RE: st: RE: SIPP survey in stata - calculations

[Jimmy Jin <Jimmy.Jin@cbo.gov>]

Hi Austin,

Thanks for your correction. I meant: to apply the probability weight as an analytic weight in summarize/tabulate to obtain a count, with the resulting estimate being rough--extreme emphasis on "rough." DEFINITELY not kosher, but a jury rig nevertheless.

At any rate, I agree that the actual variables and methods involved are probably not appropriate for Statalist. As for the CBO website, you probably won't find much there about conducting the analysis in Stata either. You might find the analysis itself, though. Worth a look, for sure.

-Jimmy

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Austin Nichols
Sent: Tuesday, February 15, 2011 11:23 AM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: RE: SIPP survey in stata - calculations

Jimmy Jin <Jimmy.Jin@cbo.gov>
The proper weight variable to use depends on the use to which you are putting the data, but you cannot properly "apply the weight as a probability weight using summarize or tabulate" since those commands do not accept pweights; you must use -svy- commands, e.g.
svyset ssuid [pw=wpfinwgt]
svy: tab epensnyn

Much of the detail on the survey is on its website
(http://www.census.gov/sipp/) including a Users' Guide. If you do ssc inst ddf2dct help ddf2dct you will see links to the NBER site you refer to; however the most recent versions of the files and the fastest download times are always from BLS (http://www.bls.census.gov/sipp_ftp.html).

However, it's not clear that the original poster who wants to "calculate the nr of people and the percentage of working age who are employed and are included in a retirement plan offered on their job and not through the social security" can do so using the SIPP.  Is the idea to find the number of uncovered workers (mostly state gov't
employees) who have other pensions?  This kind of question is probably outside the scope of Statalist, though I am guessing there is something apropos on the CBO website...

On Tue, Feb 15, 2011 at 11:02 AM, Jimmy Jin <Jimmy.Jin@cbo.gov> wrote:
> Tharshini,
>
> First of all, I recommend getting SIPP data from the NBER SIPP repository, which has Stata do/dct files ready-made so you don't accidentally mess up the first step of importing the data:
>
> http://www.nber.org/data/sipp.html
>
> Second of all, you do need to use weights. The weight variable at the person level is WPFINWGT. You can apply the weight as a probability weight using summarize or tabulate, or you can sum the weight variable itself over the group you're interested in to obtain a rough estimate.
>
> Best,
> Jimmy
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Tharshini 
> Thangavelu
> Sent: Sunday, February 13, 2011 2:30 PM
> To: statalist@hsphsun2.harvard.edu
> Subject: st: SIPP survey in stata - calculations
>
> Hi statalist,
>
> Have someone worked with the SIPP = U.S Survey of Income and Program 
> Participation, Topical Module wave 3 year 2008
>
> I have few questions concerning the data. When you work with the SIPP, do I have to take into account the weighting procedure when I want to calculate indicators.
>
> Ex: I want to calculate the nr of people and the percentage of working age who are employed and are included in a retirement plan offered on their job and not through the social security.
>
> If anyone who can provide me with their knowledge to understand the SIPP data set, I would be more than happy.
>
> Thanks in advance !

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