Hi, i am working on a data set with EDA. That includes QQ-Plots of residuals vs expected normal distribution. What puzzles me is that the range of ordinate and abscissae is so different: while the theoretical quantiles range from [-2, 2] the sample quantiles on the ordinate do extent from [-20, 50]. Quite obviously some kind of transformation is done. Although i intensively RTFM i could not find, what is done here. What exactly characterizes the range of the ordinate in QQ-Plots ? cheers, Fred -- Fred Jopp | fjopp at zedat.fu-berlin.de Lab.f. Soil Zoology & Ecology | www.biologie.fu-berlin.de/bodenzoo/ Grunewaldstr. 34 | Tel. +49.30.8385.5946 D-12165 Berlin | FAX +49.30.8385.3886
On 8 Dec 02, at 18:31, Fred Jopp wrote:> Hi, > > i am working on a data set with EDA. That includes QQ-Plots of > residuals vs expected normal distribution. > What puzzles me is that the range of ordinate and abscissae is > so different: while the theoretical quantiles range from [-2, 2] > the sample quantiles on the ordinate do extent from [-20, 50]. > Quite obviously some kind of transformation is done.?? What transformation? The ordinate is simplest the variable you supplied, and its units are what they are. And the range of the normal quantiles (the abscissa) is driven by the number of observations you supplied.> > Although i intensively RTFM i could not find, what is done here. > What exactly characterizes the range of the ordinate in QQ-Plots ? >Rather than TFM, perhaps an elementary introduction to QQ-plots? ---JRG> cheers, > Fred > -- > Fred Jopp | fjopp at zedat.fu-berlin.de > Lab.f. Soil Zoology & Ecology | www.biologie.fu-berlin.de/bodenzoo/ > Grunewaldstr. 34 | Tel. +49.30.8385.5946 > D-12165 Berlin | FAX +49.30.8385.3886 > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > http://www.stat.math.ethz.ch/mailman/listinfo/r-helpJohn R. Gleason Syracuse University 430 Huntington Hall Voice: 315-443-3107 Syracuse, NY 13244-2340 USA FAX: 315-443-4085 PGP public key at keyservers
Dear Fred, At 06:31 PM 12/8/2002 +0100, you wrote:>i am working on a data set with EDA. That includes QQ-Plots of >residuals vs expected normal distribution. >What puzzles me is that the range of ordinate and abscissae is >so different: while the theoretical quantiles range from [-2, 2] >the sample quantiles on the ordinate do extent from [-20, 50]. >Quite obviously some kind of transformation is done. > >Although i intensively RTFM i could not find, what is done here. >What exactly characterizes the range of the ordinate in QQ-Plots ?I assume that you're plotting against the quantiles of the standard normal distribution. Unless your residuals are standardized, there's no reason to suppose that the scales would be similar. Moreover, one generally looks simply for a linear pattern in the QQ plot, suggesting that the data might come from the reference distribution, though possibly with a different centre and scale. Various departures from linearity in the plot suggest skewness, heavy tails, outliers, etc., relative to the reference distribution. I hope that this helps, John ----------------------------------------------------- John Fox Department of Sociology McMaster University Hamilton, Ontario, Canada L8S 4M4 email: jfox at mcmaster.ca phone: 905-525-9140x23604 web: www.socsci.mcmaster.ca/jfox -----------------------------------------------------
On Sun, 8 Dec 2002, Fred Jopp wrote:> i am working on a data set with EDA. That includes QQ-Plots of > residuals vs expected normal distribution. > What puzzles me is that the range of ordinate and abscissae is > so different: while the theoretical quantiles range from [-2, 2] > the sample quantiles on the ordinate do extent from [-20, 50]. > Quite obviously some kind of transformation is done. > > Although i intensively RTFM i could not find, what is done here. > What exactly characterizes the range of the ordinate in QQ-Plots ?A qqnorm plot (is that what you meant?) is of a sample against a *standard* normal. One expects a straight line of slope sigma and intercept mu. The answer to your final question is `whatever the users specifies'. QQ plots are much more general than you seem to think. -- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595