Elite athletes' estimates of the prevalence of illicit drug use: Evidence for the false consensus effect
Matthew Dunn BA (Psych), PostGradDipPsych, GCertPopH, PhD, Lecturer, Johanna O. Thomas BA (Psych), MPH, Research Officer, Wendy Swift BA (Hons), MPH, PhD, Senior Lecturer, Lucinda Burns BA (Hons), MPH, PhD, Grad Cert Health Policy, Senior Lecturer.
Abstract
Introduction and Aims. The false consensus effect (FCE) is the tendency for people to assume that others share their attitudes and behaviours to a greater extent than they actually do. The FCE has been demonstrated for a range of health behaviours, including substance use. The study aimed to explore the relationship between elite athlete's engagement in recreational drug use and their consensus estimates (the FCE) and to determine whether those who engage in the behaviour overestimate the use of others around them.
Design and Method. The FCE was investigated among 974 elite Australian athletes who were classified according to their drug use history.
Results. Participants tended to report that there was a higher prevalence of drug use among athletes in general compared with athletes in their sport, and these estimates appeared to be influenced by participants' drug use history. While overestimation of drug use by participants was not common, this overestimation also appeared to be influenced by athletes' drug use history.
Discussion and Conclusions. The results suggest that athletes who have a history of illicit drug use overestimate the prevalence of drug use among athletes. These findings may be helpful in the formulation of normative education initiatives.[Dunn M, Thomas JO, Swift W, Burns L. Elite athletes' estimates of the prevalence of illicit drug use: Evidence for the false consensus effect. Drug Alcohol Rev 2012;31:27–32]
Introduction
The false consensus effect (FCE) is the tendency for people to assume that others share their attitudes and behaviours to a greater extent than they actually do [1]. The FCE is thought to result from a bias in social perception [2] and multiple theories have been examined in an attempt to discern how it operates. Marks and Miller in 1987 examined the FCE in relation to four theoretical perspectives, which included selective exposure and cognitive availability, whereby perceptions of similarity are affected by the ease with which evidence of similarity is accessed from memory; salience and focus of attention, in which the focus of attention on one's preferred position may increase estimates of consensus for one's position; logical information processing, where active reasoning and rational processes in the form of casual attribution underlies perceptions of similarities; and motivation, whereby perceiving similarity between self and particular targets may bolster perception social support and validate the correctness or appropriateness of a position [2]. The authors concluded that biases are influenced by a host of variables and that each theoretical perspective appears to have its own domain of application, albeit with some degree of overlap into other domains [2]. Furthermore, they concluded that two or more mechanisms may operate at the same time to produce assumed similarity [2].
The FCE has been demonstrated for a range of health behaviours, including substance use [3–6]. The FCE has been investigated in the sporting context, most notably with investigating substance use and tertiary student athletes. These studies have found that student athletes with high rates of alcohol use overestimate alcohol consumption among their athletic and non-athletic peers [7–10]. Recent research has also investigated the FCE regarding doping in elite athletes [11,12]. For instance, Petróczi and colleagues [12] found that the estimation of doping made by those athletes using performance-enhancing drugs (PED) exceeded that made by non-PED users, and the authors concluded that the FCE may be a successful way of estimating the prevalence of PED use where self-report may be distorted by socially desirable responses.
To date, there has been a lack of research investigating the FCE in regards to illicit drug use among elite athletes. Compared with PED, such as anabolic-androgenic steroids (AAS) and human growth hormones (HGH), athletes are more likely to come into contact with illicit drugs, such as ecstasy, cannabis and cocaine [13]; thus, their perception of the number of other athletes engaging in the use of these drugs could influence their own decision to use. The aim of the present research was therefore to explore the relationship between elite athlete's engagement in illicit drug use and their consensus estimates (the FCE) and to determine whether those who engage in the behaviour overestimate the use of others around them. If this is found to be the case, it will provide direction for more effective prevention campaigns that disavow this myth.
Method
Participants and procedure
Participants were elite athletes from eight national sporting organisations (rugby league, rugby union, athletics, hockey, softball, netball, diving and triathlon) and one national sporting institute (the Australian Institute of Sport) who were recruited as part of a larger study investigating illicit drug use in sport. An athlete was considered ‘elite’ if eligible for state or national selection in their sport. During the survey period (July 2008 to May 2009), a total of 974 usable surveys were returned and are used in the current analysis.
Surveys were completed at team meetings or competitions. A member of the research team informed athletes the purpose of the project; that all information provided was confidential and anonymous and that they could decline to participate at any stage. Surveys were self-completed and returned to the researcher. The recruitment and survey implementation process was consistent with two exceptions, where survey completion for netball and softball was coordinated through team managers at team meetings prior to a national tournament and returned to the research team on the day of or after competition. More information on the methodology can be found elsewhere [13,14].
Measures
Standard demographic information was collected. Lifetime and recent drug use information was collected for six substances (cannabis, ecstasy, meth/amphetamine, cocaine, GHB and ketamine) using questions adapted from the National Drug Strategy Household Survey (Australian Institute of Health and Welfare, 2009).
Participants were asked ‘In your experience, what proportion of athletes in your sport do you feel use illicit drugs?’ followed by ‘What proportion of athletes in general do you feel use illicit drugs?’ Categorical responses for both questions included ‘None’, ‘Under 2%’, ‘3–5%’, ‘6–10%’, ‘11–20%’ and ‘20%+’. These questions were adapted from a study investigating substance use among English footballers [15].
The layout of the survey was such that participants made their estimations prior to questions about their own drug use; however, participants were advised prior to the commencement of the survey that there were questions about their own use in the survey.
Data analysis
Percentages are presented for categorical variables and means or medians presented for continuous variables. Multinomial logistic regressions were conducted to make comparisons between those who reported never using a drug, those who had used a drug but not in the past year and those who had used a drug in the past year. All analyses were conducted using PASW Statistics Version 18 (SPSS Inc., Chicago, IL, USA).
Results
Characteristics of the sample
The majority of the sample was male (76%) with a mean age of 23.1 years (range 18–44 years); 78% were aged 20–29 years. The majority (96%) spoke English as their main language at home. Most had completed secondary education (66%) and one-quarter had obtained a tertiary qualification (28%). Virtually all (92%) participated in a team sport and most (76%) trained all of the time with other athletes. Half (51%) indicated that they were a ‘full-time athlete’, one-third (29%) indicated they were a ‘full-time athlete also engaging in other work’ and one-fifth (19%) indicated that they were a ‘part-time athlete’.
Previous analysis of data obtained for this study showed that 7% of participants reported the use of at least one of the six illicit drugs under investigation in the past year (see Dunn et al. [13] for more information). Thus, participants were divided into three groups: those who had never used any of the six drugs under investigation (‘no-users’; n = 744; 76%); those who had used at least one of the six drugs under investigation but not in the past year (‘lifetime users’; n = 163; 17%); and those who had used at least one of the six drugs under investigation in the past year (‘recent users’; n = 67; 7%).
Table 1 presents the demographic characteristics for the three groups. Compared with the non-users, lifetime users were more likely to know other athletes who use drugs. Compared with non-users, recent users were more likely to be a full-time athlete and to also know other athletes who use drugs. Both lifetime users and recent users were older than the non-users.
| Variable | Non-users n = 744 (95% CI) | Lifetime users n = 163 (95% CI) | Recent users n = 67 (95% CI) |
|---|---|---|---|
| Male (%) | 74.6 (72–78) | 79.1 (73–85) | 79.1 (69–89) |
| Age (mean) [SD; range] | 23 [3.6; 18–39] | 25*[4.1; 18–44] | 24*[2.6; 18–29] |
| Completed secondary education or obtained tertiary qualification (%) | 94.1 (94–96) | 94.5 (91–98) | 88.1 (80–96) |
| Full-time athlete (%) | 50 (46–54) | 48.5 (41–56) | 73.1 (63–84)*** |
| Know other athletes who use drugs (%) | 18.1 (15–18) | 37.4 (30–45)*** | 62.7 (51–74)*** |
- *Significant at the P < 0.05 level, using non-users as the baseline.**Significant at the P < 0.01 level. ***Significant at the P < 0.001 level. CI, confidence interval.
False consensus effect
Participants were asked to indicate the proportion of athletes both in their sport, and in sports in general, they felt used illicit drugs. There was a trend for participants to report that more athletes in general (i.e. those involved in sports other than their own) used illicit drugs than athletes in their sport—13% felt that no athletes in their sport used illicit drugs; 28% felt that under 2% used such drugs; 17% felt that 3–5% used such drugs; 13% felt that 6–10% used such drugs; 8% felt that 11–20% used such drugs; and 7% felt that 20%+ used such drugs. In comparison, 7% felt that, in general, no athletes used illicit drugs; 13% felt that under 2% used such drugs; 18% felt that 3–5% used such drugs; 20% felt that 6–10% used such drugs; 14% felt that 11–20% used such drugs; and 14% felt that 20%+ used such drugs.
Table 2 presents estimates of drug use for the three groups. Compared with non-users, both lifetime and recent users were more likely to believe that athletes in their sport were using illicit drugs. Similarly, compared with non-users, lifetime users were more likely to note that athletes in general were using illicit drugs. Compared with non-users and lifetime users, recent users were more likely to indicate that 20%+ of athletes in general were using illicit drugs.
| Variable | Non-users n = 744 | Lifetime users n = 163 | Recent users n = 67 |
|---|---|---|---|
| Proportion of athletes in your sport who use drugs (%) | |||
| None | 15.3 | 4.3** | 3.0*** |
| Less 2% | 28.0 | 36.2 | 11.9 |
| 3–5% | 16.4 | 21.5 | 16.4 |
| 6–10% | 11.8 | 19.0 | 16.4 |
| 11–20% | 7.4 | 6.7 | 22.4 |
| 20%+ | 5.1 | 10.4** | 22.4*** |
| Proportion of athletes in sport generally who use drugs (%) | |||
| None | 9.0 | 3.1** | 0 |
| Less 2% | 12.5 | 14.1 | 9.0 |
| 3–5% | 17.7 | 23.9 | 9.0 |
| 6–10% | 19.6 | 22.7 | 22.4 |
| 11–20% | 12.9 | 17.2 | 20.9 |
| 20%+ | 12.1 | 13.5 | 29.9*** |
- *Significant at the P < 0.05 level, using non-users as the baseline. **Significant at the P < 0.01 level.***Significant at the P < 0.001 level.
Directional accuracy
The proportion of athletes who reported using illicit drugs in the past year was 7% (95% confidence interval = 5–9%). Using this as the ‘true’ level of illicit drug use, the proportion of the sample that overestimated the prevalence of drug use among athletes in their sport and in general was 16% and 28% respectively. Figure 1 presents the proportion of athletes who estimated the prevalence of athletes in their sport and athletes in general who use illicit drugs as being over 11%+ for the three groups. Compared with the proportion of non-users who estimated the prevalence of athletes in their sport who use illicit drugs as being over 11%+, recent users were more likely to endorse the proportion as being over 11%+ (odds ratio = 0.2, 95% confidence interval = 0.1–0.3; P < 0.001) (Figure 1). Similarly, compared with the proportion of non-users who estimated the prevalence of athletes in general who use illicit drugs as being over 11%+, recent users were more likely to endorse the proportion as being over 11%+ (odds ratio = 0.3, 95% confidence interval = 0.2–0.5; P < 0.001) (Figure 1). No effect was found for lifetime use.
Prevalence estimates of other athletes who use illicit drugs as being over 11%+. 
Athletes in their sport,
Athletes in general.
Discussion
The current study provided evidence for the FCE in regards to estimates of illicit drug use among elite athletes. Participants tended to report that there was a higher prevalence of drug use among athletes in general compared with athletes in their sport, and these estimates appeared to be influenced by participants' own drug use history. Similarly, while overestimation of drug use by participants was not common, this overestimation also appeared to be influenced by athletes' drug use history. The findings from this study can help influence the construction of normative educational messages aimed not only at athletes competing at the elite level but also junior athletes who may have had limited occasion to come into contact with recreational drugs.
Overestimation of drug use by athletes in participants' sport and athletes generally was not common; the prevalence of athletes overestimating drug use was lower than that found in other studies investigating estimations of substance use. Findings from college-based samples have found that those who report illicit drug use overestimate the proportion of other students who engage in use [3–5,16–20], with some studies showing that between 50–70% of participants overestimate peer substance use [4,21]. Despite this, when athlete estimations were considered according to their drug use history, there was a higher proportion of participants overestimating athlete drug use among those who had ever or recently used illicit drugs.
In accordance with the World Anti-Doping Agency's various education commitments, many sporting organisations in Australia conduct annual drug information seminars for their athletes. Australian athletes have indicated that presentations and pamphlets are their preferred formats of education [14] and these formats are ideal for presenting normative education messages. For instance, presentations to athletes could present prevalence figures regarding drug use among athletes to help correct biases that athletes might hold.
Understanding how the FCE operates can assist not only in helping to understand why a particular behaviour may occur but also, as mentioned, help in the design of normative education on drug use in sport. It has been suggested that those who engage in a particular behaviour inflate the real consensus so that it normalises their behaviour, thus making it easier to engage in that behaviour [22]. In the context of the current study, this ‘safety in numbers’ theory [22] suggests that those athletes who engage in drug use may believe that their behaviour is not bad if other athletes are using drugs. While it would seem advantageous for normative education messages to present actual consensus figures for drug use among athletes, there are impediments. First, it has been suggested that there is an absence of a reliable estimate for drug use among athletes [12], and while data from anti-doping organisations [23], national sporting organisations [24] and public health research [13] suggest that low numbers of athletes engage in drug use, these methods rely on either biomedical detection or self-report and thus may be an underestimation.
Second, knowing another athlete who uses drugs may have a stronger impact on the decision than can be countered by educational messages. In the current sample, athletes who had used drugs were more likely to indicate knowing other athletes who also use drugs [13]. A study investigating intentions to use AAS among current non-users found that knowing someone who used AAS was related to positive intentions to engage in future use [25]. These issues may need to be taken into account when considering normative education initiatives.
Limitations
There are limitations to the current study. First, athletes were asked to make estimates about the use of ‘illicit drugs’ and were not provided with a definition as to what could or could not constitute an ‘illicit drug’; therefore, it is not possible to investigate whether the distortion observed was drug-specific. For example, Wolfson [3] found that cannabis and amphetamine users made significantly higher estimates of cannabis use than non-users, whereas amphetamine users gave significantly higher estimates of amphetamine use than non-users and cannabis-only users. Future research may wish to investigate whether the current findings are replicated when applied to drug-specific scenarios. Second, while this study was specifically investigating the use of ecstasy, meth/amphetamine, cannabis, cocaine, ketamine and GHB, participants may have been considering drugs more synonymous with performance enhancement, such as AAS and HGH when making their estimates. Third, this study relied on self-reported drug use, and future research may wish to combine physiological measures to validate self-report. Fourth, the study used self-reported drug use as the ‘true’ level of drug use. Again, future research may wish to determine this true level of drug use using physiological measures to confirm use. Finally, many of the athletes came from team sports and indicated that they trained most of the times with other athletes; as such, caution should be given when generalising the findings to athletes who compete in sports with a more individual focus.
Conclusions
The present study found that elite athletes tended to report that there was a higher prevalence of drug use among athletes in general compared with athletes in their sport, and these estimates appeared to be influenced by participants' own drug use history. Similarly, while overestimation of drug use by participants was not common, this overestimation also appeared to be influenced by athletes' drug use history. The findings from this study may be helpful in the formulation of normative education initiatives.
Acknowledgements
The current project was funded by the Australian Government Department of Health and Ageing. The authors would like to thank the athletes who participated in the study, as well as staff from the national sporting organisations and Australian Institute of Sport who assisted with this project. The authors would like to thank Mr Mark Deady and Ms Chiara Bucello of NDARC for comments on earlier drafts of this manuscript.
