Why the contrast matrices are different for order and unordered factored? On Mon, Nov 9, 2009 at 1:12 PM, Greg Snow <greg.s...@imail.org> wrote: > Mostly it is a conceptual difference. An unordered factor is one where there > is no inherent order to the levels, examples: > > Color of car > Race > Nationality > Sex > State/Country of birth > Etc. > > In the above, the order of the levels could be changed without it really > changing the meaning (think of the order of bars in a bar chart). We may > want to print/plot in some specific order such as alphabetic for easy lookup > or based on the summary values of another vector for nice looking plots, but > there is no overriding reason why we would order color as blue/green/red vs. > green/red/blue, etc. > > Ordered factors have some natural order, for example maybe you are studying a > drug and have doses labeled as Low, Medium, and High. It makes the most > sense to print and plot in that order rather than alphabetically (High, Low, > Medium). Any continuous variable that has been cut into categories (best not > to do this, but if done) has a natural order. Survey questions where you > response can range from strongly disagree to strongly agree are usually > ordered (but there may be disagreement on what the correct ordering is). > > In R the most apparent effects of using ordered vs. factor is in how they > print out and how some modeling functions default to handling them (the > default contrasts for ordered factors is different, rpart treats ordered > factors differently). > > Hope this helps, > > -- > Gregory (Greg) L. Snow Ph.D. > Statistical Data Center > Intermountain Healthcare > greg.s...@imail.org > 801.408.8111 > > >> -----Original Message----- >> From: r-help-boun...@r-project.org [mailto:r-help-boun...@r- >> project.org] On Behalf Of Peng Yu >> Sent: Sunday, November 08, 2009 4:28 PM >> To: r-h...@stat.math.ethz.ch >> Subject: [R] ordered factor and unordered factor >> >> I don't understand under what situation ordered factor rather than >> unordered factor should be used. Could somebody give me some examples? >> What are the implications of order vs. unordered factors? Could >> somebody recommend a textbook to me? >> >> ______________________________________________ >> 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. >
______________________________________________ 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.
