Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: interviewereffect-multilevelmodeling


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: interviewereffect-multilevelmodeling
Date   Wed, 24 Oct 2012 06:09:17 -0500

How are interviewers related to the sample design/clusters? Are they
nested into clusters (fewer clusters than interviewers), are clusters
nested into interviewers (fewer interviewers than clusters), or are
they crossed (interviewers worked at different sites)? 30 clusters,
let alone 17 clusters, is not a great sample size, neither for
design-based inference for logistic regression (via -svy: logit-), and
even less so for multilevel model (-xtmelogit-) where you would have
to work with the actual 17 locations, rather than nominal 17 PSUWRs.

-- 
-- Stas Kolenikov, PhD, PStat (SSC)  ::  http://stas.kolenikov.name
-- Senior Survey Statistician, Abt SRBI  ::  work email kolenikovs at
srbi dot com
-- Opinions stated in this email are mine only, and do not reflect the
position of my employer


On Wed, Oct 24, 2012 at 5:47 AM, Srinivas <geevasan@yahoo.com> wrote:
> Dear all
> I am fairly new to STATA and multilevel modeling.
> I recently conducted a survey in India on elder abuse. The sample size is 902 (31 respondents from 30 clusters) elderly persons, one elderly person from one household.
>
> I employed a two stage sampling design. The initial number of villages ([PSUs], lowest rural administrative unit) was 30. But probability proportional to sampling size yielded only 17 villages i.e. villages with larger population size were considered as more than one site.
>
> The dependent variable is elder mistreatment (presence of abuse-1)
> I intend to find out risk factors of elder mistreatment using logistic regression
> My questions are as follows
> 1.      I understand that I could control for the cluster effects using Svyset command
> 2.      I also need to control for interviewer effects. Could someone suggest as to how I should do it in STATA 12.
> 3.      If I were to consider going for multilevel analysis (cluster level data) Should I consider the PSUs as only 17 villages or should I consider it as 30 clusters?
>
>
> I would greatly appreciate if you could clarify my doubts.
> Thanking you in advance for your time.
>
> Sincerely
> Srinivasan Chokkanathan
> NUS
> *
> *   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index