similar to: Creating data for logistic regression and Cox proportional hazards regression

I am trying to run a Cox regression (coxph) to predict survival after surgery and xray therapy. The model violates the proportionality assumption (tested using time-dependent covariates). I have tried to use strata to solve the problem but the stratified model still violates the assumption. Does the violation of proportionality preclude use of a Kaplan-Meier analysis? Is there any other possible

Given a square matrix of variables X, is there any way to compute a multivariable correlation among all the variables in the array? It is possible to calculate the correlation of all pairs of variables in the array, but I want to know the correlation of all the variables taken together, i.e. for the matrix X=x1 x2 x3 x4 x5 x6 x7 x8 x9 I don't want the pair-wise correlations

I hope one and all will allow a stats question: When running a cox proportional hazards model ,there are two ways to deal with age, including age as a covariate, or to include age as part of the follow-up time, viz, Age as a covariate: tetest1 <- list(time= c(4, 3,1,1,2,2,3), status=c(1,NA,1,0,1,1,0), age= c(0, 2,1,1,1,0,0),

R 2.2.0 windows XP How can I perform a regression analyses using a vector of means, a variance-covariance matrix? I looked at the help screen for lm and did not see any option for using the afore mentioned structures as input to lm. Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and Informatics Baltimore VA Medical Center GRECC, University of Maryland School of Medicine Claude D.

Windows Vista R 2.9.1 When trying to install Rcmdr I get the message shown below. What does it mean, do I need to do anything? The downloaded packages are in C:\Users\John Sorkin\AppData\Local\Temp\RtmpRCLbhe\downloaded_packages updating HTML package descriptions Warning message: In file.create(f.tg) : cannot create file 'C:\PROGRA~1\R\R-29~1.1/doc/html/packages.html', reason

R 2.7.2 Windows XP I am using apply to compute a series of Student's t-test from two matrices, sample1 and sample2. boo<-apply(sample1,1,t.test,sample2) I want to pick of the p-values from the tests, but can't seem to get it to work. I have tried several methods to get the values including: boo<-apply(sample1,1,t.test$t.test,sample2) boo<-apply(sample1,1,t.test,sample2)$t.test

R 2.3.1 Windows XP Question: How can I obtain the values of BIAS and STD. ERROR from a bootstrap. Background: I am running a bootstrap: result2<-boot(1:400,regSEvssample,R=5000) and obtain the following results: Bootstrap Statistics : original bias std. error t1* 1.876602 -0.0001368616 0.1630380 I would like to get the values of ORIGINAL and BIAS. I can get the value of

Is there an R function that can be used to calculate a sample size for a linear regression: i.e. Given a linear regression: y=f(x,z) i.e. fit1<-lm(y~x+z) is there a function which can be passed coef(z), SE(z), and other parameters to determine a sample size based on z, SEz and other information from the regression? Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and Informatics

At the suggestion of many people, I have installed emacs on my linux (Fedora 8.0) computer with the intention of using emacs as window interface to R (2.6.0). I have gone though the emacs tutorial and don't see any information about how I should use emacs to run R. Can anyone suggest a document that I might read? In the past I have used R on a Windows XP system and used the built-in windowing

Fedora 10 R 2.8.1 I hope someone can tell me the meaning of error I received trying to install RODBC and how I can get around the problem. n.b. The error originally occurred then I was installing Rcmdr. I then tried to install RODBC separately and received the same error. checking sql.h usability... no checking sql.h presence... no checking for sql.h... no checking sqlext.h usability... no

I am trying to model data in which subjects are followed through time to determine if they fall, or do not fall. Some of the subjects fall once, some fall several times. Follow-up time varies from subject to subject. I know how to model time to the first fall (e.g. Cox Proportional Hazards, Kaplan-Meir analyses, etc.) but I am not sure how I can model the data if I include the data for those

Can anyone suggest a good non-parametric test, and an R implementation of the test, that allows for two or more factors? Wilcoxon signed rank allows for only one. Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and Informatics Baltimore VA Medical Center GRECC and University of Maryland School of Medicine Claude Pepper OAIC University of Maryland School of Medicine Division of

I am having trouble selecting rows of a dataframe that will be included in a regression. I am trying to select those rows for which the variable Meno equals PRE. I have used the code below: difffitPre<-lm(data[,"diff"]~data[,"Age"]+data[,"Race"],data=data[data[,"Meno"]=="PRE",]) summary(difffitPre) The output from the summary indicates that

Fellow Rers, Please forgive me if I have posted this to the wrong R list serve. Over the course of the years that I have used R and participated in this list server, I have noted a large number of questions and answers that direct people to specific packages. The multitude of packages is one of the great strengths of R. Unfortunately there is no (or at least I am not aware of) any single source

Can someone tell me why apply (and apply) are faster in performing repeated operations than a for (or do) loop? I am looking for a technical explanation. Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and Informatics Baltimore VA Medical Center GRECC and University of Maryland School of Medicine Claude Pepper OAIC University of Maryland School of Medicine Division of Gerontology

R 2.1.1 Win 2k Would someone suggest a method (or methods) that can be used to determine ntile cutpoints of a vector, i.e. to determine values that can be used to divide a vector into thirds such as 0-33 centile, 34-66 centile, 67-100 centile. If for example I had a vector: 1,2,3,4,5,6,7,8,9 and wanted to divide the vector into thirds I would have cut-points of 3, and 6. Thanks, John John

R 2.2.0 windows XP I am beginning to explore the mitools package contributed by Thomas Lumley (thank you Thomas) and I have a few questions: (1) In the examples given in the mitools documentation, the only family argument used is family=binomial. Does the package support family=gaussian and other link functions? I ran the with function with family=gaussian and I obtained results, but I am not

R 2.8.1 windows XP I don't understand the output from x[iS+1:iE] produced by the code below: x = c(1,2,3,4,5) x [1] 1 2 3 4 5 iS=2 # start position iE=4 # end position [iS:iE] [1] 2 3 4 # I don't understand the results of the command below. I would expect to see 3, 4, not 3, 4, 5, NA x[iS+1:iE] [1] 3 4 5 NA Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and

Peter You question is difficult to answer without more information about the distribution of your residuals. Different residual patterns call for different transformations to stabilize the variance. One very common form of heterocedasticity is increasing variance with increasing values of an independent predictor, i.e. the variance of the residuals of y=x increase as x increases. In this case a

Is there any way to load a package using R code rather than the Load Package command in the Package menu? I would like to load the boot package. R.2.1.0 Patched Win 2k. Thanks, John John Sorkin M.D., Ph.D. Chief, Biostatistics and Informatics Baltimore VA Medical Center GRECC and University of Maryland School of Medicine Claude Pepper OAIC University of Maryland School of Medicine Division of