Appendix E. Mixed-model analysis of the probability of emergence in the first growing season for transplanted G. tricolor seeds.
Fixed effects for the split-plot model are shown here, although the full model also included transect (nested within boundary type) and its interactions with population zone and removal treatment as random effects. Significance testing is based on restricted maximum likelihood and a binomial error structure. Population zone is a fixed factor with three levels: core population sites, edge sites, and peripheral sites beyond major population boundary. There are two population boundary types: the oak-canopy boundary and the valley bottom boundary. (N = 2816)
Effect |
Numerator df |
Denominator df |
F |
P |
Population zone |
2 |
11.3 |
4.09 |
0.0463 |
Removal treatment |
1 |
19.4 |
22.13 |
0.0001 |
Population zone × removal treatment |
2 |
19.3 |
6.82 |
0.0058 |
Boundary type |
1 |
7.5 |
4.21 |
0.0767 |
Population zone × boundary type |
2 |
11.3 |
2.24 |
0.1520 |
Removal treatment × boundary type |
1 |
19.4 |
0.78 |
0.3893 |
Zone × removal × boundary type |
2 |
19.3 |
1.38 |
0.2757 |