Heteroscedasticity

April 6, 2008
By

(This article was first published on V. » R, and kindly contributed to R-bloggers)

If a model is estimated using the following code:
lm(y~x1+x2)->p

1. bptest(p) does the Breuch Pagan test to formally check presence of heteroscedasticity. To use bptest, you will have to call lmtest library.

2. If the test is positive (low p value), you should see if any transformation of the dependent variable helps you eliminate heteroscedasticity. Also check if the right hand side of the model is okay.

3. If 2 does not work, you can use the white’s heteroscedasticity-corrected covariance matrices to make inference. Package car has a function hccm that gives you the heteroscedasticity-corrected covariance matrix (there is a similar function in package sandwich also). coeftest(p,vcov=hccm(p)) will give you the results of the tests using this matrix. Use these results instead of summary(p).

library(lmtest)
library(car)
bptest(p)
coeftest(p,vcov=hccm(p))

Tags:

To leave a comment for the author, please follow the link and comment on his blog: V. » R.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.