Abstract
Background and aims:
Little is known about how perceived norms about alcohol consumption may influence high alcohol consumption rates in Uganda. This study estimated the accuracy of perceived norms about men’s alcohol consumption and estimated the association between perceived norms and personal consumption.
Design:
Cross-sectional, whole-population, sociocentric social network study.
Setting:
Eight rural villages in Rwampara District in southwestern Uganda in 2016–2018.
Participants:
A total of 719 men aged 18 years and older (representing 91% of permanent resident men).
Measurements:
Self-reported frequent (≥4 days per week) and heavy alcohol consumption (six or more drinks on one occasion, three or more occasions of intoxication, or spending an excessive amount on alcohol). Participants also reported whether they thought most other men in their village engaged in frequent and heavy alcohol consumption (perceived norms). Using the network study design, we calculated alcohol consumption behavior within villages and social networks. Perceived norms were compared with aggregated self-reports. Multivariable Poisson regression models were used to estimate the association between perceived norms and individual behavior.
Findings:
Throughout villages, frequent and heavy alcohol consumption ranged from 7 to 37%. However, 527 (74%) participants perceived, contrary to fact, that most other men in their villages frequently consumed alcohol, and 576 (81%) perceived that most others heavily consumed alcohol. Overestimation of alcohol consumption by others was pervasive among sociodemographic subgroups and was present irrespective of the actual consumption behavior at the village level and within social networks. Men who misperceived these alcohol consumption behaviors as being common were more likely to engage in frequent (adjusted relative risk [aRR] = 3.98; 95% confidence interval [CI] = 1.69–9.34) and heavy (aRR = 4.75; 95% CI, 2.33–9.69) alcohol consumption themselves.
Conclusions:
Most men in eight rural Ugandan villages incorrectly thought that frequent and heavy alcohol consumption were common among men in their villages. These misperceived norms had a strong positive association with individual drinking behavior.
Keywords: social norms, social networks, descriptive norms, perceived norms, binge drinking, alcohol consumption, sub-Saharan Africa, Uganda, alcohol use, misperception
INTRODUCTION
During the last several decades, the commercial availability of alcohol has expanded across sub-Saharan Africa. Increases in alcohol marketing, alcohol industry influence, and alcohol consumption have accompanied this expansion.[1–3] Updates to national alcohol regulatory frameworks, however, have lagged.[4] Meanwhile, the availability of treatment and support for people with alcohol use disorders remains limited.[5] Moreover, stigma attached to alcohol use and mental disorders is pervasive and undermines treatment seeking.[6–8]
Uganda has one of the highest per capita rates of alcohol consumption among men in sub-Saharan Africa.[9] While fewer than 8% of women report heavy drinking in Uganda, one in three men report heavy episodic drinking and 1 in 10 are classified as having alcohol use disorder.[9, 10] Despite these high rates, Uganda has few alcohol use disorder treatment programs and most are particularly inaccessible in rural areas.[11] Identifying novel entry points to reduce harmful levels of alcohol consumption among men in this context is critical for reducing personal alcohol-related harms associated with consumption[12] as well as reducing negative consequences for others, such as HIV transmission[13] and intimate partner violence.[14] Perceived norms about alcohol consumption in this context may be an understudied key factor in determining alcohol consumption behaviors among men in Uganda.
Conceptual Framework
People often incorrectly perceive the extent to which health behaviors among peers are normative, i.e. which health behaviors or health risk behaviors are widely prevalent among their peers.[15–18] College students in the United States and Europe represent the canonical example of this phenomenon: studies have consistently found that college students believe heavy alcohol consumption is more common among their peers than is actually true.[17] Moreover, individuals’ perceptions concerning typical alcohol consumption behavior are strongly associated with their own alcohol consumption behavior.[19–24] This body of research has focused primarily on studying alcohol consumption among young adults within educational or other institutional settings in high-income countries.
Emerging literature from sub-Saharan Africa has provided parallel evidence consistent with these findings. First, people overestimate the prevalence of high-risk or harmful behaviors, and underestimate the prevalence of protective behaviors, in relation to HIV[25–29] and other health conditions of interest.[30] These studies also show evidence of associations between norm perceptions and individual behavior. Second, field experiments investigating expressions of dissent, conflict resolution, corruption, and violence against women find associations between changes in perceived norms and changes in individual attitudes and behaviors.[31–34] However, the extent to which alcohol consumption norms are misperceived and the role of this misperception in driving individual alcohol consumption have not been studied within a general population sample in sub-Saharan Africa.
Additionally, no studies have assessed the extent to which exposure to alcohol consumption by others, either during childhood or adulthood, may confound estimates of the association between perceived norms about alcohol consumption and individual alcohol consumption. However, exposure to others’ alcohol consumption is associated with both perceived norms [35, 36] and individual alcohol consumption.[37–39] Failure to properly adjust for exposure to alcohol consumption by others may bias (away from the null) estimates of the association between perceived norms and individual consumption.
The relevance of this body of research for prevention of alcohol-related harms is clear: if men in Uganda overestimate alcohol consumption norms among their peers, and this misperception is a significant driver of their own heavy alcohol consumption, then the “social norms approach” might be used to correct misperceived alcohol consumption norms and, ultimately, change behavior.[40] In contrast to interventions that aim to educate people about the harmful biological effects of alcohol consumption or to persuade people about the risks of heavy consumption, ‘social norms’ interventions provide information about how peers actually consume less alcohol than is commonly believed.[40–43] This intervention approach has primarily been used to reduce high-risk alcohol consumption among college students,[17, 18, 44–46] but has also gained traction in other contexts.[47–57]
For this study, we analysed data about perceived alcohol consumption norms and personal reports of alcohol consumption from individuals and peers in an ongoing whole-population, sociocentric social network study of adults in rural Uganda.[58] The study design uniquely allowed us to collect and compare measures of exposure with perceived norms and estimate their associations with personal consumption. This study focused only on alcohol consumption behavior among men because self-report of potentially harmful levels of alcohol use is rare among women in this setting in Uganda. We hypothesized that most men would not report frequent or heavy alcohol consumption, that men across the whole population perceive that most other men in their villages engage in frequent and heavy consumption of alcohol, and that individual alcohol consumption behavior would be associated with one’s perception of alcohol consumption norms (after adjusting for exposure and other factors).
METHODS
Study Population and Procedure
We conducted a cross-sectional, whole-population study among all adult residents aged 18 years and older in 8 villages in one administrative parish in Rwampara District, a rural region in southwestern Uganda. This parish, located approximately 20 kilometers from the local commercial hub of Mbarara Town, was selected due to its tractable population and geographic size, the lack of non-profit and intervention presence in the parish, and its similarity to other rural areas in Uganda where most Ugandans reside.[59] Additionally, this region is similar to other low-resource rural contexts in sub-Saharan Africa in that most households engage in an agriculture-based economy or small-scale trading/enterprise, household food and water insecurity are common, and access to electricity and piped water is rare.[59–62]
Research assistants who spoke the local language (Runyankore) collected data in 2016–2018. Using a continuously updated parish census list of eligible adult residents, a research assistant approached a potential participant typically at their home and asked the person to participate in a study about health and wellbeing after undergoing an informed consent process. A signature or a thumbprint indicating consent to participate was obtained. Data were collected with a computer-assisted, survey-based interview tool. Survey questions had been written in English, translated into Runyankore, and then back-translated to English to verify the translation’s fidelity. Question piloting and translation followed an iterative process.
The network study design entailed using name generator questions to elicit the names of other adult parish residents with whom a participant directly interacted.[63] These questions focused on social interactions related to social exchange, food exchange, financial discussions, health discussions, and emotional support.[58, 64] Another question elicited the names of any spouses. Participants provided confirmation of responses via a photographic search function on the computer. All responses represented out-going personal network ties, and were collapsed across the six name generator questions.[65] Information was also available regarding in-coming personal network ties because all the eligible nominations were also eligible study participants due to the sociocentric network design of the study, which targets everyone within a specific boundary.[63, 64] Thus, the set of unique direct ties to a participant, regardless of direction, represented a participant’s personal network.[66] Ties to people who did not participate in the study were excluded, as were ties to women, because this study focused on men. Additionally, data self-reported by participants within a participant’s personal network were linked to the index participant.
Ethical approval was granted by the Partners Human Research Committee at Massachusetts General Hospital, the Research Ethics Committee at Mbarara University of Science and Technology, and the Vanderbilt Human Research Protections Program. We also received clearance from the Uganda National Council of Science and Technology and the Research Secretariat in the Office of the President of the Republic of Uganda. The analyses were not pre-registered; therefore, the results should be considered exploratory.
Measures
Alcohol consumption behavior.
The primary outcomes for this analysis were frequent alcohol consumption and heavy alcohol consumption. Participants reported their own frequency of alcohol consumption using the following categories as response options: within the past year (once a month or less, 2–4 times per month, 2–3 times per week, and ≥4 times per week), 1–5 years prior to interview, more than 5 years prior, and never. We defined frequent alcohol consumption as consuming alcohol ≥4 times per week (vs. 2–3 times per week or less). This threshold was based on a prior study of alcohol consumption across 35 countries (including Uganda) classifying high-frequency drinking as ≥5 times per week.[67]
Among participants who reported alcohol consumption within the past year, we administered a locally derived scale consisting of three items about alcohol consumption behavior [68]: consuming 6 or more drinks in a single sitting in the past 12 months; spending ≥25,000 Ugandan shillings (approximately $10 at the time of the survey) on alcohol in the past 30 days; or being intoxicated on 3 or more of the past 30 days. The first item was the only one to refer to “drinks” and therefore also included these additional instructions: “I understand that you may share drinks and that some drinks have different sizes. For the purposes of this question, “one drink” should be considered equal to 1 shot or 1 tot of a strong alcoholic drink like waragi or vodka, or 1 full glass of a light alcoholic drink like beer”. This instruction was pilot-tested. The definition gave participants a sense of the large quantity of alcohol that 6 or more drinks represented when thinking about their response to the first item. We defined heavy alcohol consumption as reporting at least one of these three behaviors.[68] This study did not collect number of alcoholic drinks typically consumed because alcoholic drinks of non-standardized quantities and types are commonly consumed in this context.[69]
We used these individual-level data to calculate population-level norms of alcohol consumption behavior at the village level. Frequent alcohol consumption was considered normative in a village if >50% of men in the village reported frequent consumption of alcohol. The population-level norm for heavy alcohol consumption was defined similarly. This >50% threshold to identify population norms has been used in the social norms literature, both in Uganda[25, 30] and in other settings.[70–72]
We then combined the individual-level data with personal network data to calculate the number of male alters (i.e., men in the index participant’s personal network) who reported frequent alcohol consumption and the number of male alters who reported heavy alcohol consumption. We also created two binary variables: exposure (versus no exposure) to at least one male alter who reported frequent alcohol consumption and exposure (versus no exposure) to at least one male alter who reported heavy alcohol consumption.
Perceived norms about alcohol consumption behavior.
Four questions elicited participants’ perceptions about norms of alcohol consumption among other men in their villages. They were asked to report their perceptions of the frequency with which most other adult men in their village consumed any alcohol, consumed 6 or more drinks in a single sitting, spent ≥25,000 Ugandan shillings on alcohol in the past 30 days, or were intoxicated on 3 or more of the past 30 days. “Other adult men in your village” was set as the social reference group.[73, 74] Our pre-testing questionnaire suggested that this reference group was easily understood by participants as a group to which they belonged.[50] Responses about perceived norms were re-coded in the same way as personal reports.
To identify misperceived norms, we compared participants’ perceptions about most other men’s alcohol consumption behavior against the actual population norms of alcohol consumption behavior at the village level. Take, for example, a participant who lived in a village where heavy alcohol consumption was not normative. If this participant mistakenly perceived that most other men in their village engaged in heavy alcohol consumption, then this participant would be considered as misperceiving the norm (in this case, overestimating the extent to which heavy alcohol consumption was normative). This approach to measuring misperceived norms draws on methods from previous studies in this setting[8, 25, 75] and elsewhere.[30, 76–79]
Additional covariates.
To adjust for other sources of potential exposure, we asked participants to report whether they had lived during childhood with an adult who consumed alcohol excessively or misused drugs. Married or cohabiting participants were also asked whether they believed their partner engaged in any past year alcohol consumption.
To adjust for personal attitude about alcohol consumption, we asked participants how they personally felt about intoxication among men, using a Likert scale. Due to skewness in responses, we grouped together the responses ‘never okay to drink’ and ‘okay to drink but not to get drunk’ into one category (‘thought intoxication was not okay’) and ‘okay to get occasionally drunk as long as it does not interfere with responsibilities’, ‘okay to get drunk even if it interferes with responsibilities’, and ‘a frequent drunk is okay if that is what the individual wants to do’ into the other category (‘thought intoxication was okay’). This question was adapted by locally piloting a version similar to one used in the United States.[80]
Finally, participants reported their religion, education (completed primary school versus did not complete primary school), marital status, and HIV status (negative, positive, unknown). Household wealth quintile was created using a household asset-based index.[81] A positive symptom screen for depression was measured using a locally adapted depression subscale of the Hopkins Symptom Checklist.[82] The total number of male alters in a participant’s personal network and village of residence were also recorded.
Statistical analysis
We first examined the prevalence of alcohol consumption behavior and perceived norms of alcohol consumption behavior within socio-demographic, exposure, and attitude categories. To estimate the association between reported consumption and perceived norms, we fitted Poisson regression models with cluster-correlated robust estimates of variance to adjust for clustering at the village level. The primary outcomes for these analyses were the binary variables for frequent and heavy alcohol consumption. With a binary dependent variable, the modified Poisson regression model has been shown to yield estimated incidence rate ratios that can be interpreted straightforwardly as relative risk ratios.[83] The primary explanatory variable of interest was misperceiving alcohol consumption norms (binary). Regression models adjusted for sociodemographic factors, exposure to alcohol consumption behavior of others, and personal attitudes about intoxication. Analyses were conducted with Stata version 16.[84]
We conducted several sensitivity analyses to assess the robustness of our findings. First, we excluded 39 participants who reported belonging to a traditionally abstinent religion (Born Again Pentecostal, Muslim, Seventh Day Adventist).[85–89] We then included religion (Catholic, Protestant, other) in the regression models. Second, we re-fitted the regression models to data for only married/cohabiting men. In these regression models, we additionally adjusted for perception of their partner’s alcohol consumption. Third, we reclassified frequent alcohol consumption as ≥2 times per week for both the outcome and main explanatory variable and re-fitted the general model. Fourth, we fitted linear probability models with identical outcomes and explanatory variables and village fixed effects, and accounted for network autocorrelation by specifying a weight based off an adjacency matrix (using the lnam package in R).[90] Finally, based on results from the primary regression analyses, we used methods proposed by Vanderweele and Ding to calculate the e-value,[91, 92] which represents the minimum strength of association (on the risk ratio scale) that an unobserved confounder would need to have with both the exposure (norm misperception) and the outcomes (frequent or heavy alcohol consumption) to completely account for the estimated associations, conditional on the included covariates.
RESULTS
From 2016 to 2018, 719 men were interviewed (91% response rate), with 57 to 117 men interviewed per village. One person was deemed ineligible for the study because he was acutely intoxicated at each of multiple interview attempts. While everyone responded to the items about personal alcohol consumption, five did not report their perception about norms of frequent alcohol consumption and six did not report their perception about norms of heavy alcohol consumption. Fewer than 1% of observations for other variables were missing. The mean age was 40 years [standard deviation (SD) = 16], most participants [489 (68%))]had completed primary school or more and most [468 (65%)] were married/cohabiting. The median number of male alters was 5 [interquartile range (IQR) = 3 – 8]. While most men identified as Protestant [503 (70%)] or Catholic [174 (24%)], 39 (5%) identified as adherents of traditionally abstinent religions [Born-Again Pentecostal (n=29), Muslim (n=9), and Seventh-Day Adventist (n=1)]. None of the participants from traditionally abstinent religions reported frequent alcohol consumption and only one reported heavy alcohol consumption. Most participants [455 (63%)] did not think intoxication was okay.
Overall, 377 men (52%) reported consuming alcohol within the past 12 months, 86 men (12%) reported frequent alcohol consumption, and 181 (25%) reported heavy alcohol consumption. Nineteen men (3%) reported frequent but not heavy alcohol consumption, 114 (16%) men reported heavy but not frequent alcohol consumption, and 67 men (9%) reported both frequent and heavy alcohol consumption. The median number of male alters who reported frequent alcohol consumption was 0 (IQR = 0–1) and the median number of male alters who reported heavy consumption was 1 (IQR = 0–2).
The prevalence of self-reported frequent and heavy alcohol consumption ranged from 0–44% across sociodemographic subgroups (Table 1). Across villages, 7–24% of men reported frequent alcohol consumption and 16–37% of men reported heavy alcohol consumption. Thus, by definition, neither frequent nor heavy alcohol consumption were normative (because each was reported by fewer than 50% of men within each village). Moreover, even when a less stringent threshold was used to define frequent consumption (two or more times per week), this behavior was not normative in any village (i.e., the prevalence of frequent alcohol consumption according to this lower threshold ranged from 22–43% across villages)
Table 1.
Prevalence of self-reported frequent and heavy alcohol consumption among men in eight villages in Rwampara District, southwest Uganda (N=719).
| N (%) Participants | n (%) Reporting personal frequent alcohol consumption | n (%) Reporting personal heavy alcohol consumption | ||||
|---|---|---|---|---|---|---|
| Total | 719 | (100%) | 86 | (12%) | 181 | (25%) |
| Age (years) | ||||||
| 18–25 | 160 | (22%) | 11 | (7%) | 36 | (23%) |
| 26–35 | 174 | (24%) | 21 | (12%) | 44 | (25%) |
| 36–45 | 139 | (19%) | 15 | (11%) | 39 | (28%) |
| 46–55 | 126 | (18%) | 21 | (17%) | 34 | (27%) |
| ≥56 | 115 | (16%) | 17 | (15%) | 26 | (23%) |
| Marital status | ||||||
| Not married / cohabiting | 251 | (35%) | 30 | (12%) | 68 | (27%) |
| Married / cohabiting as if married | 468 | (65%) | 56 | (12%) | 113 | (24%) |
| Religion | ||||||
| Protestant | 503 | (70%) | 58 | (12%) | 120 | (24%) |
| Catholic | 174 | (24%) | 28 | (16%) | 60 | (34%) |
| Born-again Pentecostal | 9 | (1%) | 0 | (0%) | 1 | (11%) |
| Muslim | 29 | (4%) | 0 | (0%) | 0 | (0%) |
| Other (Not religious; Other; Seventh Day Adventist) | 4 | (1%) | 0 | (0%) | 0 | (0%) |
| Education | ||||||
| None / some primary education | 230 | (32%) | 36 | (16%) | 64 | (28%) |
| Completed primary education or more | 489 | (68%) | 50 | (10%) | 117 | (24%) |
| Household asset wealth | ||||||
| 1st quintile (poorest) | 123 | (17%) | 26 | (21%) | 44 | (36%) |
| 2nd quintile | 145 | (20%) | 18 | (12%) | 43 | (30%) |
| 3rd quintile | 136 | (19%) | 12 | (9%) | 27 | (20%) |
| 4th quintile | 156 | (22%) | 20 | (13%) | 40 | (26%) |
| 5th quintile (least poor) | 159 | (22%) | 10 | (6%) | 27 | (17%) |
| Personal attitudes about intoxication | ||||||
| Did not think intoxication was okay | 455 | (63%) | 27 | (6%) | 74 | (16%) |
| Thought intoxication was okay | 264 | (37%) | 59 | (22%) | 107 | (41%) |
| Had childhood exposure to adult who consumed alcohol excessively or who misused drugs | ||||||
| No | 297 | (41%) | 30 | (10%) | 57 | (19%) |
| Yes | 421 | (59%) | 56 | (13%) | 123 | (29%) |
| Participant reported spouse/partner to have consumed alcohol in past 12 months (among married or cohabiting men only, n=468) | ||||||
| No | 266 | (58%) | 24 | (9%) | 53 | (20%) |
| Yes | 194 | (42%) | 30 | (15%) | 58 | (30%) |
| Personal network size | ||||||
| 0–2 male alters | 107 | (15%) | 7 | (7%) | 16 | (15%) |
| 3–8 male alters | 433 | (60%) | 56 | (13%) | 125 | (29%) |
| ≥9 male alters | 179 | (25%) | 23 | (13%) | 40 | (22%) |
| Exposure to frequent alcohol consumption by men in personal network | ||||||
| No male alters reported frequent alcohol consumption | 370 | (51%) | 23 | (6%) | - | - |
| At least one male alter reported frequent consumption | 349 | (49%) | 63 | (18%) | - | - |
| Exposure to heavy alcohol consumption by men in personal network | ||||||
| No male alters reported heavy alcohol consumption | 192 | (27%) | - | - | 29 | (15%) |
| At least one male alter reported heavy alcohol consumption | 527 | (73%) | - | - | 152 | (29%) |
Notes: Frequent alcohol consumption was defined as alcohol consumption ≥4 times per week. Heavy alcohol consumption was defined as reporting consumption of ≥6 drinks on one occasion at least once in the past 12 months, spending excessive money on alcohol in the past 30 days, or being intoxicated 3 or more times in the past 30 days.
While the population data demonstrated that frequent and heavy alcohol consumption were not normative, 527 men (74%) incorrectly thought that most adult men in their own village engaged in frequent alcohol consumption and 576 men (81%) incorrectly thought that most adult men in their own village engaged in heavy alcohol consumption. Most men misperceived alcohol consumption norms, by overestimating rates of potentially harmful consumption, irrespective of sociodemographic and other subgroups (Table 2).
Table 2.
Misperceptions of normative alcohol consumption behavior among men in eight villages in Rwampara District, southwest Uganda (N=719).
| Participants who misperceived frequent alcohol consumption as the norm | Participants who misperceived heavy alcohol consumption as the norm | |||
|---|---|---|---|---|
|
|
||||
| n | (%) | n | (%) | |
| Total | 527 | (74%) | 576 | (81%) |
| Age (years) | ||||
| 18–25 | 119 | (75%) | 136 | (85%) |
| 26–35 | 131 | (76%) | 144 | (85%) |
| 36–45 | 96 | (70%) | 108 | (78%) |
| 46–55 | 92 | (73%) | 99 | (79%) |
| ≥56 | 85 | (75%) | 84 | (74%) |
| Marital status | ||||
| Not married / cohabiting | 183 | (73%) | 206 | (83%) |
| Married / cohabiting as if married | 344 | (74%) | 370 | (80%) |
| Religion | ||||
| Protestant | 353 | (71%) | 392 | (79%) |
| Catholic | 142 | (82%) | 151 | (87%) |
| Born-again Pentecostal | 8 | (89%) | 9 | (100%) |
| Muslim | 22 | (76%) | 22 | (79%) |
| Other (Not religious; Other; Seventh Day Adventist) | 2 | (50%) | 2 | (50%) |
| Education | ||||
| None / some primary education | 186 | (81%) | 197 | (86%) |
| Completed primary education or more | 341 | (70%) | 379 | (78%) |
| Household asset wealth | ||||
| 1st quintile (poorest) | 97 | (80%) | 104 | (85%) |
| 2nd quintile | 110 | (76%) | 118 | (82%) |
| 3rd quintile | 94 | (70%) | 108 | (80%) |
| 4th quintile | 107 | (69%) | 120 | (77%) |
| 5th quintile (least poor) | 119 | (75%) | 126 | (80%) |
| Personal attitudes about intoxication | ||||
| Did not think intoxication was okay | 333 | (74%) | 357 | (79%) |
| Thought intoxication was okay | 194 | (74%) | 219 | (84%) |
| Had childhood exposure to adult who consumed alcohol excessively or who misused drugs | ||||
| No | 212 | (72%) | 230 | (78%) |
| Yes | 314 | (75%) | 345 | (83%) |
| Participant reported spouse/partner to have consumed alcohol in past 12 months (among married or cohabiting men only, n=467) | ||||
| No | 189 | (72%) | 209 | (79%) |
| Yes | 151 | (78%) | 157 | (81%) |
| Personal network size | ||||
| 0–2 male alters | 80 | (76%) | 85 | (82%) |
| 3–8 male alters | 325 | (75%) | 347 | (81%) |
| ≥9 male alters | 122 | (69%) | 144 | (81%) |
| Exposure to frequent alcohol consumption by men in personal network | ||||
| No male alters reported frequent alcohol consumption | 263 | (72%) | - | - |
| At least one male alter reported frequent consumption | 264 | (76%) | - | - |
| Exposure to heavy alcohol consumption by men in personal network | ||||
| No male alters reported heavy alcohol consumption | - | - | 150 | (80%) |
| At least one male alter reported heavy alcohol consumption | - | - | 426 | (81%) |
Notes: Frequent alcohol consumption was defined as alcohol consumption ≥4 times per week. Heavy alcohol consumption was defined as reporting consumption of ≥6 drinks on one occasion at least once in the past 12 months, spending excessive money on alcohol in the past 30 days, or being intoxicated 3 or more times in the past 30 days.
In a multivariable Poisson regression model, misperceiving the norm was associated with a greater risk of frequent alcohol consumption after adjustment for sociodemographic factors (adjusted relative risk [aRR]=3.82; 95% confidence interval [CI] = 1.61–9.09). After additionally adjusting for personal attitudes about intoxication and exposure to others’ consumption behavior, the estimated association remained statistically significant (aRR=3.98; 95% CI, 1.69–9.34) (Table 3). Similarly, misperceiving the norm was associated with a greater risk of heavy alcohol consumption in the fully adjusted multivariable regression model (aRR=4.75; 95% CI 2.33–9.69) (Table 4).
Table 3.
Modified Poisson regression models estimating associations between perceived norms and frequent alcohol consumption among men in eight villages in Rwampara District, southwestern Uganda.
| Frequent alcohol consumption | |||
|---|---|---|---|
| (n=708) | |||
|
|
|||
| aRR | (95% CI) | p-value | |
| Perceived norm | |||
| Incorrectly thought that most men in own village consume alcohol frequently | 3.98 | (1.69, 9.34) | 0.002 |
| Correctly thought that frequent alcohol consumption was not the norm | REF | ||
| Exposure to frequent alcohol consumption by men in personal network | |||
| At least one male alter reported frequent alcohol consumption | 2.43 | (1.56, 3.79) | <0.001 |
| No male alters reported frequent alcohol consumption | REF | ||
| Personal attitude about intoxication | |||
| Thought intoxication was okay | 3.58 | (2.50, 5.12) | <0.001 |
| Did not think intoxication was okay | REF | ||
| Number of male alters in personal social network | 1.01 | (1.00, 1.03) | 0.121 |
| Childhood exposure to adult who consumed alcohol excessively or who misused drugs | |||
| Had exposure | 1.22 | (0.85, 1.76) | 0.278 |
| Did not have exposure | REF | ||
| Self-reported HIV serostatus | |||
| HIV serostatus unknown | 1.95 | (1.50, 2.52) | <0.001 |
| HIV positive | 1.60 | (1.02, 2.50) | 0.039 |
| HIV negative | REF | ||
| Depression status | |||
| Symptoms indicate probable depression | 0.96 | (0.48, 1.92) | 0.918 |
| Symptoms did not indicate probable depression | REF | ||
| Age (in years) | 1.01 | (1.00, 1.02) | 0.036 |
| Marital status | |||
| Married / cohabiting | 0.85 | (0.69, 1.05) | 0.133 |
| Divorced/separated/single | REF | ||
| Education level | |||
| Completed primary education | 0.96 | (0.64, 1.43) | 0.828 |
| Did not complete primary education | REF | ||
| Household asset wealth | |||
| 1st quintile (poorest) | 2.23 | (1.34, 3.71) | 0.002 |
| 2nd quintile | 1.46 | (0.97, 2.21) | 0.072 |
| 3rd quintile | 1.35 | (0.94, 1.93) | 0.104 |
| 4th quintile | 1.87 | (1.04, 3.38) | 0.037 |
| 5th quintile (least poor) | REF | ||
Notes: Each column represents one multivariable Poisson regression model fitted to the data. aRR = Adjusted relative risk ratio. CI = Confidence Interval. REF = Reference category for dichotomous and categorical variables. Frequent alcohol consumption was defined as alcohol consumption ≥4 times per week.
Table 4.
Modified Poisson regression models estimating associations between perceived norms and heavy alcohol consumption among men in eight villages in Rwampara District, southwestern Uganda.
| Heavy alcohol consumption | |||
|---|---|---|---|
| (n=707) | |||
|
|
|||
| aRR | (95% CI) | p-value | |
| Perceived norm | |||
| Incorrectly thought that most men in own village consume alcohol heavily | 4.75 | (2.33, 9.69) | <0.001 |
| Correctly thought that heavy alcohol consumption was not the norm | REF | ||
| Exposure to heavy alcohol consumption by men in personal network | |||
| At least one male alter reported heavy alcohol consumption | 1.59 | (0.95, 2.64) | 0.075 |
| No male alters reported heavy alcohol consumption | REF | ||
| Personal attitude about intoxication | |||
| Thought intoxication was okay | 2.21 | (1.98, 2.48) | <0.001 |
| Did not think intoxication was okay | REF | ||
| Number of male alters in personal social network | 0.99 | (0.98, 1.01) | 0.483 |
| Childhood exposure to adult who consumed alcohol excessively or who misused drugs | |||
| Had exposure | 1.34 | (0.94, 1.91) | 0.106 |
| Did not have exposure | REF | ||
| Self-reported HIV serostatus | |||
| HIV serostatus unknown | 1.73 | (1.53, 1.97) | <0.001 |
| HIV positive | 0.92 | (0.57, 1.49) | 0.743 |
| HIV negative | REF | ||
| Depression status | |||
| Symptoms indicate probable depression | 1.19 | (0.80, 1.78) | 0.389 |
| Symptoms did not indicate probable depression | REF | ||
| Age (in years) | 1.00 | (1.00, 1.01) | 0.182 |
| Marital status | |||
| Married / cohabiting | 0.95 | (0.75, 1.20) | 0.658 |
| Divorced/separated/single | REF | ||
| Education level | |||
| Completed primary education | 1.10 | (0.74, 1.64) | 0.640 |
| Did not complete primary education | REF | ||
| Household asset wealth | |||
| 1st quintile (poorest) | 1.47 | (1.15, 1.89) | 0.003 |
| 2nd quintile | 1.41 | (1.00, 2.00) | 0.051 |
| 3rd quintile | 1.05 | (0.78, 1.43) | 0.740 |
| 4th quintile | 1.36 | (1.04, 1.77) | 0.024 |
| 5th quintile (least poor) | REF | ||
Notes: Each column represents one multivariable Poisson regression model fitted to the data. aRR = Adjusted relative risk ratio. CI = Confidence Interval. REF = Reference category for dichotomous and categorical variables. Heavy alcohol consumption was defined as reporting consumption of ≥6 drinks on one occasion at least once in the past 12 months, spending excessive money on alcohol in the past 30 days, or being intoxicated 3 or more times in the past 30 days.
Results from sensitivity analyses suggest that our findings remained qualitatively similar after excluding men belonging to traditionally abstinent religions (Supporting information, Table S1) and in the subgroup of married/cohabiting men (Supporting information, Table S2). Additionally, the pattern of results remained similar when we defined frequent consumption as ≥2 times per week for the whole population (Supporting information, Table S3), when we included the number of male alters who reported frequent consumption and heavy consumption (instead of the binary variables for exposure to at least one male alter who reported frequent consumption and heavy alcohol consumption) (Supporting information, Table S4), and when we accounted for network autocorrelation (Supporting information, Table S5). Finally, we calculated an e-value of 7.42 for the frequent alcohol consumption model and an e-value of 8.97 for the heavy alcohol consumption model. These e-values suggest that an unobserved confounder would need to have an estimated association with both one’s own alcohol consumption behavior and perceived norm exceeding 7, on the risk ratio scale, to shift the associations estimated in our study to a null risk ratio of 1. Alternatively, an unobserved confounder would need to have an estimated association with both one’s own alcohol consumption behavior and perceived norm exceeding 2.5, on the risk ratio scale, to shift the associations estimated in our study to such a degree that the 95% CI could completely exclude a null risk ratio of 1.
DISCUSSION
In this whole-population network study of men across eight villages in rural Uganda, 25% reported heavy alcohol consumption and 12% reported frequent alcohol consumption. However, most men overestimated local population rates of potentially harmful alcohol consumption behavior and mistakenly thought that heavy and frequent alcohol consumption were normative among men in their village. These norm misperceptions were pervasive across social strata and strongly correlated with individual alcohol consumption behavior. The estimated associations were statistically significant, large in magnitude, and robust to potential confounding by unobserved variables.
Our study extends the literature on perceived norms and alcohol consumption in two key ways. First, we provide strong evidence of these phenomena in a general adult population of men in a low-resource setting in sub-Saharan Africa. Our findings are consistent with research on alcohol consumption among college students in other contexts [17, 93] and with perceived social pressure to consume alcohol in Uganda.[94] Second, we provide initial evidence that perceptions concerning consumption within a population-based social reference group substantially matter for individual behavior regardless of consumption behavior by family members or network ties.
Taken together, these novel findings offer considerable support for a population-wide social norms approach to reduce frequent and heavy alcohol consumption among men in rural Uganda. Communicating true alcohol consumption norms that represent moderate (or less) consumption behavior across the population could be especially effective given that most men already think intoxication is not acceptable, especially when it interferes with responsibilities. Additionally, reducing alcohol consumption by focusing on positive community norms may be more palatable in this context than discussing alcohol use disorder directly, given high stigma associated with mental and behavioral health issues.[6–8] Exemplar messages might take the form of ‘Most men in your village choose to drink alcohol three or fewer times per week or not at all,’ ‘Most single (or young or married, etc.) men in your village never have six or more drinks when consuming alcohol,’ and ‘Most men do not think it is okay to get drunk’. Such messages could be shared via various platforms such as billboards, radio shows, community meetings, social media, and text messages. This information could also be embedded in peer-based counseling interventions, financial incentive programs, or couples-based support programs.[95–97] Creation of such messages in collaboration with community members may maximize their credibility and reach.[17, 43, 98]
This kind of messaging could influence behavior across the population in multiple ways. First, it would encourage reductions in alcohol consumption among people who consume alcohol frequently or heavily. Second, men who had consumed alcohol in moderate amounts, and who had previously misperceived most other men to engage in frequent or heavy alcohol consumption, would be supported to continue consuming alcohol moderately or even less often because they would learn that most men typically consume alcohol less than they had thought. Similarly, this kind of messaging would also support men who do not consume alcohol and who had overestimated the norms so that they could continue to remain abstinent. Both the moderate consumption group and the abstinent group who had misperceived norms would learn that they are part of a larger moderately consuming or abstinent population. Additionally, male alters who had overestimated the norms might become less supportive of their contacts who do engage in heavy or frequent alcohol consumption. Finally, applying a social norms approach to reduce harmful levels of alcohol consumption in this context would not ignore the long cultural tradition of alcohol consumption during ceremonies and social events in Uganda,[99] nor would it preclude or stigmatize the informal production and selling of alcohol as an income-generating activity for some households.[100]
Limitations
First, although the population prevalence of alcohol consumption in this parish is similar to that reported elsewhere in Uganda,[9, 10] personal alcohol consumption may be under-reported.[101–103] Research using objective biomarkers to assess consumption levels would be helpful to determine actual population norms for comparison with perceived norms. However, population norms based on objective measurements would be unlikely to differ substantively. Moreover, some evidence suggests that, although there may be errors at the individual level, norms based on aggregated measures from survey data closely reflect norms based on aggregate measures of objective markers.[104] Second, our estimates could be subject to confounding by a variable unmeasured in our surveys. However, the estimated e-value suggests that any such confounding would need to be extremely strong in order to explain away the observed estimates. Third, this study was conducted among a small set of villages so we cannot claim generalizability of findings to the national population or other countries. However, findings provide a foundation for conducting research on misperceived alcohol use norms in similar sub-Saharan contexts (e.g., in South Africa which has similar national rates of heavy episodic drinking and alcohol use disorder among men).[105]
Conclusion
In this whole-population study of alcohol consumption behavior among men in eight villages in rural Uganda, we found that most men incorrectly thought that frequent and heavy alcohol consumption were common among men in their villages when, in fact, such behaviors were not typical at the village level nor within personal networks. These misperceived norms were strongly associated with individual consumption behavior. Interventions to correct these misperceptions may hold promise for reducing problematic or hazardous alcohol use among men in this setting.
Supplementary Material
Acknowledgments:
We thank the HopeNet cohort study participants, without whom this research would not be possible. We also thank members of the HopeNet study team for research assistance; in addition to the named study authors, HopeNet and other collaborative team members who contributed to data collection and/or study administration during all or any part of the study were as follows: Owen Alleluya, Patience Ayebare, Dickson Beinomugisha, Bridget Burns, Patrick Gumisiriza, Clare Kamagara, Justus Kananura, Allen Kiconco, Juliet Mercy, Patrick Lukwago Muleke, Rhina Mushagara, Rumbidzai Mushavi, Elijah Musinguzi, Moran Owembabazi, Immaculate Ninsiima, Mellon Tayebwa, and Dagmar Vořechovská. We also thank Roger Hofmann of West Portal Software Corporation (San Francisco, CA, USA), for developing and customizing the CASIC Builder software program used for survey administration. We also thank Claire Evans who contributed research assistance in finalizing the manuscript.
Funding Support: This study was funded by Friends of a Healthy Uganda and U.S. National Institutes of Health (NIH) R01MH113494 and R01MH125667. JMP acknowledges salary support from NIH K01MH115811.
Footnotes
Conflict of Interest: None
REFERENCES
- 1.Bakke Ø, Endal D. Vested interests in addiction research and policy alcohol policies out of context: drinks industry supplanting government role in alcohol policies in sub-Saharan Africa. Addiction. 2010;105(1):22–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Babor TF, Robaina K, Jernigan D. The influence of industry actions on the availability of alcoholic beverages in the African region. Addiction. 2015;110(4):561–71. [DOI] [PubMed] [Google Scholar]
- 3.Shield K, Manthey J, Rylett M, et al. National, regional, and global burdens of disease from 2000 to 2016 attributable to alcohol use: a comparative risk assessment study. Lancet Public Health. 2020;5(1):e51–e61. [DOI] [PubMed] [Google Scholar]
- 4.Ferreira-Borges C, Esser MB, Dias S, et al. Alcohol control policies in 46 African countries: opportunities for improvement. Alcohol Alcohol. 2015;50(4):470–6. [DOI] [PubMed] [Google Scholar]
- 5.Francis JM, Cook S, Morojele NK, et al. Rarity and limited geographical coverage of individual level alcohol interventions in sub Saharan Africa: findings from a scoping review. J Subst Use. 2020;25(1):11–9. [Google Scholar]
- 6.McCann TV, Renzaho A, Mugavin J, et al. Stigma of mental illness and substance misuse in sub-Saharan African migrants: a qualitative study. Int J Ment Health Nurs. 2018;27(3):956–65. [DOI] [PubMed] [Google Scholar]
- 7.Kalichman SC, Banas E, Kalichman M, et al. Stigmatisation of alcohol use among people receiving antiretroviral therapy for HIV infection, Cape Town, South Africa. Glob Public Health. 2020;15(7). [DOI] [PubMed] [Google Scholar]
- 8.Rasmussen JD, Kakuhikire B, Baguma C, et al. Portrayals of mental illness, treatment, and relapse and their effects on the stigma of mental illness: population-based, randomized survey experiment in rural Uganda. PLoS Med. 2019;16(9):1–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.WHO. Uganda Alcohol Consumption: Levels and Patterns. Global status report on alcohol and health 2018. Geneva: World Health Organization; 2018. [Google Scholar]
- 10.Kabwama SN, Ndyanabangi S, Mutungi G, et al. Alcohol use among adults in Uganda: findings from the countrywide non-communicable diseases risk factor cross-sectional survey. Glob Health Action. 2016;9(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kuule Y, Dobson AE, Harries AD, et al. Screening, diagnosis, and management of patients with alcohol use disorders at Bwindi Community Hospital, Uganda. Front Public Health. 2018;6(MAY):1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Graham K, Bernards S, Knibbe R, et al. Alcohol-related negative consequences among drinkers around the world. Addiction. 2011;106(8):1391–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hahn JA, Woolf-King SE, Muyindike W. Adding fuel to the fire: alcohol’s effect on the HIV epidemic in Sub-Saharan Africa. Curr HIV/AIDS Rep. 2011;8(3):172. [DOI] [PubMed] [Google Scholar]
- 14.Greene MC, Kane JC, Tol WA. Alcohol use and intimate partner violence among women and their partners in sub-Saharan Africa. Global Mental Health. 2017;4:E13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Perkins HW. Misperceptions of peer substance use among youth are real. Addiction. 2012;107(5):885–6. [DOI] [PubMed] [Google Scholar]
- 16.Perkins HW. Misperception is reality: the “Reign of Error” about peer risk behaviour norms among youth and young adults. In: Xenitidou M, Edmonds B, editors. The Complexity of Social Norms. New York: Springer; 2014. p. 11–36. [Google Scholar]
- 17.Perkins HW, Perkins JM. Using the social norms approach to promote health and reduce risk among college students. In: Cimini MD, Rivero EM, editors. Promoting Behavioral Health and Reducing Risk among College Students. New York: Routledge; 2018. p. 127–44. [Google Scholar]
- 18.Dempsey RC, McAlaney J, Bewick BM. A critical appraisal of the social norms approach as an interventional strategy for health-related behavior and attitude change. Front Psychol. 2018;9:1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Borsari B, Carey KB. Descriptive and injunctive norms in college drinking: a meta-analytic integration. J Stud Alcohol. 2003;64(3):331–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Perkins HW. Misperceptions of peer drinking norms in Canada: another look at the “reign of error” and its consequences among college students. Addict Behav. 2007;32(11):2645–56. [DOI] [PubMed] [Google Scholar]
- 21.Lewis MA, Neighbors C, Geisner IM, et al. Examining the associations among severity of injunctive drinking norms, alcohol consumption, and alcohol-related negative consequences: the moderating roles of alcohol consumption and identity. Psychol Addict Behav. 2010;24(2):177–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Neighbors C, Dillard AJ, Lewis MA, et al. Normative misperceptions and temporal precedence of perceived norms and drinking. J Stud Alcohol. 2006;67(2):290–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Perkins HW, Haines MP, Rice R. Misperceiving the college drinking norm and related problems: a nationwide study of exposure to prevention information, perceived norms and student alcohol misuse. J Stud Alcohol Drugs. 2005;66(4):470–8. [DOI] [PubMed] [Google Scholar]
- 24.Prentice DA, Miller DT. Pluralistic ignorance and alcohol use on campus: some consequences of misperceiving the social norm. J Pers Soc Psychol. 1993;64(2):243–56. [DOI] [PubMed] [Google Scholar]
- 25.Perkins JM, Nyakato VN, Kakuhikire B, et al. Actual versus perceived HIV testing norms, and personal HIV testing uptake: a cross-sectional, population-based study in rural Uganda. AIDS Behav. 2018;22(2):616–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Carey KB, Scott-Sheldon LAJ, Carey MP, et al. Community norms for HIV risk behaviors among men in a South African township. J Behav Med. 2011;34(1):32–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mulawa M, Yamanis TJ, Balvanz P, et al. Comparing perceptions with actual reports of close friend’s HIV testing behavior among urban Tanzanian men. AIDS Behav. 2016;20(9):2014–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Eggers SM, Mathews C, Aarø LE, et al. Predicting primary and secondary abstinence among adolescent boys and girls in the Western Cape, South Africa. AIDS Behav. 2017;21(5):1417–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hill LM, Moody J, Gottfredson NC, et al. Peer norms moderate the association between mental health and sexual risk behaviors among young men living in Dar es Salaam, Tanzania. Soc Sci Med. 2018;196:77–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Perkins JM, Krezanoski P, Takada S, et al. Social norms, misperceptions, and mosquito net use: a population-based, cross-sectional study in rural Uganda. Malar J. 2019;18(1):1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Blair G, Littman R, Paluck EL. Motivating the adoption of new community-minded behaviors: an empirical test in Nigeria. Sci Adv. 2019;5(3):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Paluck EL, Green DP. Deference, dissent, and dispute resolution: an experimental intervention using mass media to change norms and behavior in Rwanda. Am Polit Sci Rev. 2009;103(4):622–44. [Google Scholar]
- 33.Paluck EL. Reducing intergroup prejudice and conflict using the media: a field experiment in Rwanda. J Pers Soc Psychol. 2009;96(3):574–87. [DOI] [PubMed] [Google Scholar]
- 34.Green DP, Wilke A, Cooper J. Countering violence against women at scale: a mass media experiment in rural Uganda. 2019. [Google Scholar]
- 35.Lerman K, Yan X, Wu X-Z. The “Majority Illusion” in social networks. PLoS One. 2016;11(2):e0147617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Alipourfard N, Nettasinghe B, Abeliuk A, et al. Friendship paradox biases perceptions in directed networks. Nat Commun. 2020;11(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Anda RF, Whitfield CL, Felitti VJ, et al. Adverse childhood experiences, alcoholic parents, and later risk of alcoholism and depression. Psychiatr Serv. 2002;53(8):1001–9. [DOI] [PubMed] [Google Scholar]
- 38.Polenick CA, Birditt KS, Blow FC. Couples’ alcohol use in middle and later life: stability and mutual influence. J Stud Alcohol Drugs. 2018;79(1):111–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Rosenquist JN, Murabito J, Fowler JH, et al. The spread of alcohol consumption in a large social network. Ann Intern Med. 2010;152(7):426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Perkins HW, editor. The Social Norms Approach to Preventing School and College Age Substance Abuse: A Handbook for Educators, Counselors, and Clinicians. San Francisco: Jossey-Bass; 2003. [Google Scholar]
- 41.Perkins HW. College student misperceptions of alcohol and other drug norms among peers: exploring causes, consequences, and implications for prevention programs. Designing Alcohol and Other Drug Prevention Programs in Higher Education: Bringing Theory Into Practice. Newton, MA: Higher Education Center for Alcohol and Other Drug Prevention; 1997. p. 177–206. [Google Scholar]
- 42.Perkins H, Craig DW. A multifaceted social norms approach to reduce high-risk drinking: lessons from Hobart and William Smith Colleges. Newton, MA: Higher Education Center for Alcohol and Other Drug Prevention and U.S. Department of Education; 2002. [Google Scholar]
- 43.Haines MP, Perkins HW, Rice RM, et al. A guide to marketing social norms for health promotion in schools and communities. East Lansing, MI: National Social Norms Resource Center; 2005. [Google Scholar]
- 44.Miller DT, Prentice DA. Changing norms to change behavior. Annu Rev Psychol. 2016;67:339–61. [DOI] [PubMed] [Google Scholar]
- 45.Bewick BM, Bell D, Crosby S, et al. Promoting improvements in public health: using a social norms approach to reduce use of alcohol, tobacco and other drugs. Drugs (Abingdon Engl). 2013;20(4):322–30. [Google Scholar]
- 46.DeJong W, Schneider SK, Towvim LG, et al. A multisite randomized trial of social norms marketing campaigns to reduce college student drinking. J Stud Alcohol. 2006;67(6):868–79. [DOI] [PubMed] [Google Scholar]
- 47.Pedersen ER, Parast L, Marshall GN, et al. A randomized controlled trial of a web-based, personalized normative feedback alcohol intervention for young-adult veterans. J Consult Clin Psychol. 2017;85(5):459–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Prentice D, Paluck EL. Engineering social change using social norms: lessons from the study of collective action. Curr Opin Psychol. 2020;35:138–42. [DOI] [PubMed] [Google Scholar]
- 49.Bursztyn L, González AL, Yanagizawa-Drott D. Misperceived social norms: women working outside the home in Saudi Arabia. Am Econ Rev. 2020;110(10):2997–3029. [Google Scholar]
- 50.Tankard ME, Paluck EL. Norm perception as a vehicle for social change. Soc Issues Policy Rev. 2016;10(1):181–211. [Google Scholar]
- 51.Linnemayr S, Rice T. Insights from behavioral economics to design more effective incentives for improving chronic health behaviors, with an application to adherence to antiretrovirals. J Acquir Immune Defic Syndr. 2016;72(2):e50–e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Zapp D, Buelow R, Soutiea L, et al. Exploring the potential campus-level impact of online universal sexual assault prevention education. J Interpers Violence. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Perkins HW, Linkenbach JW, Lewis MA, et al. Effectiveness of social norms media marketing in reducing drinking and driving: a statewide campaign. Addict Behav. 2010;35(10):866–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Neighbors C, Rodriguez LM, Rinker DV, et al. Efficacy of personalized normative feedback as a brief intervention for college student gambling: a randomized controlled trial. J Consult Clin Psychol. 2015;83(3):500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Reid AE, Aiken LS. Correcting injunctive norm misperceptions motivates behavior change: a randomized controlled sun protection intervention. Health Psychol. 2013;32(5):551–60. [DOI] [PubMed] [Google Scholar]
- 56.Testa M, Livingston JA, Wang W, et al. Preventing college sexual victimization by reducing hookups: a randomized controlled trial of a personalized normative feedback intervention. Prev Sci. 2020;21(3):388–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Orchowski LM. “Trouble in Paradigm” and the social norms approach to violence prevention. Violence Against Women. 2019;25(14):1672–81. [DOI] [PubMed] [Google Scholar]
- 58.Takada S, Nyakato V, Nishi A, et al. The social network context of HIV stigma: population-based, sociocentric network study in rural Uganda. Soc Sci Med. 2019;233(April):229–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Uganda Bureau of Statistics - UBOS and ICF. Uganda Demographic and Health Survey 2016. Kampala, Uganda: UBOS and ICF; 2018. [Google Scholar]
- 60.Perkins JM, Nyakato VN, Kakuhikire B, et al. Food insecurity, social networks, and symptoms of depression among men and women in rural Uganda: a cross-sectional, population-based study. Public Health Nutr. 2018;21(5):838–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Mushavi RC, Burns BFO, Kakuhikire B, et al. “When you have no water, it means you have no peace”: a mixed-methods, whole-population study of water insecurity and depression in rural Uganda. Soc Sci Med. 2020;245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Tsai AC, Kakuhikire B, Mushavi R, et al. Population-based study of intra-household gender differences in water insecurity: reliability and validity of a survey instrument for use in rural Uganda. J Water Health. 2016;14(2):280–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.Marsden PV. Network Data and Measurement. Annu Rev Sociol. 1990;16(1):435–63. [Google Scholar]
- 64.Perkins JM, Subramanian SV, Christakis NA. Social networks and health: a systematic review of sociocentric network studies in low- and middle-income countries. Soc Sci Med. 2015;125:60–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Marin A, Hampton KN. Simplifying the personal network name generator: alternatives to traditional multiple and single name generators. Field Methods. 2007;19(2):163–93. [Google Scholar]
- 66.Marsden PV. Egocentric and sociocentric measures of network centrality. Social Networks. 2002;24(4):407–22. [Google Scholar]
- 67.Wilsnack RW, Wilsnack SC, Kristjanson AF, et al. Gender and alcohol consumption: patterns from the multinational GENACIS project. Addiction (Abingdon, England). 2009;104(9):1487–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 68.Fatch R, Bellows B, Bagenda F, et al. Alcohol consumption as a barrier to prior HIV testing in a population-based study in rural Uganda. AIDS Behav. 2013;17(5):1713–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Papas RK, Sidle JE, Wamalwa ES, et al. Estimating alcohol content of traditional brew in Western Kenya using culturally relevant methods: the case for cost over volume. AIDS Behav. 2010;14(4):836–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Perkins HW, Craig DW. Student-athletes’ misperceptions of male and female peer drinking norms: a multi-site investigation of the “Reign of Error”. J Coll Stud Dev. 2012;53(3):367–82. [Google Scholar]
- 71.Perkins HW, Craig DW, Perkins JM. Using social norms to reduce bullying: a research intervention among adolescents in five middle schools. Group Process Intergroup Relat. 2011;14(5):703–22. [Google Scholar]
- 72.Perkins HW, Meilman PW, Leichliter JS, et al. Misperceptions of the norms for the frequency of alcohol and other drug use on college campuses. J Am Coll Health Assoc. 1999;47(6):253–8. [DOI] [PubMed] [Google Scholar]
- 73.Shibutani T. Reference groups as perspectives. Am J Sociol. 1955;60(6):562–9. [Google Scholar]
- 74.Sherif M. The concept of reference groups in human relations. Group Relations at the Crossroads. Oxford: Harper; 1953. p. 203–31. [Google Scholar]
- 75.Tsai AC, Kakuhikire B, Perkins JM, et al. Measuring personal beliefs and perceived norms about intimate partner violence: population-based survey experiment in rural Uganda. PLoS Med. 2017;14(5):1–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 76.Perkins JM, Perkins HW, Craig DW. Misperceived norms and personal sugar-sweetened beverage consumption and fruit and vegetable intake among students in the United States. Appetite. 2018;129(June):82–93. [DOI] [PubMed] [Google Scholar]
- 77.Dempsey RC, McAlaney J, Helmer SM, et al. Normative perceptions of cannabis use among European university students: associations of perceived peer use and peer attitudes with personal use and attitudes. J Stud Alcohol Drugs. 2016;77(5):740–8. [DOI] [PubMed] [Google Scholar]
- 78.McAlaney J, Helmer SM, Stock C, et al. Personal and perceived peer use of and attitudes toward alcohol among university and college students in seven EU countries: Project SNIPE. J Stud Alcohol Drugs. 2015;76(3):430–7. [DOI] [PubMed] [Google Scholar]
- 79.Perkins JM, Perkins HW, Craig DW. Misperceiving a code of silence: peer support for telling authorities about weapons at school among middle school and high school students in the United States. Youth Soc. 2019;51(6):814–39. [Google Scholar]
- 80.Perkins HW, Berkowitz AD. Perceiving the community norms of alcohol use among students: some Research implications for campus alcohol education programming. Subst Use Misuse. 1986;21(9–10):961–76. [DOI] [PubMed] [Google Scholar]
- 81.Smith ML, Kakuhikire B, Baguma C, et al. Do household asset wealth measurements depend on who is surveyed? Asset reporting concordance within multi-adult households in rural Uganda. J Glob Health. 2020;10(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 82.Ashaba S, Kakuhikire B, Vořechovská D, et al. Reliability, validity, and factor structure of the Hopkins Symptom Checklist-25: population-based study of persons living with HIV in rural Uganda. AIDS Behav. 2018;22(5):1467–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 83.Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702–6. [DOI] [PubMed] [Google Scholar]
- 84.StataCorp. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC; 2019. [Google Scholar]
- 85.Seventh-day Adventist Church. Drugs: General Conference of Seventh-day Adventists; 2021. Available from: https://www.adventist.org/articles/drugs/.
- 86.Dudley RL, Mutch PB, Cruise RJ. Religious factors and drug usage among Seventh-day Adventist youth in North America. Journal for the Scientific Study of Religion. 1987;26(2):218–33. [Google Scholar]
- 87.Pew Forum on R, Public L. Tolerance and tension: Islam and Christianity in sub-Saharan Africa. Washington, DC, Pew Research Center. 2010;147. [Google Scholar]
- 88.Al-Ansari B, Thow A-M, Day CA, et al. Extent of alcohol prohibition in civil policy in Muslim majority countries: the impact of globalization. Addiction. 2016;111(10):1703–13. [DOI] [PubMed] [Google Scholar]
- 89.Assemblies of God. Abstinence from Alcohol: General Council of the Assemblies of God; 2016. Available from: https://ag.org/Beliefs/Position-Papers/Abstinence-from-Alcohol.
- 90.Butts CT. sna: Tools for Social Network Analysis. R package version 2.6 ed2020. [Google Scholar]
- 91.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the e-value. Ann Intern Med. 2017;167(4):268–74. [DOI] [PubMed] [Google Scholar]
- 92.Haneuse S, VanderWeele TJ, Arterburn D. Using the e-value to assess the potential effect of unmeasured confounding in observational studies. JAMA. 2019;321(6):602–3. [DOI] [PubMed] [Google Scholar]
- 93.Steyl T, Phillips J. Actual and perceived substance use of health science students at a university in the Western Cape, South Africa. Afr Health Sci. 2011;11(3):329–33. [PMC free article] [PubMed] [Google Scholar]
- 94.Breuer C, Bloom B, Miller AP, et al. “The bottle is my wife”: exploring reasons why men drink alcohol in Ugandan fishing communities. Soc Work Public Health. 2019;34(8):657–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 95.Haberer JE, Sabin L, Amico KR, et al. Improving antiretroviral therapy adherence in resource-limited settings at scale: a discussion of interventions and recommendations. J Int AIDS Soc. 2017;20(1):1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 96.Bassett IV, Wilson D, Taaffe J, et al. Financial incentives to improve progression through the HIV treatment cascade. Curr Opin HIV AIDS. 2015;10(6):451–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 97.Remien RH, Stirratt MJ, Dolezal C, et al. Couple-focused support to improve HIV medication adherence: a randomized controlled trial. AIDS. 2005;19(8):807–14. [DOI] [PubMed] [Google Scholar]
- 98.Hummer JF, Davison GC. Examining the role of source credibility and reference group proximity on personalized normative feedback interventions for college student alcohol use: a randomized laboratory experiment. Subst Use Misuse. 2016:1–15. [DOI] [PubMed] [Google Scholar]
- 99.Willis J. Potent brews: a social history of alcohol in East Africa 1850–1999. Oxford: James Currey; 2002. [Google Scholar]
- 100.Holmila M, Mäkelä P, Österberg E. A window to alcohol problems and politics in Africa. Kettil Bruun society meeting in Uganda. NAD Publication. 2011;28(1):97–9. [Google Scholar]
- 101.Adong J, Fatch R, Emenyonu NI, et al. Social desirability bias impacts self-reported alcohol use among persons with HIV in Uganda. Alcohol Clin Exp Res. 2019;43(12):2591–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 102.Bajunirwe F, Haberer JE, Boum Y, et al. Comparison of self-reported alcohol consumption to phosphatidylethanol measurement among HIV-infected patients initiating antiretroviral treatment in Southwestern Uganda. PLoS One. 2014;9(12):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 103.Muyindike WR, Lloyd-Travaglini C, Fatch R, et al. Phosphatidylethanol confirmed alcohol use among ART-naïve HIV-infected persons who denied consumption in rural Uganda. AIDS Care. 2017;29(11):1442–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 104.Craig DW, Perkins HW. Accuracy of estimated blood alcohol concentration norms from college student drinking survey data: verification using matched late-night breath measurements. J Stud Alcohol Drugs. 2018;79(3):455–64. [PubMed] [Google Scholar]
- 105.WHO. South Africa Alcohol Consumption: Levels and Patterns. Geneva: World Health Organization; 2018. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.