Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: R: Imputation vs substitution with mean


From   James Bernard <jamesstatalist@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: R: Imputation vs substitution with mean
Date   Fri, 18 Oct 2013 16:04:41 +0800

Thanks Carlo,

Very helpful, indeed.

James

On Fri, Oct 18, 2013 at 3:12 PM, Carlo Lazzaro
<carlo.lazzaro@tiscalinet.it> wrote:
> Dear James,
> sample mean substitution is a very ill-advised approach (even if you have
> few missing values).
> In the best case, replacing missing values with the mean of the existing
> observation will reduce the variance, making all the subsequent statistics
> biased.
> Also last observation carried forward (LOCF) belongs to the "don't do it"
> group.
>
> You can find wide coverage of this topic in:
>
> - Roderick J. A. Little, Donald B. Rubin. Statistical Analysis with Missing
> Data, 2nd Edition. Wiley Series in Probability and Statistics, 2002.
> - John W. Graham. Missing Data: Analysis and Design. Springer, 2012.
>
> There are other textbooks on dealing with missing data; a Google search can
> give you a comprehensive picture of what has been published so far.
>
> Multiple imputation is the way to go; I would point you to -[MI] intro
> substantive - in Stata .pdf manual (as usual, a very good first step).
>
>
> I do hope this helps.
>
> Best regards,
> Carlo
>
>
>
> -----Messaggio originale-----
> Da: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu] Per conto di James Bernard
> Inviato: venerdě 18 ottobre 2013 07:38
> A: statalist@hsphsun2.harvard.edu
> Oggetto: st: Imputation vs substitution with mean
>
> Hi All,
>
> I have a longitudinal dataset with a dependent variable of count type.
> Mu independent variable are of continuous and categorical type.
>
> My independent variables have many missing values. I want to substitute the
> missing values.
>
> One way is to go through what I found out to be "imputation".
>
> I heard that an easier approach is substituting the values with group
> mean/population mean. Is this very different form imputation done by
> -mi-command in Stata?
>
> Thanks in advance,
>
> James
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index