Model averaging for the BEWR suggested that cavity type was 15 times stronger at predicting successful nests than vegetation cover, though both were significant (Table 2; see Table S1 for candidate model selection), with nests more likely to be successful as vegetation cover increased (β = 0.06), and if nests were built in an abandoned woodpecker cavity over a natural cavity (β = 0.95). Model averaging for both the BCTI and the ATBC suggested that decay and the cavity type were significant predictors for nest success. As with the GFWO, with every unit increase in decay, nest success dropped 0.19 for ATBC and 0.41 for BCTI. Again, cavity type was the strongest predictor; cavity type was 3 times stronger at predicting nest success than decay for the BCTI, and was 4 times stronger than decay for the ATBC. Across SCB species, cavity type was the strongest predictor of nest success.