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: Pooled fixed effect model
From
Alfonso Sanchez-Penalver <[email protected]>
To
"[email protected]" <[email protected]>
Subject
Re: st: Pooled fixed effect model
Date
Sun, 5 Jan 2014 06:29:50 -0500
Hi Lukas & Joerg,
Let me see if I understand your problem well. You have two datasets that you want to pool. And each dataset would be a different year. The problem that you have is that you cannot identify what regions are the same in the different datasets and want to keep each region as a new cluster in the pooled data.
My first problem with this is that if you do that there is no guarantee that the estimates of the parameters are consistent, or that they make any sense at all, since there is no guarantee that the sample cluster average is the same for those regions that are present in the different years. The second problem I have is that I don't see the benefit of pooling the data in this case because you can't account for the correlation of the residuals across time of those observations that belong to a region that is present in the different years. Thus if each year's dataset is large there would be very little gain of estimates' efficiency by pooling the data.
Please let me know if I have misunderstood the problem you're facing.
Best,
Alfonso Sanchez-Penalver
> On Jan 5, 2014, at 5:13 AM, Jörg Kattner <[email protected]> wrote:
>
> Dear Statalist members,
>
> We have question regarding a fixed effect model to account for regional/com=
> munity effects in our regression.
>
> We would like to run the following regression:
>
> Svy: clogit y =3D x1 x2 x3 group(v001)
>
> V001 are the cluster units in one DHS survey.
>
> Now we would like to pool datasets from different years for the same country. Unfortunately, the number of clusters differs between datasets and it cannot be seen which clusters are overlapping.
>
> Let's imagine that dataset 1 has 100 clusters and dataset 2 has 200 clusters. Would it be possible to pretend to have 300 different clusters for the pooled fixed effect model? The studies may be 15 years apart, so that the characteristics of a cluster might have changed over time.
>
> Any help would be much appreciated!
>
> Best regards,
>
> Lukas & Joerg
>
>
>
>
> *
> * 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/