Recent content by noetsi

"For virtually every logistic regression model that we estimate in the real world, there will be some uncorrelated covariates that are statistically associated with the binary outcome, but that we couldn’t observe to include in the model. In other words, there’s always unobserved heterogeneity...

thanks a lot jake. Although now I have two statistical experts (you and Allison) disagreeing entirely. :p I never know quite what to do about that - since I am not an expert :) We won't be running ordinal or multinomial logistic regression so we will never have more than two levels of the DV. I...

These are all binomial response variables. What hlsmith called rate my books and links call odds, but I assume they are the same.

I am pretty sure when I researched this last year I found the diagnostics don't exist in PROC GLM. That they generate the same results is great to know. Thanks. I have never heard of glmselect. I will have to look this up and see if we have that PROC.

Only the last variable is statistically significant (_cons). Its basically impossible to interpret substantively slopes in logistic regression other than their sign and if they are significant. They refer to a change in the logit for a one unit change in X controlling for other predictors. Not...

Thanks. I got over some of my physical problems which made it impossible to post until recently. A reference, from a book I think you have, is Allison's book "Logistic Regression Using SAS." 2nd ed p 98. Some authors raise more concern than he does, I list him because I think you are familiar...

How do you test for these in logistic regression. I don't know of any test if they are occurring or not.

Multicolinearity is similar in linear or logistic regression. However, many feel (for example John Fox) that its impact is limited in most cases even if it exists and few of the solutions offered for it work well, They may lead to issues such as biased parameters that are worse than...

I can't get access to that article spunky. But in any case is explained variance the same as the impact of an x on Y (relative to another x)? I would think this would be tied to change in Y given an X (standardized slope) not explained variance. But maybe they are the same thing....

I am frequently asked to rank the relative impact of predictors on a dependent variable based on the results of a regression. That is, does this variable influence (cause change in) the dependent variable more than another predictor. For linear regression I judge relative impact by standardized...

PROC GLM is able to do more with categorical predictor variables than PROC REG (which lacks a class statement). But many assumption diagnostics exist in PROC REG that do not as best I can tell in PROC GLM. So my question is can you specify a model in PROC GLM with dummy variables and then...

At the bottom of the link below the author says (without details) An alternative method for interpreting Cook's distance that is sometimes used is to relate the measure to the F(p, n–p) distribution and to find the corresponding percentile value. If this percentile is less than about 10 or 20...

Here is a more general question. Some authors in forecasting argue that you should average all the ESM models. Some say that you should use the best, with best being determined through MAPE against a holdout data set (the last 12 months are predicted, you pick the mode that best predicted them)...

It turns out PROC AUTOREG does do this after all (and this deals with hetero which is why they are called Newey West HAC sometimes PROC AUTOREG .... Model..../COVEST =NeweyWest;

Thanks for your comments ondansetron. He believes that unless there is a theoretical reason for something to be true you should not model it. My guess is that someone made this point to him in the context of methods like regression (where many argue a variable should only be in a model if it...