Table 1 Description of independent variables examined in analysis

Gender

Male or female, as established during recruitment

Current age

Age in years

Education level

Highest level of formal school attended, if any

Exposure to messages in last six months

Whether respondent had seen or heard any messages about malaria in the past six months

Household size

Sum of household listing from household questionnaire

Number of individuals, of any age, living in the household in question

Number of nets in the household

Sum of net listing from household questionnaire, total number of mosquito nets in the household

Household wealth quintile

Categorical variable with five potential values representing poorest to wealthiest household, a relative measure of wealth constructed using primary component analysis from questions related to characteristics of the interviewee’s dwelling (e.g., roof, wall, ceiling materials, access to utilities, and ownership of electronic devices, land, cattle)

Urban vs. rural setting

Refers to the relative population density of the enumeration cluster as determined by the institution who provided the enumeration maps for survey sampling

Region/ district/zone

Largest subnational administrative unit from which household were sampled

Ideational variables

 Positive net care attitudes*

Each respondent’s index score was summed across the statements. The resulting score for the two statements varied from -2 to + 2. A binary variable was created by classifying those with a negative or zero score as (0) not having positive net care attitudes. Those with a positive score (above zero, 1) were considered to have positive net care attitudes

 Positive attitudes towards net use*

Each respondent’s index score was summed across the six statements. A binary variable was created by classifying those with a negative or zero score as (0) not having positive attitudes towards net use. Those with a positive score (above zero, 1) were considered to have positive attitudes towards net use

 Severity of malaria *

Each respondent’s index score was summed across the four statements. A binary variable was created by classifying those with a negative or zero score as (0) not perceiving the consequences of malaria infection as severe (i.e., perceived severity of malaria). Those with a positive score (above zero, 1) were considered to perceive the consequence of malaria as severe

 Perceived susceptibility to malaria*

Each respondent’s index score was summed across the four statements. A binary variable was created by classifying those with a negative or zero score as (0) not perceiving themselves as susceptible to malaria infection. Those with a positive score (above zero, 1) were considered to perceive themselves as susceptible to malaria infection

 Discussed malaria with spouse, friends or relations in the last six months

Responses to the two questions were collapsed into a binary variable such that a respondent who answered “Yes” to both questions was assigned a value of (1) and a respondent who answered “No” to at least one question was assigned a value of (0)

 Perceived response efficacy of nets*

Each respondent’s index score was summed across the three statements. A binary variable was created by classifying those with a negative or zero score as (0) not perceiving nets as an effective way to prevent malaria. Those with a positive score (above zero, 1) were considered to perceive nets as an effective way to prevent malaria

 Perceived self-efficacy for net use*

Unlike the other Likert statements, respondents were asked to indicate if they were confident in their ability to perform a particular action by selecting “could,” or “could not.” Each respondent received a score for their response: (− 1) could not, (0) don’t know/not sure/missing, and (1) could. Each respondent’s index score was summed across the four statements. A binary variable was created by classifying those with a negative or zero score as (0) not being confident that they could use a net consistently. Those with a positive score (above zero, 1) were considered to be confident that they could use a net consistently

 Perceived net use as the norm in one’s community

Responses to this question were re-coded such that respondents who indicated one of the first three options were coded (1) and respondents who indicated one of the last two options, or for whom the response was missing, were coded (0)