Heterozygosity-fitness correlations (HFCs) have been examined in a wide diversity of contexts, and the results are often used to infer the role of inbreeding in natural populations. Although population demography, reflected in population-level genetic parameters such as allelic diversity or identity disequilibrium, is expected to play a role in the emergence and detectability of HFCs, direct comparisons of variation in HFCs across many populations of the same species, with different genetic histories, are rare. Here, we examined the relationship between individual microsatellite heterozygosity and a range of sexually selected traits in 660 male guppies from 22 natural populations in Trinidad. Similar to previous studies, observed HFCs were weak overall. However, variation in HFCs among populations was high for some traits (although these variances were not statistically different from zero). Population-level genetic parameters, specifically genetic diversity levels (number of alleles, observed/expected heterozygosity) and measures of identity disequilibrium (g2 and heterozygosity-heterozygosity correlations) were not associated with variation in population-level HFCs. This latter result indicates that these metrics do not necessarily provide a reliable predictor of HFC effect sizes across populations. Importantly, diversity and identity disequilibrium statistics were not correlated, providing empirical evidence that these metrics capture different essential characteristics of populations. A complex genetic architecture likely underpins multiple fitness traits, including those associated with male fitness, which may have reduced our ability to detect HFCs in guppy populations. Further advances in this field would benefit from additional research to determine the demographic contexts in which HFCs are most likely to occur.