It is often assumed that indicator species can act as surrogates for conservation of other taxa because the habitat characteristics that limit the distribution of indicators should also affect multiple co-occurring native species. However, this assumption is rarely explicitly tested. Here, we test whether four fish species are potential indicators of overall fish diversity in the Bear River drainage (UT and WY, USA). We then examine whether biodiversity indicator relationships are driven by shared habitat associations among species by examining species-habitat models. We use these models to test for correlations between indicator species and common co-occurring species in predicted occurrence and density. Fish diversity was higher at sites with potential indicators than sites without, indicating that conservation aimed at any of these species is likely to affect a broad number of co-occurring taxa. However, with minor exceptions, habitat correlations were inconsistent between indicator and co-occurring species. Instead, most species showed idiosyncratic habitat relationships. Thus, while conservation of habitats containing indicator species is likely to positively affect the distribution of a minority of co-occurring taxa, they may have neutral (or even negative) effects on the distribution or abundance of a majority of other native taxa that are common to the region. These data suggest that reliance on indicator taxa in conservation can be misleading because they obscure important ecological information about affected non-target species.