Depensatory population dynamics (Allee effects) may be defined simply as a decrease in population growth rate at low density.  Although these types of dynamics have played an increasingly important role in population dynamics modeling, a recent meta-analysis by Sibly et al. (2005) suggested that less than 1% of 3200+ time series supported models that allowed for depensation.  Previous analyses have focused on traditional hypothesis testing (e.g. regressing per capita growth rate against data).  This type of analyses has a number of problems, the largest being that not all data are included in the likelihood calculation.  In this talk, I present a Bayesian state-space solution to this problem.  Treating the unobserved states as latent variables, I illustrate the benefits of considering both observation and process error.  Because the focus of this analysis is on the relationship between growth and density (and not on selecting a particular population model), the results may be applied generally to any population.  These methods are applied to thousands of simulated data sets, and finally to 1400 records from the Global Population Dynamics Database.  I explore the effects of time series length, population parameters, and magnitude of errors on the evidence supporting Allee dynamics.