Weighting Data without SPSS/Stata [message #4358] |
Tue, 12 May 2015 15:19 ![Go to next message Go to next message](/theme/default/images/down.png) |
jd.weaver
Messages: 5 Registered: April 2015 Location: Seattle, WA
|
Member |
|
|
Hi All,
I'm working with multi-year DHS data from 9 countries to look at trends in modern contraception usage. Because of the size of the data I am importing it into a relational database (SQL Server) for ease of storage and access (people other than me need to access this data).
My question is what is the best way to add the sample weights (V005) to the data without using SPSS/Stata?
Alternatively, is there a way to apply sample weights in SPSS/Stata and then save the dataset as a flat file (e.g. XLS) with appropriate weights still applied to the data?
thanks very much!
|
|
|
Re: Weighting Data without SPSS/Stata [message #4363 is a reply to message #4358] |
Wed, 13 May 2015 13:19 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Trevor-DHS
Messages: 802 Registered: January 2013
|
Senior Member |
|
|
The weighting needs to be applied when you do your analysis, and how you store the data is not affected - just make sure you include v005 as one of the variables in your database. When analyzing the data you would use the weight, that is, whenever you count cases for numerators or denominators of indicators, instead of counting each case once, you count each case as being the value of the weight (divided by 1000000). For example, if the weight in v005 is 875000, then count the case as 0.875 people. You will need to ensure that you always use the weights in this way.
There is a much less efficient way of weighting the data, that avoids having to apply the weights in the analysis. This requires creating multiple copies of each record, and if the weight in v005 was 875000, then you would have to create 875000 copies of that particular record. I don't think this is a realistic approach, but i thought I would mention it.
Let me know if any of this is not clear.
|
|
|
|