Perform Spread-Level test¶

Source: man/performSpreadLevelTest.Rd · Help: ?performSpreadLevelTest

Tests the slope of log(|residuals|) on log(fitted).

Usage¶

performSpreadLevelTest(model, data = NULL)

Arguments¶

model: an object of class lm.
data: optional data frame the model was fitted to. When NULL, the default, the model frame of model is used. The statistic is computed from the residuals and fitted values of model, so data does not change the result; it is accepted so that every test takes (model, data, ...).

Value¶

An object of class htest.

Details¶

A slope near 0 indicates constant variance. A significant slope suggests a power transformation may stabilize the spread.

References¶

Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association, 74(368), 829–836.

Examples¶

 data(mtcars)
 m <- lm(mpg ~ wt + qsec, data = mtcars)
 performSpreadLevelTest(m)