In mist-net studies, it is often difficult to use capture-recapture methods to estimate number of birds present. Many investigators use number of birds captured as an index of population size. We investigate the consequences of using indices of bird abundance as surrogates for population size in hypothesis tests. Unless all of the birds present are captured, indices are biased estimates of local population size, and the amount of bias depends on the proportion of birds captured. We demonstrate the potential effects of bias on hypothesis tests based on indices. The bias generally causes type I error rates to be inflated. Investigators should either estimate the proportion of animals captured using capture-recapture methods or demonstrate that results of hypothesis tests based on indices are not consequences of bias in the indices.