The DHS Program User Forum: General » Multiple Sexual Partner in Husband

Home » Topics » General » Multiple Sexual Partner in Husband

Show: Today's Messages :: Show Polls :: Message Navigator

Multiple Sexual Partner in Husband [message #24218]

Fri, 18 March 2022 07:00

Nike2021
Messages: 8
Registered: October 2021

Member

I am using the variable v836 "Total lifetime number of sexual partners" of women as one of the risk factors for cervical cancer. I also want to include history of multiple sexual partners in the woman's husband. I saw 'mv836' in the men's recode. I am using R statitical software for my analysis. And I am looking at the survey of 8 West African countries 2017-2020 (Benin 2017-2018, Cameroon 2018, Gambia 2019/2020, Guinea 2018, Liberia 2019/2020, Mali 2018, Nigeria 2018, and Sierra Leone 2019).
I merged the women and men dataset by:
1. selecting the need variables from the two dataset
2. Adding a matching variable to both data frame
3. Merge the men and women data
4. Specify the survey design

I noticed that the total from the merged data is lower than the original total of the women data. E.g. in Benin the total number of women surveyed is 15928; after I merged it with the men data it became 10294. It means some of the records cannot be merged. What do you suggest please?

[Updated on: Tue, 22 March 2022 05:43]

Report message to a moderator

Re: Multiple Sexual Partner in Husband [message #24612 is a reply to message #24218]

Tue, 07 June 2022 20:02

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

Following is a response from Senior Demographic and Health Research Analyst, Jeffrey Edmeades:

The men's datasets (indicated by 'MR' in the dataset name e.g. SLMR7AFL.dta) do include a variable that is the equivalent of v836. This variable is mv836 in the men's data and is available for all the surveys you are using for your analysis. However, you cannot only use either the women's (IR) or men's (MR) data to link data from husbands to their wives to do this, you will need to use the Couple's Data (indicated by 'CR' in the dataset name e.g. BJCR71FL.dta). These data files include data for men and women in each household that each declared that they were married/living together to each other and links their data. The unit of analysis in the couple file is the couple, rather than an individual woman or man. This means that you will need to merge the women's file (IR) with the couple file (CR) (see guidance here). In addition the information for merging, the only information you will need from the CR file is mv836. Please keep in mind that in many surveys, data were collected from a smaller group of men than women, meaning that your sample size will be smaller after combining the women's and couple's data.

Report message to a moderator

Re: Multiple Sexual Partner in Husband [message #24641 is a reply to message #24612]

Mon, 13 June 2022 05:37

Nike2021
Messages: 8
Registered: October 2021

Member

Dear Jeffrey Edmeades,
Thanks for your detailed response. The couples file does not include data for formerly in union/living with a woman. In R statistical software, I am able to concatenate mv836 with v502. Will it then be appropriate to link mv836 to v502 in the country to get the multiple partners for the different marital status of the woman?
Kind regards,
Nike Olajide

Report message to a moderator

Re: Multiple Sexual Partner in Husband [message #24652 is a reply to message #24641]

Tue, 14 June 2022 16:49

Janet-DHS
Messages: 888
Registered: April 2022

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

If the term "concatenate" refers to constructing a new variable for different combinations of an mv variable and v variable, yes, you can do that. However, I don't see what that will give you. In the CR file, all women have v502=1 and all men have mv502=1. That is, both partners are currently married/cohabiting. Concatenation is more useful for something like the combinations of v106 and mv106 (education of the partners).

Report message to a moderator

Re: Multiple Sexual Partner in Husband [message #24660 is a reply to message #24652]

Thu, 16 June 2022 05:01

Nike2021
Messages: 8
Registered: October 2021

Member

Dear Tom Pullum,
Thanks for your prompt response. Yes, I linked mv836 and v502 because I am interested in the >5 sexual partner of the husband/partner of women who are in a union/ formaerly in a union. I need to group them as exposed and those <5 or never in a union as unexposed for each of the countries in the analysis. I want the total to be the total of the women 15-49 years for each country for further analysis.
Kind regards,
Nike

Report message to a moderator

Re: Multiple Sexual Partner in Husband [message #24686 is a reply to message #24660]

Tue, 21 June 2022 16:39

Janet-DHS
Messages: 888
Registered: April 2022

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

You want to analyze a v variable for women together with an mv variable for men. Are you trying to pair women with men? The couples file consists of pairs of women and men who identify each other as current partners. We do not have a way to construct any other pairings of women and men. Former partners are not likely even to be in the sample, because they will probably not be in the same household. Or are you trying to construct a cross-tabulation of women and men who are NOT linked as pairs? I can't visualize such a table.

Report message to a moderator