These functions calculate measures of the change in the covariance
matrices for the fixed effects based on the deletion of an
observation, or group of observations, for a hierarchical
linear model fit using lmer.
# S3 method for default covratio(object, ...) # S3 method for default covtrace(object, ...) # S3 method for mer covratio(object, level = 1, delete = NULL, ...) # S3 method for lmerMod covratio(object, level = 1, delete = NULL, ...) # S3 method for lme covratio(object, level = 1, delete = NULL, ...) # S3 method for mer covtrace(object, level = 1, delete = NULL, ...) # S3 method for lmerMod covtrace(object, level = 1, delete = NULL, ...) # S3 method for lme covtrace(object, level = 1, delete = NULL, ...)
| object | fitted object of class |
|---|---|
| ... | do not use |
| level | variable used to define the group for which cases will be
deleted. If |
| delete | index of individual cases to be deleted. To delete specific
observations the row number must be specified. To delete higher level
units the group ID and |
If delete = NULL then a vector corresponding to each deleted
observation/group is returned.
If delete is specified then a single value is returned corresponding
to the deleted subset specified.
Both the covariance ratio (covratio) and the covariance trace
(covtrace) measure the change in the covariance matrix
of the fixed effects based on the deletion of a subset of observations.
The key difference is how the variance covariance matrices are compared:
covratio compares the ratio of the determinants while covtrace
compares the trace of the ratio.
Christensen, R., Pearson, L., & Johnson, W. (1992) Case-deletion diagnostics for mixed models. Technometrics, 34(1), 38--45.
Schabenberger, O. (2004) Mixed Model Influence Diagnostics, in Proceedings of the Twenty-Ninth SAS Users Group International Conference, SAS Users Group International.
Adam Loy loyad01@gmail.com
data(sleepstudy, package = 'lme4') ss <- lme4::lmer(Reaction ~ Days + (Days | Subject), data = sleepstudy) # covratio for individual observations ss.cr1 <- covratio(ss) # covratio for subject-level deletion ss.cr2 <- covratio(ss, level = "Subject") if (FALSE) { ## A larger example data(Exam, package = 'mlmRev') fm <- lme4::lmer(normexam ~ standLRT * schavg + (standLRT | school), data = Exam) # covratio for individual observations cr1 <- covratio(fm) # covratio for school-level deletion cr2 <- covratio(fm, level = "school") } # covtrace for individual observations ss.ct1 <- covtrace(ss) # covtrace for subject-level deletion ss.ct2 <- covtrace(ss, level = "Subject") if (FALSE) { ## Returning to the larger example # covtrace for individual observations ct1 <- covtrace(fm) # covtrace for school-level deletion ct2 <- covtrace(fm, level = "school") }