Table 1.

Measurement of variables

Intimate Partner Violence (IPV)	Generated from yes/no questions suggesting emotional, physical, and sexual violence. A composite factor analysis was performed on the three forms of violence and coded as experienced, and not experienced.
Contraceptive discontinuation	Defined as the interruption of contraceptive use for one month or longer by women who had used a contraceptive method in the past 12 months but discontinued at least once without switching to another method. It was coded as not discontinued, and discontinued.
Discontinuation while still at risk	It was defined as discontinuing for reasons other than wanting to become pregnant or no longer at risk of becoming pregnant (e.g. health concerns/side effects, method inconvenience, wanted a more effective method, cost, lack of access, or husband opposition). It was coded as yes, and no.
Discontinuation while not at risk	It was defined as women who discontinue because they want to become pregnant or for other fertility-related reasons (e.g. infrequent sex/husband away, marital dissolution/separation, difficult to get pregnant). It was coded as yes, and no.
Marital duration	Marital duration is calculated by subtracted age at the time of marriage from age at the time of the survey, in completed years. This measure in whole years was further reduced to 5-year categories as a standard variable in the standard recode datasets. The categories are 0–4 years, 5–9 years, 10–14 years, 15–19 years, 20–24 years, 25–29 years, and 30 years or more.
Marital status	Marital status was classified as: never married, married, living with a partner, divorced, widowed, and separated in the DHS. It was recoded as married or living with a partner, never married, and formerly married
Education	Women’s highest level of education attained. It was grouped as none, primary, secondary, and tertiary
Employment	Employment status of women 12 months prior to the survey. Recoded as currently working, and currently not working.
Number of living children	A number of living children at the time of the survey. It was recoded as 0, 1–4, and 5+
Husband characteristics
Husband age	This was measured in whole years in the dataset. It was recoded as 20–29 years, 30–39 years, 40–49 years, 50–59 years, and 60 years or more.