Count data are often the most available indices for bat abundance. Counts of bats are useful for estimating differences in population size, between different habitats, or at different times. But such estimates are not without complications. We used pre-existing bat count data sets, and simulated bat count data to survey a variety of methods for interpreting bat counts. These methods included, evaluation of habitat preferences through AIC model selection of generalized linear models, evaluation of differences in abundance using hierarchical models, and evaluation of spatially replicated time series dynamics using additive mixed-models. All methods proved useful under the special circumstances that accommodated model assumptions. Generalized linear models required the most restrictive assumptions, while hierarchical and additive models allowed many assumptions to be relaxed. We identified several areas where current modeling practices might be improved.