1. Species distribution models (SDMs) estimate habitat suitability for species in geographic space. They are extensively used in conservation under the assumption that there is a positive relationship between habitat suitability and species success and stability.

2. Given the difficulties in obtaining demographic data across a species’ range, this assumption is rarely tested. Here we provide a range-wide test of this relationship for the eastern subspecies of purple martin (Progne subis subis).

3. We build a well-supported SDM for the breeding range of the purple martin, and pair it with an unparalleled demographic dataset of nest success and local and regional abundance data for the species to test the proposed link between habitat suitability and fecundity and demography.

4. We find a positive relationship between regional abundance and habitat suitability but no relationship between local abundance or fecundity and habitat suitability.

5. Our data suggest that local success is driven largely by biotic and stochastic factors and raise the possibility that purple martins are experiencing a time lag in their distribution. More broadly our results call for caution in how we interpret SDMs and do not support the assumption that areas of high habitat suitability are the best areas for species persistence.

Puma_samples_10km2.csv - Purple martin occurrence dataset for North America from eBird. Dataset was filtered to remove errors and thinned to a maximum of 1 record per 10km2

MartinWatch_nestlevel.csv - Citizen science records of nest success between 1995-2016 from the Purple Martin Conservation Association. Recording errors and low record numbers are removed as detailed in the paper.

PMCA_survey.csv - Responses of Purple Martin Conservation Association members to a survey regarding management practices and nest occupancy. Answers refer to their colony success in the year 2016.