Reports the Bonferroni p-values for testing each observation in turn to be a mean-shift outliner, based Studentized residuals in linear (t-tests) and generalized linear models (normal tests).
outlierTest(model, ...)# S3 method for lm
outlierTest(model, cutoff=0.05, n.max=10, order=TRUE,
labels=names(rstudent), ...)
# S3 method for outlierTest
print(x, digits=5, ...)
an lm or glm model object.
observations with Bonferonni p-values exceeding
cutoff are not reported, unless no observations are
nominated, in which case the one with the largest Studentized
residual is reported.
maximum number of observations to report (default, 10).
report Studenized residuals in descending order of magnitude?
(default, TRUE).
an optional vector of observation names.
arguments passed down to methods functions.
outlierTest object.
number of digits for reported p-values.
an object of class outlierTest, which is normally just
printed.
For a linear model, p-values reported use the t distribution with degrees of
freedom one less than the residual df for the model. For a generalized
linear model, p-values are based on the standard-normal distribution. The Bonferroni
adjustment multiplies the usual two-sided p-value by the number of
observations. The lm method works for glm objects. To show all
of the observations set cutoff=Inf and n.max=Inf.
Cook, R. D. and Weisberg, S. (1982) Residuals and Influence in Regression. Chapman and Hall, https://conservancy.umn.edu/handle/11299/37076.
Fox, J. (2016) Applied Regression Analysis and Generalized Linear Models, Third Edition. Sage.
Fox, J. and Weisberg, S. (2019) An R Companion to Applied Regression, Third Edition, Sage.
Weisberg, S. (2014) Applied Linear Regression, Fourth Edition, Wiley.
Williams, D. A. (1987) Generalized linear model diagnostics using the deviance and single case deletions. Applied Statistics 36, 181--191.
# NOT RUN {
outlierTest(lm(prestige ~ income + education, data=Duncan))
# }
Run the code above in your browser using DataLab