Table 2.

Summary of linear regressions modeling the explanatory power of interactions between contexts differing in richness context by a single community member

Model (n = 272 paired interactions)	df	adjusted R2	p value
focal change ∼ total change	1	0.273	<2.2e−16
focal change ∼ n-contexts	4	0.026	0.024
focal change ∼ interaction effect	1	0.159	<2e−12
focal change ∼ total change + interaction effect	2	0.417	<2.2e−16

“Focal change” indicates the change in density of the focal isolate in the predicted context. “Total change” indicates the change in total density between the community contexts of the interaction. “n-contexts” indicates the richness contexts from the pair of interactions (e.g., 1=>2 & 2=>3). “Interaction effect” indicates the change in density of the focal isolate in the interaction context which was not being predicted. We modeled all 272 interactions that were observed between richness contexts differing by a single community member. We considered the explanatory power of interactions from the bottom-up, i.e., “interaction effects” came from the lower richness context (continuing the example above, 1=>2), while “total change” came from the predicted higher richness context (2=>3), as in the models described in Table 1. A “+” in the model indicates the separate effects with no interaction term. The “p value” column indicates the p value of the model itself.