Yes, I want the same test as is done for b[1] and b[2] in the summary table -- for H0: b[]==0.
OK, do F-test on full model with cf and sf vs reduced model with intercept only. I want to test y~cf + sf vs y~ intercept (ie mean) -- I guess I just use var(y) Thanks Duncan for your help Bill On Thu, Jun 17, 2010 at 11:18 AM, Duncan Murdoch <murdoch.dun...@gmail.com> wrote: > William Simpson wrote: >> >> Suppose I do a trigonometric regression >> fit<-lm(y~ cf + sf) >> where cf and sf are the cos and sine components. >> >> b<-coef(fit) >> I have the fitted sine component b[2] and the cos component b[3]. >> Doing summary(fit) gives me the p-values and SEs for b[2] and b[3]. >> >> But I want the amplitude of the fitted waveform >> amp<-sqrt(b[2]^2+b[3]^2) >> >> Can someone please tell me how to get the p-value for amp? >> > > "the p-value for amp" is ambiguous; p-values refer to tests, not functions. > But let's assume you want to test whether amp = 0. Then this is achieved > by an F test comparing the model with cf and sf versus one without it. > You'll see it in summary(fit) at the bottom of the display. If you want to > include other covariates in the model, you can use anova, e.g. > > anova(lm(y ~ other), lm(y ~ cf + sf + other)) > > Duncan Murdoch > > > > > ______________________________________________ R-help@r-project.org 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.