Monitoring long—term population change is an integral part of effective conservation—oriented research and management, and is central to the current debate on the status of Neotropical migrant land birds. However, the analysis of count data such as the Breeding Bird Survey is complicated by the subjective nature of trend estimation, and by limitations inherent to extensive, volunteer—based surveys, such as measurement error and missing data. A number of analysis methods have been used that differ in their approach to dealing with these complications and produce different estimates of population change when applied to the same data. There is, however, no consensus as to which method is the most suitable. Many analytical issues remain unresolved, such as model of trend, observer effects, treatment of missing observations, distribution of counts, and data selection criteria. These issues make it difficult to evaluate the relative merits of the methods, although a number of new approaches (nonlinear regression, Poisson regression, estimating equations estimates) offer promising solutions to some problems. I suggest the use of Monte Carlo simulations to empirically test the performance of the methods under realistic, spatially explicit scenarios of population change, and provide an example of the approach.