I have just performed my first multiply imputed multiple regression analyses (using Schafer's freestanding version of NORM for Windows), and the results have brought up a question for me that I'm hoping listmembers will have some thoughts about.
The multiple regression analyses all involved the same set of 6 predictors with a number of different dependent variables, using a dataset with a sample size of 613 cases. I conducted each regression model five times using the five imputed datasets that I generated with NORM. I should note that most of the missing data in these analyses were in the dependent variables and not the predictors. For these variables, data were available from only 569 to 578 participants. What was most surprising to me was the huge variability in the degrees of freedom generated by the analyses. For example, age was one of the predictors, and df associated with this predictor varied from 21 to 7012 for different dependent variables. The "missing information" statistic for age was similarly variable. Neither the df nor the missing information statistic seemed to correspond to the actual percentage of missing values in the predictor or the DV. I'd be grateful if folks on this list could help me interpret such results. For example, what does it mean that the missing information statistic can vary so widely for a predictor when the actual % of missing values is constant among DVs? Thanks in advance for your thoughts on what is probably a very basic question! Best Jon __________________________________ Jonathan Mohr, Ph.D. Assistant Professor Department of Psychology Loyola College 4501 North Charles Street Baltimore, MD 21210-2699 E-mail: [email protected] Phone: 410-617-2452 Fax: 410-617-5341 __________________________________ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.utsouthwestern.edu/pipermail/impute/attachments/20031211/a7f7b7db/attachment.htm From jmohr <@t> loyola.edu Wed Dec 17 10:32:10 2003 From: jmohr <@t> loyola.edu (Jonathan Mohr) Date: Sun Jun 26 08:25:01 2005 Subject: IMPUTE: more basic questions Message-ID: <[email protected]> I'm still immersed in my first multiple imputation analyses, and a couple more questions have arisen for me: 1. Say that one of my goals is to estimate means/sds of variables with missing data by gender (along with overall means/sds of those variables). I can think of a few approaches to conducting the multiple imputation: (a) in addition to the variables of interest, include gender in the imputation model. (b) in addition to the variables of interest, include gender and the interactions of gender with those variables in the imputation model. (c) conduct separate imputation analyses by gender, then recombine the imputed women's and men's dataset. Any opinions as to which strategy is best? 2. I am interested in conducting a multiple regression analysis with interaction terms, using multiply imputed datasets. I understand that I need to include these interaction terms in the imputation model (along with the "main effect" variables). What isn't clear to me is which of the following two "versions" of the interaction term xz I should use: (a) the imputed interaction terms (i.e., estimates of the missing xz values generated by the MCMC imputation method) (b) the interaction terms computed by taking the product of the imputed x values and the imputed z values. Any thoughts about which might be the preferred strategy? Thanks in advance for your thoughts! Best, Jon __________________________________ Jonathan Mohr, Ph.D. Assistant Professor Department of Psychology Loyola College 4501 North Charles Street Baltimore, MD 21210-2699 E-mail: [email protected] Phone: 410-617-2452 Fax: 410-617-5341 __________________________________ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.utsouthwestern.edu/pipermail/impute/attachments/20031217/09c04cf9/attachment.htm From Howells_W <@t> bmc.wustl.edu Thu Dec 18 17:06:44 2003 From: Howells_W <@t> bmc.wustl.edu ([email protected]) Date: Sun Jun 26 08:25:01 2005 Subject: IMPUTE: Is there a list archive for Impute? Message-ID: <of6447043e.ebd81817-on86256e00.007edcda-86256e00.007ef...@wustl.edu> New to list and don't want to duplicate questions already asked and answered. Bill H.
