The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Ethiopia » Trend Analysis (Trend Analysis using Multilevel logistic regression )
Trend Analysis [message #26350] Sat, 11 March 2023 06:02 Go to next message
Jeremy is currently offline  Jeremy
Messages: 9
Registered: February 2022
Member
Hello Dear DHS experts and specialists,

My deepest greeting to you, I came back again, I want to do trend analysis on contraceptive use, obesity and smoking using the
EDHS 2000, 2005,2011, 2016 and 2019 datasets. I am planning to use the IR file for the first two studies. I tried to read the
different responses you gave on "trend" related issues here on the forum and I have got some useful points related to my study,
but I need more help on the following areas:

a) The different survey years have different cluster size, is it ok to do the trend analysis on
varying size or should I take the one with minimum size as a base.
b) Can I simply merge the files using one common variable say v001,
c) What basic advice do you give me to do the analysis.

Blessings!

Jeremy
Re: Trend Analysis [message #26355 is a reply to message #26350] Mon, 13 March 2023 11:18 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Following is a response from Senior DHS staff member, Tom Pullum:

Successive DHS surveys in the same country are completely independent of each other. The clusters are census enumeration areas, sampled from a huge list, and the ones sampled will be different, just as the households and individuals are different. They are numbered 1, 2, 3, etc. for convenience, but cluster 1 in one survey is not the same as cluster 1 in another survey--in fact, cluster 1 in one survey will not be the same as ANY cluster in ANY of the other surveys.

When combining surveys into a single large file for a trend analysis, use "append", not "merge". Hope this is clear. Good luck with your research.
Re: Trend Analysis [message #26368 is a reply to message #26355] Tue, 14 March 2023 04:41 Go to previous messageGo to next message
Jeremy is currently offline  Jeremy
Messages: 9
Registered: February 2022
Member
Ok thank you so much Sir, I really appreciate your kindly support.
Have a Blessed day, Stay Safe!

Blessings!

Jeremy
Re: Trend Analysis [message #26454 is a reply to message #26355] Wed, 22 March 2023 08:53 Go to previous messageGo to next message
Jeremy is currently offline  Jeremy
Messages: 9
Registered: February 2022
Member
Hello Dear Tom Pullum, I am back again with few questions. For my analysis I am using SPSS and R.
I am new to Stata, and it is not working on my PC. As you advised me I tried to append the files
on SPSS but I could not do it, SPSS has merge option and you told me to append not to merge.
My basic question is that, for two of my trend studies on contraceptive use and overweight/obesity,
I want to do multilevel logistic regression analysis using IR file of the 2000,2005,2011, 2016 and
2019 EDHS. I want to append all the survey files into one so that I can do the analysis.
Here are my questions:
1) How can I append the files using SPSS?,
2) Can I do multilevel logistic regression analysis after appending or do I need further data screening?,
3) What advice can you give me on the weighting issue on the multilevel logistic regression analysis?,

Thank you so much. Have a great day. Stay Safe!

Blessings!

Jeremy
Re: Trend Analysis [message #26457 is a reply to message #26454] Wed, 22 March 2023 11:12 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

We (the DHS staff) do not use SPSS. You can find many posts on the forum that describe how to append and how to adjust the weights in pooled data using Stata. The commands would be different in SPSS but the concepts are the same.

I just went to google and entered "how to append files in SPSS". The command came up immediately. You should be able to find tutorials on how to use weights in SPSS. Sorry that we cannot be more specific. Perhaps other SPSS users can add something.
Re: Trend Analysis [message #26471 is a reply to message #26457] Fri, 24 March 2023 00:59 Go to previous message
Jeremy is currently offline  Jeremy
Messages: 9
Registered: February 2022
Member
Ok, thank you so much Dear Tom Pullum.

Blessings!

Jeremy
Previous Topic: Problem with dates in the Ethiopia datasets
Next Topic: Selecting appropriate weights when IR and MR files are pooled
Goto Forum:
  


Current Time: Thu Mar 28 15:55:16 Coordinated Universal Time 2024