Dear Kathryn, On Tue, 10 Mar 2015 03:24:03 -0700 (PDT) kat123 <kathryn.harr...@northampton.ac.uk> wrote: > I have run a logit data transformation in R using the logit function in the > package car. > > http://cran.r-project.org/web/packages/car/car.pdf > > If i run logit on a column of data that contains a 0 value it makes and > adjustment according to the literature of 0.025. > > I thought this meant that it was running the transformation as > > log((p+0.025)/ (1-(p+0.025)))
It's not that simple -- think what would happen if there were both 0s and 1s in the data: > p <- 0:1 > log((p+0.025)/ (1-(p+0.025))) [1] -3.663562 NaN Warning message: In log((p + 0.025)/(1 - (p + 0.025))) : NaNs produced > logit(p) [1] -3.663562 3.663562 Warning message: In logit(p) : proportions remapped to (0.025, 0.975) > log(c(0.025, 0.975)/(1 - c(0.025, 0.975))) [1] -3.663562 3.663562 To see what logit() does simply print it by typing logit at the command prompt. I hope this helps, John ------------------------------------------------ John Fox, Professor McMaster University Hamilton, Ontario, Canada http://socserv.mcmaster.ca/jfox/ > > However, if I run individual values through this equation they do not match > up to the output of the logit function. > > Any suggestions? > > > > > > -- > View this message in context: > http://r.789695.n4.nabble.com/logit-in-car-package-tp4704408.html > Sent from the R help mailing list archive at Nabble.com. > > ______________________________________________ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > 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. ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.