| Descriptives and chisq [message #8881] |
Mon, 11 January 2016 09:53  |
cbdolan
Messages: 17
Registered: March 2013
Location: Williamsburg, VA
|
Member |
|
|
For this analysis, I'm using the 2007 and 2013/14 DRC.
In order to provide unbiased estimates, sampling weights based on the probabilities of selection from the underlying population and non-responses were taken into account using the following code and then applying it through the svy: command
gen wgt=v005/1000000
egen stratum=group (v024 v025)
svyset [pw=wgt], psu(v021) strata(stratum)
Now I am creating my descriptive table. The DHS weighting videos recommend this syntax:
tab female treat [iweight=wgt]
However, this doesn't allow for a chisq option. Is the following syntax a correct alternative? This gives me proportions as well as the associated chisq:
svy: tab female treat
Thanks.
Carrie
|
|
|
|
| Re: Descriptives and chisq [message #8883 is a reply to message #8881] |
Mon, 11 January 2016 13:31  |
Reduced-For(u)m
Messages: 292
Registered: March 2013
|
Senior Member |
|
|
Yes, you can use the "svy" prefix instead of "manually" weighting to tabulation. The "svy" gets you the weighting and also inference (p-values/standard errors/confidence intervals) adjustments for stratification and clustered sampling.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
