Table 1 Independent variables included in the estimated multilevel models

Respondent’s current age

Measured in single years

Respondent’s level of education

Defined as none, primary, and secondary education or higher

Perceived severity of malaria

Based on four items that assessed the perceived severity of malaria (Additional file 1: Table S1). Responses that reflect the perception that malaria is a serious condition were scored + 1, those that reflect lack of perceived severity were scored -1 while neutral responses were scored 0. The scores for all items were combined, and the total score split at 0 to denote perceived versus non-perceived severity

Perceived susceptibility to malaria

Based on four items that assessed the extent to which the respondents perceived themselves to be at risk for malaria (Additional file 1: Table S1). The items were scored, combined and split at 0

Attitudes towards bed nets

Derived from six items measured on the Likert scale (Additional file 1: Table S1). The items were scored, combined, and split at the median

Perceived response-efficacy of nets

Derived from three items that assessed beliefs about the efficacy of bed nets (Additional file 1: Table S1). The items were scored, combined, and split at 0

Perceived self-efficacy for net use

Derived from four items that assessed the respondent’s level of confidence in their ability to use nets for themselves or their children (Additional file 1: Table S1). The items were scored, combined, and split at the median

Descriptive norm about net use

Derived from a question that asked what proportion of people in the respondent’s community sleeps under a net each night. Response denoting at least half of community members sleep under a net were scored 1 while other responses were scored 0

Interpersonal communication about malaria

Discussion of malaria with others, including spouse, friends and family

Exposure to malaria-related messages

Exposure to malaria messages from the media and community sources

Integrity of the ceiling of the dwelling unit

Defined as completely sealed versus partially sealed or absent ceiling

Household wealth quintile

Derived from selected household assets and dwelling characteristics using principal component analysis

Presence of under-five children in the household

Variable derived from the household schedule

Household net coverage

Defined as whether the household has at least one ITN for every two members

Type of place of residence

Defined as urban versus rural residence

Region/zone/district of residence

Defined as study regions in Cameroon, study zones in Côte d’Ivoire, and study districts in Sierra Leone