Social norms, misperceptions, and mosquito net use: a population-based, cross-sectional study in rural Uganda

Malaria is a leading cause of morbidity and mortality in the world. To prevent malaria transmission, the World Health Organization (WHO) recommends sleeping under a mosquito net on a daily basis for the 3.2 billion people worldwide who remain at risk of malaria [1]. Despite evidence that nets are effective in preventing malaria transmission and that most people have a net under which to sleep, some individuals still do not sleep under a mosquito net [2, 3].

Past research on protective behaviours and attitudes has found that many people underestimate the prevalence of a protective behaviour or attitude in a reference group and believe it to rarely occur or be held (i.e., they perceive a minority to do it or hold that belief) even when the protective behaviour or attitude is present among a majority in that reference group [4‐9]. Likewise, people often overestimate the prevalence of risky behaviours and believe them to be common when they are not [10‐21]. That is, perceived descriptive norms, which are the behaviours an individual believes to be the most common within a reference group [22], may not be actual descriptive norms (also known as collective norms), which are the behaviours actually engaged in by a majority of the reference group [23‐26].

A similar phenomenon may occur with mosquito net use whereby people underestimate use among peers. For example, if people do not talk about sleeping habits with peers, then they may simply lack information about typical mosquito net use behaviour. Furthermore, if sleeping under a mosquito net is only discussed when someone has malaria symptoms, when mosquitoes have been noticed, or when there is a problem with a net, then using a mosquito net may seem less common than it is in reality. Similarly, local media (e.g., billboards, newspapers, television, radio) may disproportionately display or discuss people who are sick with malaria rather than people who are healthy, which also makes the protective behaviour seem less common. Thus, it is plausible that even in contexts where more than 50% of people sleep under a mosquito net every night, there may be many people who do not perceive this behaviour to be normative.

Humans have a tendency to follow the herd. Misperceiving a protective behaviour as uncommon when the behaviour actually is common is problematic if people’s choices and actions are motivated by their perception of what is typical behaviour. Indeed, the classic sociological dictum suggests that: ‘what is perceived as real is real in its consequences’ [27]. In addition, decades of research in social psychology have demonstrated this conformity to social norms. People adopt behaviours even when the behaviour goes against what an individual would otherwise do and conform to what they perceive is accepted by others [28‐31]. In general, humans’ desire to avoid being viewed as different or as an outcast and seek to avoid being socially sanctioned [32]. Thus, social norms, and more specifically, perceived social norms, may in part drive personally sleeping under a mosquito net every night.

The Theory of Planned Behaviour and the subsequent Integrated Behavioural Model conceptualize perceived norms (also labelled as subjective norms) as an important antecedent to intentions and behaviours [33‐35]. Starting from these psychological frameworks, Social Norms Theory then applies a sociological lens to the role of social norms in predicting behaviour by identifying both perceived norms at the micro level (i.e., an individual’s perceptions about typical behaviour among a reference group) and actual norms at a meso level (i.e., actual typical behaviour among reference groups) as potentially important predictors of attitudes and behaviours [23]. This theory emphasizes separate measurement of these social norm constructs to allow for identification of norm misperception and to be able to assess the extent to which perceived norms predict behaviour above and beyond the predictive contribution of actual norms. The Social Norms Approach applies this theory to behavioural research and intervention by calculating and describing the extent of norm misperceptions and then, assuming misperceptions exist, designing messages around positive actual norms as the basis of a social norms intervention to change misperceived norms and ultimately change attitudes and increase healthy or positive behaviour [23, 26, 36‐38]. This intervention approach focuses on existing positive norms among a peer group or community as opposed to showing negative messages, which may make harmful behaviour seem more common than is true and inadvertently increase the harmful behaviour. Additional theoretical discussion from the communication discipline builds upon these theories by focusing on moderators of the relationship between perceived norms and personal attitudes and behaviour [24, 39‐41].

A large body of observational studies on topics of public health importance provides empirical support for these theories and frameworks by finding strong associations between perceived norms and personal behaviour (e.g., see these reviews for some examples [42‐44]). In addition, studies using longitudinal analyses, quasi-experimental designs, and randomized controlled trials have shown that changes in individuals’ perceived norms about a variety of risky behaviours determine, at least in part, whether the individual personally engages in the behaviour [18, 21, 45‐63]. Given this theoretical and empirical history, several recent review articles discuss the importance of changing perceived social norms (i.e., correcting misperceived norms) to change behaviour and also discuss issues to consider as part of developing interventions that try to change perceived norms by directly or indirectly providing social norms messages based on actual norms [44, 64‐67]. Moreover, social norms messaging has been utilized within recent behavioural science-based interventions as an effective form of social nudging [68].

Although most of this work has primarily been conducted in high-income countries, a few studies from sub-Saharan Africa have identified misperceptions of social norms and associations between norm perception and personal behaviour about HIV-prevention related behaviours [9, 69‐72] and environmental-related attitudes and behaviours [73]. In addition, a few experimental studies on other topics in Africa have found that changes in perceived norms were associated with changes in personal behaviour [74‐77]. However, no studies on mosquito net use have compared the gap between an individual’s perception about whether daily mosquito net use was normative in a given population and the actual norm for mosquito net use in that population. If there are people who misperceive an existing positive mosquito net use norm, then an opportunity to correct the misperception is presented. And, if perception is associated with personal behavior, then intervening on individuals or a population to correct norm misperceptions might then increase daily mosquito net use among people who do not sleep under mosquito nets every night. In addition, people who sleep under a mosquito net for other reasons but who had misperceived the norm (and thus were ‘carriers of the misperception’) may feel more supported to continue engaging in the protective behaviour upon learning the true norm. They may also be less likely to spread incorrect assumptions about normative behaviour in their community during general conversation and support an overall climate of using mosquito nets.

In Uganda, the vast majority of households now own mosquito nets due to a nationwide campaign that was conducted in 2013–2014 to ensure universal free access to mosquito nets and provide a ratio of at least one net for every two people in a household [78]. Yet, ownership has not translated into consistent use: according to 2014–2015 nationally representative data, one-quarter of people in Uganda with at least one net at home report not having slept under it the previous night [79]. Thus, a cross-sectional, population-based study on social norms and mosquito net use was conducted in rural Uganda. The study aimed to (1) quantify the prevalence of people who misperceived the daily mosquito net use norm; and, (2) determine the extent to which perception of the norm was associated with personal daily mosquito net use. It was hypothesized that a substantial amount of people would erroneously perceive that daily mosquito net use was not normative in their community despite most people sleeping under a mosquito net every night. The second hypothesis was that people who thought daily mosquito net use was normative would be much more likely to personally sleep under a mosquito net every night as compared to people who misperceived that most adults in their community did not do so.

Participant characteristics are presented in Table 1. More than 60% were under 40 years old, 58% were married, and 69% had completed primary education or less. The average number of adults per household was 1.8 (SD = 0.5). Eighty-six percent of participants personally reported having at least one mosquito net in their household. Overall, 68% of participants reported having enough mosquito nets in their household to indicate a ratio of one or more mosquito nets per every two people in their household (the national target). Forty percent of participants reported having three or more mosquito nets in their household.

Daily mosquito net use among adults was the actual norm in this parish as 65% of all participants reported sleeping under a mosquito net every night. (This percentage incorporates participants who had no mosquito nets). Among participants reporting at least one mosquito net in the household, 75% reported sleeping under a net every night. Furthermore, sleeping under a mosquito net every night was also the actual norm for many sub-groups as 54–85% of people in different sociodemographic categories reported sleeping under a net every night (except for people aged 80 years or more and people without a net in the household) (Table 1). Even the majority of participants living in a household with fewer than one net for every two people (but at least one net in the household) reported sleeping under a mosquito net every night. Similarly, 57–81% of adults across each village reported doing so each night (Table 2).

Yet, 23% of participants erroneously thought that the majority of adults in their parish do not sleep under a mosquito net every night, and thus misperceived the norm. This level of norm misperception was consistent as 20–30% of participants across sociodemographic sub-categories believed that daily mosquito net use was not normative in the parish (Table 3).

For example, 24% of participants with zero nets in their household misperceived daily mosquito net use as uncommon in their parish as did 22% of participants in a household with a ratio of nets:people of 1.0 or greater. In addition, 8% of participants indicated not knowing how often adults in their parish sleep under a mosquito net. Thus, almost one-third of all participants did not realize that daily mosquito net use was the norm in their parish when combining participants who misperceived the norm with participants who were unsure about the norm. Similarly, 17–38% across villages misperceived the norm (Table 4).

Excluding participants who reported zero nets in their household, a simple bivariate association showed that among participants who perceived daily mosquito net use as normative, 81% reported personally sleeping under a mosquito net seven nights per week. In contrast among participants who thought daily mosquito net use was not normative, 61% reported personally sleeping under a mosquito net every night. Further, among participants who did not know their perception of the norm, 63% reported daily mosquito net use. Subsequent regression analyses found that an individual’s perception about the normative mosquito net use behaviour had a statistically significant association with personal mosquito net use after adjusting for several other explanatory variables (Table 5). Participants who perceived daily use as normative in their parish were 2.94 times more likely (95% CI 2.09–4.14, p < 0.001) to sleep under a mosquito net every night compared to participants who perceived that most adults in their parish did not do so while adjusting for several other variables. Other factors that were associated with daily mosquito net use included the ratio of nets:people in the household, the number of children in the household, the number of additional adults in the household, being female, being married, and having postgraduate education. For example, compared to having access to fewer than one net for every two people in the household, having a ratio of ‘0.5 to < 1.0’ nets per person in the household was associated with a 2.21 greater likelihood (95% CI 1.47–3.32, p < 0.001) of personally sleeping under a net every night. A supplemental analysis found similar estimates of the association between perception of the norm and personal use for both men and women.

In this study, 23% of adults in a malaria-endemic region misperceived sleeping under a mosquito net every night as rarely occurring among most adults in their community even though most adults personally reported sleeping under a mosquito net every night. Moreover, 5–25% of adults (depending on the sociodemographic category) indicated that they did not know how many nights per week most adults in their parish slept under a mosquito net. This novel identification of misperceived norms about regular mosquito net use is similar to results demonstrating the discrepancy between actual and perceived behavioural norms from social norms studies on other health-related behaviours and attitudes [4‐7, 10‐21, 25, 26, 65, 94‐96]. In addition, underestimating the normativity of a protective behaviour is similar to results about misperception of HIV testing norms within this same population in rural Uganda [9] and about HIV prevention behaviour in Tanzania and South Africa [69, 70, 72].

The second novel contribution of this study was finding that perceiving daily mosquito net use to be normative was a strong protective factor for personally sleeping under a mosquito net every night. The finding that the perceived norm is strongly associated with personal behaviour is consistent with results from studies on other topics [18, 21, 45‐58, 84, 97, 98]. In follow-up research, it will be important to assess whether malaria-specific knowledge and perceptions of risk, and belief in the effectiveness of mosquito nets to prevent malaria, confound or modify the association between perception and behaviour [84, 99‐104].

Interpretation of these findings is subject to limitations. First, the data are cross-sectional so causal direction between norm perception and personal behaviour cannot be determined. However, previous studies using longitudinal and experimental designs provide extensive evidence that change in an individual’s perception of the norm has led to changes in personal behaviour for several health-related behaviours (as cited previously).

A second limitation is that data about personal mosquito net use are based on self-report. Although the rates reported in this study are similar to rates from past research [79], it is possible that reporting may have been influenced by social desirability bias (i.e., given the regular government campaigns) or by their perceptions of the norm (whether correctly perceived or incorrectly perceived). While the extent of mis-reporting is quantifiable and important, it is unclear whether the degree of mis-reporting could be significant enough to change the findings reported here. For example, a recent meta-analysis of studies comparing self-report and objective data found that people overestimated their mosquito net use by 13.6% [114]. Thus, even if 10–15% of participants in this study had incorrectly reported sleeping under a mosquito net seven nights per week when they actually sleep under a net less often, the majority of participants in this study would still be using a mosquito net every night. In that scenario, the actual norm would not have changed and norm misperception would still exist within this population. Moreover, as participants were asked how many nights they sleep under a mosquito net, mis-reports from participants who reported using a net less than seven nights per week but still over-estimated the number of nights they sleep under a mosquito net would not affect the results from this study as this study is discussing a binary norm about daily use. Finally, even if there were people whose under-reporting was driven by perceptions that others did not use nets, then the prevalence of actual use would be even greater and the extent of norm misperception found in this study would still exist.

Future research on mosquito net use norms would improve if objective monitoring methods were used to measure actual net use norms [101]. In addition, follow-up studies could assess the importance of norms about different social reference groups [115‐118]. Inquiring about norms about “men your age in your village” or “women with young children” could perhaps show less norm misperception. However, although the potential association between perceived norms about proximal peers and personal behaviour may be stronger than the association between perceived norms about distal peers and personal behaviour, the extent of close peer norm misperception, and thus the possible extent of perceived peer norm correction would likely be less [26]. In contrast, even though the perceived norm about distal peers may be less influential, there is likely to be more norm misperception about distal peer groups [116]. This greater misperception thus allows for more potential change to occur in the perceived norm, and ultimately, perhaps, in personal behaviour. Relatedly, collecting network data could examine whether clustering of perceived norms about mosquito net use and actual use behaviours occurs within friendship networks or other kinds of networks [119]. Social network structure and the clustering of specific behaviour among close friends might make that behaviour seem more common overall than it actually is in the larger reference group [120].

Assessing social norms about a variety of behaviours related to mosquito net use when malaria vectors are most present, such as entry and exit of nets at night or the early morning, may also be worthwhile [121, 122]. Likewise, the power of perceptions about societal expectations around mosquito net use may also play a role. Injunctive norms, that is perceptions about what people ought to do, are distinct from behavioural descriptive norms and can exert their own direct influence or a modifying influence on the relationship between descriptive norms and personal behaviour [22, 24, 123].

There are two main findings from this population-based study on mosquito net use in rural Uganda. First, even though daily mosquito net use was reported by a majority of people in the targeted community, about a quarter of people misperceived daily use as not normative in this population, and another 8% did not know whether it was normative. Second, people who thought daily mosquito net use was normative among adults in their community were much more likely to report personally sleeping under a mosquito net every night. The estimated association was statistically significant, large in magnitude, and robust. These findings suggest an opportunity for anti-malarial interventions to correct misperceived norms about daily mosquito net use in malaria-endemic regions. An increase in accurately perceiving this protective behaviour to be common and normative may help advance malaria prevention efforts in sub-Saharan Africa.