similar to: add1.lm and add1.glm not handling weights and offsets properly (PR#8049)

David, Thanks. The reason add1.lm (and drop1.lm) do not support offsets is that lm did not when they were written, and the person who added offsets to lm did not change them. (I do wish they had not added an offset arg and just used the formula as in S's glm.) That is easy to add. For the other point, some care is needed if 'x' is supplied and the upper scope reduces the number

Dear R-List users, Can anyone explain exactly the difference between Weights options in lm glm and gls? I try the following codes, but the results are different. > lm1 Call: lm(formula = y ~ x) Coefficients: (Intercept) x 0.1183 7.3075 > lm2 Call: lm(formula = y ~ x, weights = W) Coefficients: (Intercept) x 0.04193 7.30660 > lm3 Call:

Hi! I would like to perform an F-Test over more than one variable within a generalized mixed model with Gamma-distribution and log-link function. For this purpose, I use the package mgcv. Similar tests may be done using the function "anova", as for example in the case of a normal distributed response. However, if I do so, the error message "error in eval(expr, envir, enclos) :

Hello, I've written a simple (although probably overly roundabout) function to test whether two regression slope coefficients from two linear models on independent data sets are significantly different. I'm a bit concerned, because when I test it on simulated data with different sample sizes and variances, the function seems to be extremely sensitive both of these. I am wondering if

This is a question regarding the 'regTermTest' function in the 'survey' package. Imagine Z as a three level factor variable, and code ZB and ZC as the two corresponding dummy variables. X is a continuous variable. In a 'glm' of Y on Z and X, say, how do the two test specifications test.terms = c("ZB:X","ZC:X") # and test.terms = ~ ZB:X + ZC:X in

Ted Harding said: > I can get the estimated RRs from > RRs <- exp(summary(GLM)$coef[,1]) > but do not see how to implement confidence intervals based > on "robust error variances" using the output in GLM. Thanks for the link to the data. Here's my best guess. If you use the following approach, with the HC0 type of robust standard errors in the

Folks; I am having a problem with the cv.glm and would appreciate someone shedding some light here. It seems obvious but I cannot get it. I did read the manual, but I could not get more insight. This is a database containing 3363 records and I am trying a cross-validation to understand the process. When using the cv.glm, code below, I get mean of perr1 of 0.2336 and SD of 0.000139. When using a

I am attempting to run a glm with a binomial model to analyze proportion data. I have been following Crawley's book closely and am wondering if there is an accepted standard for how much is too much overdispersion? (e.g. change in AIC has an accepted standard of 2). In the example, he fits several models, binomial and quasibinomial and then accepts the quasibinomial. The output for residual

This is a continuing issue with the one on the list a long time ago (I couldn't find a solution to it from the web): -------------------------------------------------------------------------- > Using a formula converted with as.formula with lme leads > to an error message. Same works ok with lm, and with > lme and a fixed formula. > > # demonstrates problems with lme and

Dear all, I am comparing logistic regression models to evaluate if one predictor explains additional variance that is not yet explained by another predictor. As far as I understand Baron and Li describe how to do this, but my question is now: how do I report this in an article? Can anyone recommend a particular article that shows a concrete example of how the results from te following simple

dear list! i have run several glm analysises to estimate a mean rate of dung decay for independent trials. i would like to compare these results statistically but can't find any solution. the glm calls are: dung.glm1<-glm(STATE~DAYS, data=o_cov, family="binomial(link="logit")) dung.glm2<-glm(STATE~DAYS, data=o_cov_T12, family="binomial(link="logit")) as

There is a bug in AIC.default and AIC.lm, as illustrated below. (I've only checked this under 1.6.1, and can't easily check if it has already been reported since the site is down.) > lm1 <- lm(y ~ x, list(x=1:10, y=jitter(1:10))) > lm2 <- lm(y ~ x, list(x=1:10, y=jitter(1:10))) > AIC(lm1, lm2) df AIC lm1 3 -18.662493 lm2 3 -7.265906 > AIC(lm1, lm2, k = 2)

Using a formula converted with as.formula with lme leads to an error message. Same works ok with lm, and with lme and a fixed formula. # demonstrates problems with lme and as.formula demo<-data.frame(x=1:20,y=(1:20)+rnorm(20),subj=as.factor(rep(1:2,10))) demo.lm1<-lme(y~x,data=demo,random=~1|subj) print(summary(demo.lm1)) newframe<-data.frame(x=1:5,subj=rep(1,5))

Hello, I use a type III anova ("car" package) to analyse an unbalanced data design. I have two factors and I would have the effect of the interaction. I read that the result could be strongly influenced by the contrasts. I am really not an expert and I am not sure to understand indeed about what it is... Consequently, I failed to properly used the fit.contrast function (gregmisc

Hi all, In a simulation context, I'm applying some my function, "myfun" say, to a list of glm obj, "list.glm": >length(list.glm) #number of samples simulated [1] 1000 >class(list.glm[[324]]) #any component of the list [1] "glm" "lm" >length(list.glm[[290]]$y) #sample size [1] 1000 Because length(list.glm) and the sample size are rather large,

This is a question about removing the intercept in a binomial glm() model with categorical predictors. V&R (3rd Ed. Ch7) and Chambers & Hastie (1993) were very helpful but I wasn't sure I got all the answers. In a simplistic example suppose I want to explore how disability (3 levels, profound, severe, and mild) affects the dichotomized outcome. The glm1 model (see below) is

I'm sorry, maybe the question was bad posed. Ista has well described my problem. Thanks Massimo >----Messaggio originale---- >Da: izahn at psych.rochester.edu >Data: 28/07/2011 17.52 >A: "David Winsemius"<dwinsemius at comcast.net> >Cc: "m.fenati at libero.it"<m.fenati at libero.it>, <r-help at r-project.org> >Ogg: Re: [R]

Problems with predict and lines in plotting binomial glm Dear R-helpers I have found quite a lot of tips on how to work with glm through this mailing list, but still have a problem that I can't solve. I have got a data set of which the x-variable is count data and the y-variable is proportional data, and I want to know what the relationship between the variables are. The data was

Dear members of the R-help list, I have sent the email below to the R-SIG-ME list to ask for help in interpreting some R output of fitted linear models. Unfortunately, I haven't yet received any answers. As I am not sure if my email was sent successfully to the mailing list I am asking for help here: Dear members of the R-SIG-ME list, I am new to linear models and struggling with

This isn't a question about R, but I'm hoping someone will be willing to help. I've been looking at calibration plots in multiple regression (plotting observed response Y on the vertical axis versus predicted response [Y hat] on the horizontal axis). According to Frank Harrell's "Regression Modeling Strategies" book (pp. 61-63), when making such a plot on new data