Table 1.

Variables used in the latent class analysis to develop HIV risk profiles for out-of-school 15- to 24-year-old women in Kenya, Malawi, and Zambia, 2016–2017

Domains and factors	Measurement and variable definition
Household characteristics
Orphanhood	Questions asked about loss of mother and father, defined as lost both parents, lost mother, lost father, and non-orphan
Household socioeconomic status	Composite measure created of three questions about household construction, water, and sanitation facilities and divided into tertiles of low, mid, and high
Hunger in the past month	Question about going without food for a whole day in the past month, defined as often/sometimes versus rarely/never
Adult supervision	Questions asked whether an adult in the household knows the respondent’s whereabouts during day or at night, defined as yes versus no
Respondent characteristics
Mobility	Question on movement or travel outside of community, defined as weekly/monthly, yearly, never
Marital status	Question asked about marital status (formal or civil or cohabitation as if married) of the respondents, defined as currently married, formerly married, and never married
Respondent attitudes
Support for gender equitable norms	Attitudinal measures assessed support for gender equitable norms using the previously validated GEM scale. The GEM scale assesses views on dimensions on home and child care (e.g., cooking and cleaning are the wife’s responsibility), sexual relationships (e.g., men are always ready to have sex), health and disease prevention (e.g., my partner would be outraged if I asked him to use a condom), and violence (e.g., there are times when a woman deserves to be beaten). An overall GEM scale score is calculated and categorized as low, medium, and high support for equitable gender norms (Pulerwitz and Barker 2008; Vu et al. 2017; Wesson et al. 2019)
Perception on exposure to HIV	Question on self-perception of HIV risk, “How likely is it that you have been exposed to HIV,” defined as (no risk/unlikely versus somewhat likely/very likely/don’t know) (Santelli et al. 2013)
Respondent knowledge
Comprehensive knowledge about HIV transmission	Composite measure based on standard Demographic Health Survey questions of correct knowledge of two ways to prevent HIV (e.g., can people reduce their chance of getting HIV by having just one uninfected sex partner who has no other sex partners) and rejection of three misconceptions about HIV (e.g., can people get the HIV/AIDS virus from mosquito bites), defined as yes/no
Comprehensive knowledge about condoms	Composite measure based on questions around correct knowledge that condoms are an effective method of preventing pregnancy, protecting against HIV/AIDS, and protecting against STIs, defined as yes/no