAT Mathematic (PIAT-M) subscale. Study II explores the Flynn effect at the item level of the PIAT-M and identifies possible causes of the Flynn Effect using expert ratings of item contents. Results indicated interaction effects for mothers' education or household incomes, specifically, children with older college educated mothers or children born to higher income households had seen an accelerated Flynn effect in their PIAT-M scores than their peers with older lower educated mother or lower income households. At the item level, reasoning domain most consistently predicted the Flynn effect for the children with average IQ. Recall, computation, reasoning and algebra were predictive of the Flynn effect when children in both extreme ends of the IQ scale were included in the analysis. These results add to the literature in understanding the operation of the Flynn effect that had not been carefully studied in the past.