similar to: calculate response probabilities using sem-analysis

Dear list members, In sem, std.coef() will give me standardized coefficients from a sem model. But is there a trick so that path.diagram can use these coefficients rather than unstandardized ones? Thanks Steve Powell From: John Fox <jfox_at_mcmaster.ca> Date: Wed 28 Feb 2007 - 14:37:22 GMT Dear Tim, See ?standardized.coefficients (after loading the sem package). Regards, John John

Hello - Does anybody know how to get the SEM package in R to return standardized path coefficients instead of unstandardized ones? Does this involve changing the covariance matrix, or is there an argument in the SEM itself that can be changed? Thank you, Tim [[alternative HTML version deleted]]

I'm not aware of any routine that those the job, although I think that it could be relatively easily done by multiplication the manifest variable vector with the estimates for the specific effect. To make an example: v1; v2; v3; v4 are manifest variables that loads on one y latent variablein a data frame called "A" the code for the model should be like: model <-specifymodel( y

Hello, I will use the lca method in the e1071 package. But I get the following error: Error in pas[j, ] <- drop(exp(rep(1, nvar) %*% log(mp))) : number of items to replace is not a multiple of replacement length Does anybody know this error and knows what this means? Kind regards, Tryntsje

Hi, I'm using the SEM package to estimate a model with a binary variabele as dependent variable. In the literature I have to use then the correlation matrix, made by function hetcor(). Literature also says that the standard errors are not correct then. My question is if somebody knows what the impact is on the estimated coefficients. If I want to calculate the estimated probability I see a

Hi, I'm generating the name of the variable with paste function and then using that variable name further to get the specific position value from the data.frame, here is the snippet from my code: modelResults <- extractModelParameters("C:/PilotStudy/Mplus_Input/Test", recursive=TRUE) #extractModelParameters reads all the output files from the Test folder and create the

hi, I plot path diagrams with the path.diagram() function of the sem package in combination with the graphviz application. Now I want the graphviz code for a path-plot with the actual standardized coefficients on the arrows (not the names). I tried to add edge.labels="values" as an argument to path.diagram() but it's just ignored. can anyone help me on that? p.s.;

Hello. I am getting results from sem that are not correct (that's assuming that the results from my AMOS 4.0 software are correct). sem does not vary some of the parameters substantially from their starting values, and the final estimates of those parameters as well as the model chisquare value are incorrect. I've attached some code that replicates the problem. The parameters in

Hi there, as a student sociology, I'm starting to learn about SEM. The course I follow is based on LISREL, but I want to use the SEM-package on R parallel to it. Using LISREL, I found it to be very usable to be able to see the total direct and total indirect effects (standardized and unstandardized) in the output. Can I create these effects using R? I know how to calculate them

Dear useRs, I have written a function that implements a Bayesian method to compare a patient's score on two tasks with that of a small control group, as described in Crawford, J. and Garthwaite, P. (2007). Comparison of a single case to a control or normative sample in neuropsychology: Development of a bayesian approach. Cognitive Neuropsychology, 24(4):343?372. The function (see

I would like to obtain Confidence Intervals for the estimates (unstandardized beta weights) of each predictor in a WLS regression: m1 = lm(x~ x1+x2+x3, weights=W, data=D) SPSS offers that output by default, and I am not able to find a way to do this in R. I read through predict.lm, but I do not find a way to get the CIs for multiple independent variables. Thank you Torvon [[alternative HTML

Hi all, is it possible to run factor analysis in R such that the routine returns - unstandardized factor scores (according to the original scale) - rotated factor scores (these may be standardized) So far I have only found the possibility to return standardised unrotated factor scores. Thank you very much, Ursula ==================================================== NFO Infratest Ursula

I attempted to contact Drew Tyre, but the email I have for him is no longer in service. Hopefully someone can help. I'm using obs.error in R to model turtle occupancy in wetlands. I have 4 species and 20 possible patch and landscape variables, which I've been testing in smaller groups. > zib.out<-obs.error(y=painted,m=numvis,bp=zvars,pcovar=7) I get the following error

I run 9 WLS regressions in R, with 7 predictors each. What I want to do now is compare: (1) The strength of predictors within each model (assuming all predictors are significant). That is, I want to say whether x1 is stronger than x2, and also say whether it is significantly stronger. I compare strength by simply comparing standardized beta weights, correct? How do I compare if one predictor is

A short function I wrote for the purpose of evaluating scales made up of a number of questionnaire items. It provides Cronbach's Alpha, both unstandardized and based on standardized items. It also provides item statistics which include item-total correlations (corrected) and item-removed alpha. Thought some of you might find it useful. I would also appreciate any programming tips or

Sorry if this question has been asked previously, I searched but found little. There also doesn't seem to be a dedicated SEM list-serv so hopefully this will find its way to the appropriate audience. In discussing SEM with a colleague I mentioned that a model they were fitting in AMOS was equivalent to a linear regression and that the coefficients would be the same. This of course was the

Is there a "canned" function in R for finding the standard error of the mean? I have tried > sem <- function(x) c(mean =mean(x), + SEM = stdev(x)/sqrt(length(x))) > sem(pnet.lai) Error in sem(pnet.lai) : couldn't find function "stdev" It looks like there is no stdev function in R Thanks, Kirk Kirk R. Wythers email: kwythers at umn.edu University of

Hi John, Thanks for the comment... but that appears to mean that SPSS has a big problem. I have always been told that to include an interaction term in a regression, the only way is to do the multiplication by hand. But then it seems to be impossible to stop SPSS from re-standardizing the variable that corresponds to the interaction term. Am I missing something? Is there a way to perform the

Hi all, I know pnorm() is able to calculate a sequence of probabilities by providing a vector of means, and a single number for location and standard deviation. For example, pnorm(0.5,c(0.5,2),1) will produce [1] 0.5000000 0.0668072. However, I wonder if its multivariate counterpart mvtnorm can do the same thing? Namely, if I provide a sequence of (vector-)means and a single vector for

Hi there, Quick question about the output from the sem() function in the library of the same name. If I am getting probabilities >0.05 for some of my estimates of path coefficients, I'm assuming the interpretation here is that the coefficient is not significantly different from zero, correct? In that case, might it make sense that I should disregard path coefficients between