Context . Increasingly, ornithologists are being asked to identify major sources of avian mortality so as to identify conservation priorities.

Aims . Considerable evidence suggests that windows of office towers are a lethal hazard for migrating birds. The factors influencing the risk of bird–window collisions in residential settings are not understood as well.

Methods . Citizen scientists were requested to participate in an online survey that asked about characteristics concerning their homes and yards, general demographic information about participants, and whether they had observed evidence of bird–window collisions at their home.

Key results . We found that 39.0% of 1458 participants observed a bird–window collision in the previous year. The mean number of reported collisions was 1.7 ± 4.6 per residence per year, with 38% of collisions resulting in a mortality.

Conclusions . Collisions were not random, with the highest collision and mortality rates at rural residences, with bird feeders > rural residences without feeders > urban residences with feeders > urban residences without feeders > apartments. At urban houses, the age of neighbourhood was a significant predictor of collision rates, with newer neighbourhoods reporting fewer collisions than older neighbourhoods. Most people remembered collisions occurring in the summer months.

Implications . Our results are consistent with past research, suggesting that window collisions with residential homes are an important source of mortality for birds. However, we found large variation in the frequency of collisions at different types of residences. Proper stratification of residence type is crucial to getting accurate estimates of bird–window collisions when scaling local data into larger-scale mortality estimates.