Graphs and GIFs to illustrate covariances of cross-model effects
Graphs and GIFs to illustrate covariances of cross-model effects
All graphs illustrate marginal effect estimates for a single continuous independent variable across: (1) 1,000 samples, and (2) two separate models.
Graphs on the left below illustrate a situation with a high covariance of the effects across samples; the difference in effects is highly consistent across the different samples.
Graphs on the right below illustrate a situation with a low covariance of the effects across samples; the difference in effects changes across samples without a clear pattern.