Table 2. GLMM for modelling seedling survival data from the growth-room experiment experiments.
| Fixed effects | Estimated coefficients (s.e.) | z-value | P (>|z|) |
|---|---|---|---|
| (Intercept) | −1.60 (0.17) | −9.44 | <2e−16 |
| Abundance | 3.27 (0.22) | 14.88 | <2e−16 |
| Density | 0.011 (0.014) | −0.76 | 0.45 |
| Distance | 0.18 (0.0068) | 25.96 | <2e−16 |
| Fungicide | 3.58 (0.26) | 13.53 | <2e−16 |
| Seed | 0.46 (0.084) | 5.47 | 4.46e−08 |
| Abundance × distance | −0.19 (0.018) | −10.76 | <2e−16 |
| Abundance × fungicide | −2.38 (0.34) | −7.00 | 2.59e−12 |
| Distance × fungicide | −0.15 (0.010) | −14.81 | <2e−16 |
| Fungicide × density | −0.0087 (0.022) | −0.40 | 0.69 |
| Fungicide × seed | −0.34 (0.13) | −2.63 | 0.0086 |
| Abundance × distance × fungicide | 0.14 (0.027) | 5.06 | 4.22e−07 |
| Random effect | Variance | P (>χ2) | |
| Site (intercept) | 1.68e−06 | 1.00 |
GLMM, generalized linear mixed effects model.
A GLMM was used to test the effects of adult abundance (0 for abundant populations and 1 for low-density populations), distance (0, 5, 10, 15 and 20 m from a focal tree), seeding density (1, 4 and 9 seeds per vessel), seed (0 for local seeds and 1 for nonlocal seeds), soil treatment (0 for no fungicide and 1 for fungicide) and interaction terms (for example, abundance × fungicide treatment) on seedling survival of O. glaberrima. The dependent variable of the GLMM was the count of dead and live seedlings by the end of the growth-room experiment. The estimated overdispersion variance was 0.33.