similar to: Differences between SPSS and R on probit analysis

Not real data. It was gererated randomly. The original codes are the following: par(mfrow=c(2,1)) n <- 500 ######################### #DATA GENERATING PROCESS# ######################### x1 <- rnorm(n,0,1) x2 <- rchisq(n,df=3,ncp=0)-3 sigma <- 1 u1 <- rnorm(n,0,sigma) ylatent1 <-x1+x2+u1 y1 <- (ylatent1 >=0) # create the binary indicator ####################### #THE

Hi to all of you, I'm fitting an full factorial probit model from an experiment, and I've the independent variables as factors. The model is as follows: fit16<-glm(Sube ~ as.factor(CE)*as.factor(CEBO)*as.factor(Luz), family=binomial(link="probit"), data=experimento) but, when I took a look to the results I've obtained the following: glm(formula = Sube ~ CE * CEBO *

Dear all, I've got some questions, probably due to misunderstandings on my behalf, related to fitting overdispersed binomial data using glm(). 1. I can't seem to get the correct p-values from anova.glm() for the F-tests when supplying the dispersion argument and having fitted the model using family=quasibinomial. Actually the p-values for the F-tests seems identical to the p-values for

Dear all, I am new to R and my question may be trivial to you... I am doing a GLM with binomial errors to compare proportions of species in different categories of seed sizes (4 categories) between 2 sites. In the model summary the residual deviance is much higher than the degree of freedom (Residual deviance: 153.74 on 4 degrees of freedom) and even after correcting for overdispersion by

Classification: UNCLASSIFIED Caveats: NONE A similar question has been posted in the past but never answered. My question is this: for probit analysis, how do you program a 95% confidence interval for the LD50 (or LC50, ec50, etc.), including a heterogeneity factor as written about in "Probit Analysis" by Finney(1971)? The heterogeneity factor comes into play through the chi-squared

Hi! I was trying to implement a probit model on a dichotomous outcome variable and found that the predictions were outside the (0,1) interval that one should get. I later tried it with some simulated data with a similar result. Here is a toy program I wrote and I cant figure why I should be getting such odd predictions. x1<-rnorm(1000) x2<-rnorm(1000) x3<-rnorm(1000)

Just a quick point which may be easy to correct. Whilst typing the wrong thing into R 2.2.1, I noticed the following error messages, which seem to have some stray quotation marks and commas in the list of available families. Perhaps they have been corrected in the latest version (sorry, I don't want to upgrade yet, but it should be easy to check)? > glm(1 ~ 2,

I am running a binomial glm with response variable the no of mites of two species y->cbind(mitea,miteb) against two continuous variables (temperature and predatory mites) - see below. My model shows overdispersion as the residual deviance is 48.81 on 5 degrees of freedom. If I use quasibinomial to account for overdispersion the dispersion parameter estimate is 2501139, which seems

THANKS so very much for your help (previous and future!). I have a two follow-up questions. 1) You say that dispersion = 1 by definition ....dispersion changes from 1 to 13.5 when I go from binomial to quasibinomial....does this suggest that I should use the binomial? i.e., is the dispersion factor more important that the 2) Is there a cutoff for too much overdispersion - mine seems to be

I've originally made 48 GLM binomial models and compare the AIC values. But dispersion was very large: Example: Residual deviance: 8811.6 on 118 degrees of freedom I was suggested to do a quasibinomial afterwards but found that it did not help the dispersion factor of models and received a warning: Residual deviance: 3005.7 on 67 degrees of freedom AIC: NA Number of Fisher Scoring

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

Full_Name: Henric Nilsson Version: 2.3.0 Patched (2006-05-09 r38014) OS: Windows 2000 SP4 Submission from: (NULL) (83.253.9.137) When supplying an unavailable link to `quasibinomial', the error message looks strange. E.g. > quasibinomial("x") Error in quasibinomial("x") : 'x' link not available for quasibinomial family, available links are "logit",

thanks a lot Renaud. but i was interested in Finney's fiducial confidence intervals of LD50 so to obtain comparable results with SPSS. But your reply leads me to the next question: does anybody know what is the best method (asymptotic, bootstrap etc.) for calculating confidence intervals of LD50? i could "get rid" of Finney's fiducial confidence intervals but

Hello, all. I'm investigating the rate at which skeletal joint surfaces pass through a series of ordered stages (changes in morphology). Current statistical methods in this type of research use various logit or probit regression techniques (e.g., proportional odds logit/probit, forward/backward continuation ratio, or restricted/unrestricted cumulative probit). Data typically include the

I am working with Probit regression (I cannot switch to logit) can anybody help me in finding out how to obtain with R Finney's fiducial confidence intervals for the levels of the predictor (Dose) needed to produce a proportion of 50% of responses(LD50, ED50 etc.)? If the Pearson chi-square goodness-of-fit test is significant (by default), a heterogeneity factor should be used to calculate

Hello. I have yet to receive a response to my previous post, so I may have done a poor job asking the question. So, here is the general question: how can I run a run a multivariate (more than one non-independent, response variables) ordered probit regression model? I've had success doing this in the univariate case using the vglm() function in the VGAM package. For example:

Dear R users, Overdispersion is often a problem in binomial data. I attempt to model a binary response (sex-ratio) with three categorical explanatory variables, using GLM, which could assume the form: y<-cbind(sexf, sample-sexf) model<-glm(y ~ age+month+year, binomial) summary(model) Output: (Dispersion parameter for binomial family taken to be 1) Null deviance: 8956.7 on 582

 Hello! > I have a very simple set of data and I would like to analyze > them with probit analysis. > The data are: X    Event    Trial > 1210  8        8 > 121  6        8 > 60.5  6        8 > I want to estimate the value of X that will give a 95% hit > rate (Event/Trial) and the corresponding 95% CI. > you can help me? Thanks!! > Trinh  [[alternative HTML version

After looking at 48 glm binomial models I decided to try the quasibinomial with the top model 25 (lowest AIC). To try to account for overdispersion (residual deviance 2679.7/68 d.f.) After doing so the dispersion factor is the same for the quasibinomial and less sectors of the beach were significant by p-value. While the p-values in the binomial were more significant for each section of the

Hi - I am fitting a probit model using glm(), and the deviance and residual degrees of freedom are different depending on whether I use a binary response vector of length 80 or a two-column matrix response (10 rows) with the number of success and failures in each column. I would think that these would be just two different ways of specifying the same model, but this does not appear to be the case.