Mixed effects regression models tend to become common practice in the field of Social Sciences. However, diagnostic tools to evaluate these models lag behind. For instance there is no general applicable tool to check whether all units (or cases) roughly have the same influence on the regression parameters. It is however commonly accepted that tests for influential cases should be performed, especially when estimates are based on a relatively small number of cases. Testing for influence with mixed effects models is especially important in Social Science applications, for two reasons. First, models in the Social Sciences are frequently based on large numbers of individuals while the number of higher level units is often relatively small. Secondly, often the higher level units are remarkably similar, for instance in the case of neighboring countries.
Influence.ME is a new package for R which provides two innovations for evaluating influential data in mixed effects models (also called multilevel models): it extends existing procedures for use with mixed effects models, and it allows to not only search for single influential cases, but for combinations of cases that as a combination exert too much influence.