Here we use the glucocorticoid cascade hypothesis framework to address the role of baseline cortisol on changes in cognitive function over a 3-year span in non-demented rural Americans. We also determine if genotype at 4 different single nucleotide polymorphisms (SNPs) relates to change in cognitive function. We predicted 1) over time, increases in baseline cortisol will be associated with decline in cognitive function, 2) individuals homozygous for 3 CRFR1 SNP rare alleles (AA rs110402, TT rs7209436, and TT rs242924 vs. others) will show less cognitive decline and this will be particularly pronounced in those with lower baseline cortisol, and 3) FKBP5 T carriers (TT or CT vs. CC homozygotes) will have decreased cognitive performance and this will be particularly pronounced in individuals with higher baseline cortisol. Collectively, our data do not robustly support the glucocorticoid cascade hypothesis. In several cases, higher baseline cortisol related to better cognitive performance over time, but within individuals, increased cortisol over time related to decreased performance on some cognitive domains over time. Contrary to our predictions, individuals with the rare CRFR1 haplotype (AA, TT, TT) performed worse than individuals with the common haplotype across multiple domains of cognitive function. FKBP5 genotype status had minimal impacts on cognitive outcomes. Genotype effects were largely not dependent on cortisol. The Project FRONTIER dataset is supported by Texas Tech University Health Sciences Center Garrison Institute on Aging.