Michael: > I am doing a two-sided DW test: > > H0: rho = 0 > H1: rho =/= 0 > > My understanding is that most test statistics tables are one-sided. It's the > way they created the table.
...because rho > 0 is the alternative of interest in most applications. > So from online, by doing Googling, I found a bunch of DW tables for > Confidence Level 5%. Using tables for the DW test is difficult because it's distribution depends on the particular set of regressors used. The tables of DW just give upper and lower bounds. Back when the DW test was suggested, tables was the only way to make application of the test feasible. Today, you would either use the exact combination of chi-square distributions or an asymptotic approximation (both implemented in dwtest() from "lmtest") or a bootstrap approximation (implemented in durbin.watson() from "car"). For 278 observations, the normal approximation should be sufficient. hth, Z ______________________________________________ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.