PROJECT SUMMARY The protective effect of education against Alzheimer’s disease and related dementias (ADRD) is found to be smaller for racial/ethnic minority individuals than for non-Hispanic White individuals. Identifying factors that modify or mediate the education-ADRD relationship presents a unique opportunity to address the racial/ethnic disparities in ADRD and ultimately improve population health. However, most research relies on respondents’ education to assess the education-ADRD relationship, failing to consider the roles of other family member’s education or the social context and environment in which education occurs and affects ADRD. The major goal of the proposed research is to develop a novel framework that incorporates multigenerational education and multidomain contextual measures and use it to clarify the heterogeneous effects of education on ADRD risk among racial/ethnic groups by analyzing nationally representative longitudinal data. We propose to achieve this goal through three specific aims. In Aim 1, we will develop a multigenerational framework and use it to estimate the effect of multigenerational education on the respondent’s ADRD risk and the racial/ethnic disparities. Aim 2 is to construct multidomain contextual measures of structural inequality and racism and estimate their effects on individual’s ADRD risk and the racial/ethnic disparities. To provide historical and population contexts for understanding results from Aims 1 and 2, in Aim 3 we will estimate how education effects on ADRD risk may differ across four birth cohorts. The proposed research is significant because it will help reduce racial/ethnic disparities in cognitive impairment and ADRD risk by clarifying multigenerational pathways and identifying mediators and modifiers of the heterogeneous education-ADRD relationships among demographic groups. The proposed research is innovative because it will develop a new multigenerational and multilevel mediation approach to understanding why the education-ADRD risk relationship differs among racial/ethnic groups. It will also contribute new contextual data to a nationally representative data set.