Table 1.
Model comparison for predicting bee observed sex ratio.
| Predictor(s) | Residual deviance | Significance - p | Effect Size - β | AIC |
|---|---|---|---|---|
| 500 m | 20.2824 | (500 m)6.60e−5 | −0.32 ± 0.08 | 187.38 |
| 1 km | 17.6024 | (1 km)1.76e−5 | −0.40 ± 0.09 | 184.7 |
| 1.5 km | 14.9124 | (1.5 km)4.51e−6 | −0.51 ± 0.11 | 182.01 |
| 2 km | 13.4224 | (2 km)2.08e−6 | −0.57 ± 0.12 | 180.53 |
| 2 km + Floral abundance | 13.2023 | (2 km)8.83e−6 (Abund)0.84 | (2 km)−0.57 ± 0.13 (Abund)4.25e−7 ± 2.13−6 | 182.35 |
| 2 km + Floral area | 11.9723 | (2 km)6.71e−6 (Area)0.26 | (2 km)−0.54 ± 0.12 (Area)2.55e−7 ± 2.26e−6 | 181.13 |
| 2 km + Richness | 13.1923 | (2 km)1.17e−5 (Richness)0.82 | (2 km)−0.55 ± 0.13 (Richness)2.31e−4 ± 0.001 | 182.34 |
Models were GLMs with poisson distribution and log-link function. The first four rows present the effect of impervious surface cover measured at different scales (i.e. within 500 m of the garden, within 1 km, etc.). Floral metrics are season-long means (abundance and area; averaged over the number of floral surveys) or totals (richness) for an area within 20 m of the study location.