similar to: Cook's Distance in GLM (PR#9316)

Dear list, I've noticed that influence.measures and cooks.distance gives different results for non-gaussian GLMs. For example, using R-1.9.0 alpha (2003-03-17) under Windows: > ## Dobson (1990) Page 93: Randomized Controlled Trial : > counts <- c(18,17,15,20,10,20,25,13,12) > outcome <- gl(3,1,9) > treatment <- gl(3,3) > glm.D93 <- glm(counts ~ outcome +

Bug PR#9316 noted an inconsistency between the Cook's distance contours on plot.lm(x, which = 5) and the values given by cooks.distance(x) -- as shown in plot.lm(x, which = 4) -- for glms: http://bugs.r-project.org/cgi-bin/R/Analyses-fixed?id=9316;user=guest;selectid=9316 The suggested fix was to modify the contour levels by a dispersion factor, implemented as follows: dispersion <-

There may be a bug in the anova.glm function. deathstar[32] R R : Copyright 2004, The R Foundation for Statistical Computing Version 2.0.1 (2004-11-15), ISBN 3-900051-07-0 R is free software and comes with ABSOLUTELY NO WARRANTY. You are welcome to redistribute it under certain conditions. Type 'license()' or 'licence()' for distribution details. R is a collaborative project

I have a puzzle as to how R is computing Cook's distance in weighted linear regression. In this case cook's distance should be given not as in OLS case by h_ii*r_i^2/(1-hii)^2 divided by k*s^2 (1) (where r is plain unadjusted residual, k is number of parameters in model, etc. ) but rather by w_ii*h_ii*r_i^2/(1-hii)^2 divided by k*s^2,

I don't have a copy of Dobson (1990) from which the glm.D93 example is taken in example("glm"), but I'm strongly suspecting that these are made-up data rather than real data; the means of the responses within each treatment are _identical_ (equal to 16 2/3), so two of the parameters are estimated as being zero (within machine tolerance). (At this moment I don't understand

Dear useRs, I have some trouble with the calculation of Cook's distance in R. The formula for Cook's distance can be found for example here: http://en.wikipedia.org/wiki/Cook%27s_distance I tried to apply it in R: > y <- (1:400)^2 > x <- 1:100 > lm(y~x) -> linmod # just for the sake of a simple example >

Hi! The nobs() method for glm objects always returns the number of cases with non-null weights in the data, which does not correspond to the number of observations for Poisson regression/log-linear models, i.e. when family="poisson" or family="quasipoisson". This sounds dangerous since nobs() is, as the documentation states, primarily aimed at computing the Bayesian

Hi there I have a post I would like to put on the "95% confidence intercal with glm" thread. Thank-you so much! I am wondering first of all if anyone knows how to calculate confidence intervals for a GLMM? I use the lme4 library. Also, I am wondering how to predict a model mean and confidence intervals for a particular independent variable? For example in the following example:

Dear list, I just hit an error that stopped my make && make check-devel operation on my linux box (FC3, i686 P4 2GB RAM). Just to note that I've been building the development branch(?) for some time and this is the first hint of a problem. 1) updated the src tree using svn update 2) ran ../configure --with-recommended-package=no from my build directory 3) got: R is now configured

Following p.206 of "Statistical Models in S", I wish to change the code for summary.glm() so that it estimates the dispersion for binomial & poisson models when the parameter dispersion is set to zero. The following changes [insertion of ||dispersion==0 at one point; and !is.null(dispersion) at another] will do the trick: "summary.glm" <- function(object, dispersion =

Hello, I'm learning to use Sweave, and I've run into a problem: sometimes, when entering long lines of input and using long variable names, Sweave will not insert linebreaks in a way that respects the width setting. This causes undesirable overflows into the margins in the latex file. For example, consider the following document (adapted from the GLM example): \documentclass{article}

Hi there, I have the problem, that I'm not able to reproduce the SPSS residual statistics (dfbeta and cook's distance) with a simple binary logistic regression model obtained in R via the glm-function. I tried the following: fit <- glm(y ~ x1 + x2 + x3, data, family=binomial) cooks.distance(fit) dfbetas(fit) When i compare the returned values with the values that I get in SPSS,

the data file has such structure: 1992 6245 49 . . 20 1 0 0 8.739536 0 . . . . . . . . "alabama" . 0 . 1993 7677 58 . . 15 1 0 0

Well I suppose a warning's not going to hurt. Even in a case like the occupationalStatus example where you know some points have been fitted exactly, it might be useful to be reminded that the standardised residuals for these points are then NaN and cannot be displayed. Of course when you don't know in advance that this issue will arise, there is even more reason to give a warning.

Greetings, I am running glm models for species counts using a poisson link function. Normal summary functions for this provide summary statistics in the form of the deviance, AIC, and p-values for individual predictors. I would like to obtain the p-value for the overall model. So far, I have been using an analysis of deviance table to check a model against the null model with the intercept as

Hi all I am preparing a document using Sweave; a really useful tool. But I am having a problem. Consider this toy example Sweave file: \documentclass{article} \begin{document} <<echo=TRUE,results=verbatim>>= options(width=40) # Set width to 40 characters hide <- capture.output(example(glm)) # Create an example of the problem, but hide the output summary(glm.D93) #

Hello R users I like to extract z values for x1 and x2. I know how to extract coefficents using model$coef but I don't know how to extract z values for each of independent variable. I looked around using names(model) but I couldn't find how to extract z values. Any help would be appreciated. Thanks TM ######################################################### >summary(model) Call:

The following code demonstrates an annoyance with plot.lm(): library(DAAGxtras) x11(width=3.75, height=4) nihills.lm <- lm(log(time) ~ log(dist) + log(climb), data = nihills) plot(nihills.lm, which=5) OR try the following xy <- data.frame(x=c(3,1:5), y=c(-2, 1:5)) plot(lm(y ~ x, data=xy), which=5) The "Cook's distance" text overplots the label for the point with the

Hi Developers, The alpha, compiles successfully, but it is failing make check-all (on two seperate machines, both FreeBSD 6.1). Here is the version string: platform i386-unknown-freebsd6.1 arch i386 os freebsd6.1 system i386, freebsd6.1 status alpha major 2 minor 3.0 year 2006 month 03 day 27 svn rev

OK, I really dread to ask that .... much more that I know some discussion about p-values and if they are relevant for regressions were already on the list. I know to get p-val of regression coefficients - this is not a problem. But unfortunately one editor of a journal where i would like to publish some results insists in giving p-values for the squared deviance i get out from different glm and