UK adults’ implicit and explicit attitudes towards obesity: a cross-sectional study

The study examined anti-fat attitudes in a cross-section of UK adults (England, Ireland, Northern Ireland, Scotland, and Wales) and compared attitudes in relation to gender, age, BMI and exercise frequency. Implicit and explicit anti-fat attitudes were evident in our sample of UK adults in line with hypothesis 1. Anti-fat attitudes were higher in males, younger participants and more frequent exercisers, in support of hypothesis 2.

Our findings illustrate that in UK adults, anti-fat attitudes appear to be widespread. Given the stigmatisation that can result from pervasive anti-fat attitudes, interventions to modify anti-fat attitudes are required. Anti-fat attitudes appear to be robust and have proven difficult to modify [25]; however some promise has been reported in altering beliefs about the causes of obesity [26]. Current study findings suggest that particular groups could be targeted with attitude modification interventions: males, younger individuals, and frequent exercisers. There are plausible explanations for greater anti-fat attitudes in all these groups: males tend to be less empathetic than females [27], a heightened awareness of body appearance in younger individuals, and, the incidence and possible acceptance of weight-related criticism in exercise environments [28–30]. These are all modifiable factors suggesting that interventions targeting these may well be successful. Our descriptive data does not offer support for the explanations we propose. Thus they require confirmation in future work before being used to underpin interventions to address negative perceptions of obesity in these groups. Nevertheless, given that anti-fat attitudes can lead to the stigmatisation of obese people [31]; our findings highlight the need for anti-fat attitude intervention with UK adults.

Our data reveal some interesting, although possibly contradictory, findings regarding perceptions of the controllability of obesity and of the descriptors fat and obese. Females and younger respondents tended to perceive obesity as more controllable and the labels fat and obese as more insulting than males and older respondents. For younger respondents this appears logical as they reported more anti-fat attitudes, thus they perceive labels associated with the condition as insulting. In addition, correlations from the current study that support previous research [32], suggest that these anti-fat attitudes are likely to derive partially from the belief that obesity is controllable and that obese people are responsible, indeed to blame, for their condition. This interpretation does not explain the same pattern seen in females as they did not report particularly strong anti-fat attitudes. Thus it may be that the participants perception of the labels used to describe obese people are not directly related to, or derived from, their evaluative perceptions of obese people themselves.

The differences observed in perceived controllability of obesity in relation to BMI are unclear. Obese respondents reported lower perceived controllability than normal and overweight respondents. This may serve as a self-protective mechanism in obese people to maintain self-esteem as they apportion less self-blame for their obesity [17]. Or, it may reflect their lived experience of being obese, as substantial evidence suggests a role for uncontrollable factors such as genetics in becoming obese [33], and, obese people are aware of their own exercise and nutrition habits, unlike external others. Less clear is the finding that perceived controllability was lower in underweight compared with normal weight respondents. Possibly underweight people recognise that weight at both extremes of the continuum is not always within the individual’s control if they themselves suffer from an eating disorder or are not underweight through choice. These explanations are of course highly speculative given that our study did not seek to identify explanations for different obesity attitudes. Whilst they intuitively make sense future research is clearly warranted to examine these suggestions.

Interestingly, despite the differences observed in the explicit measures, as discussed above, there was a null effect in relation to implicit attitudes when compared across the demographic factors. Current study findings demonstrate that UK adults have implicit anti-fat or pro-thin bias, but no differences were observed for almost all of the demographic factors. Previously it has been suggested that implicit measures counter some of the limitations of explicit measures, such as response bias and demand characteristics [14, 15]. Thus, differences observed in explicit responses, may have been a result of participants reducing the extent of their anti-fat attitudes, whilst this was not observed via implicit measures. Thus the current study findings highlight the need to examine both implicit and explicit attitudes towards obesity.

Regardless, our findings do underscore the importance noted previously of recognising the terms used to describe overweight and obesity [34]. Although medical professionals may use the term obese in an objective sense to describe a clinical condition, for our sample and in particular younger, female respondents, this was perceived as an insulting label. This finding reinforces previous suggestions that the term obese should be avoided [35]. Moreover the findings go beyond previous suggestions that have demonstrated that the term ‘obese’ should be avoided with obese patients, as our study demonstrates that the term is perceived as insulting in participants across BMI categories. Recently, guidelines have been developed for using language more sensitively to avoid objectification of the individual and placing the condition before the person, for instance the term ‘diabetic’ has been replaced by ‘people with diabetes’ [36]. Similar adjustments would seem appropriate when discussing obese people. Studies that compare perceptions of obese people when different labels are used to describe them would be simple to conduct but may produce illuminating findings to guide the somewhat complex issue of terminology use.

Both fat phobia and anti-fat attitudes tended to be lower in overweight and obese respondents in line with previous research [7]. We might therefore suggest that obesity stigmatisation comes from non-obese people, which may serve to further alienate obese people. Interestingly though, regardless of BMI, when measured implicitly, all respondents reported an anti-fat or pro-thin bias. Even if not expressed explicitly, it appears that obese people in our sample have internalised the same anti-fat or pro-thin attitudes as have non-obese people. These findings present less apparent contradiction when we consider that self-reported attitudes are open to manipulation by the respondent, whether consciously or not [15]. In this instance, this manipulation could have occurred because obese people felt uncomfortable publicly denigrating themselves in explicitly reporting their attitudes towards obese people. Similarly, females’ implicit attitudes did not differ from males’ in their anti-fat or pro-thin bias but they explicitly reported less negative perceptions of obesity. This may reflect the greater social desirability tendency in females [37], or, as suggested above, greater empathy in females. Clearly, future studies are needed that replicate the implicit measure used here to tease out these individuals’ ‘true’ responses.

Whilst the sampling strategy has limitations, the sample was successful in other ways. For example, the sample included respondents from every country across the UK and is the first study to obtain perceptions from a large group of participants from the UK. This was made possible due to the online sampling method that offers alternative benefits, for example, internet-based studies provide an opportunity to achieve a greater diversity in their samples [38]. These authors also argue that preconceptions about internet-based research are incorrect. For instance that the resultant sample will be younger, but the sample is often similar to that observed in traditional university based samples. They also note that there is no evidence that results of internet-based research are confounded by false data or repeat responders, nor do internet-based questionnaires diminish the psychological properties reported for pen-and-paper versions, both common preconceptions. Furthermore, whilst the sampling method means the researcher is not present during data collection, some respondents did make contact with the researcher to address queries.

We do however acknowledge that there are inherent biases to this approach, which may have resulted in the greater proportion of respondents who were white, middle class, more highly educated and of a higher social economic status. The majority of respondents were female (74.2 %), aged 18–25 years (57.7 %) and were students (47.2 %). As we might expect with a volunteer, opportunistic sample, our sample composition does not exactly match that of the UK population [39]. Despite attempts to sample a varied population, a more strategic sampling approach to ensure sub-groups were more equally represented might have strengthened the conclusions drawn from these data. Our sample composition does not match the demographic profile of the UK population [39], which impacts the generalizability of the data. Nevertheless, our findings reflect those obtained with similar population subgroups, such as more anti-fat attitudes in males [10]. Thus it is likely that if a ‘representative’ sample were examined, findings would be similar to those obtained here.

The reader should be aware of these limitations when considering our findings but given the paucity of current evidence from UK samples, we offer an initial contribution to stimulate further study. It is also important to highlight that the implicit measure we employed represents both a strength and a limitation of our study. Its strength lies in offering a measure of what some authors have described as ‘true’ attitudes [15] but given the format of Implicit Association Tests responses can only indicate anti-fat or pro-thin bias and not an absolute level of anti-fat attitude.

The current study is the first to comprehensively examine obesity attitudes in the UK population, demonstrating that UK adults report both implicit and explicit anti-fat attitudes. To date, obesity stigmatisation and discrimination is not included in UK health policy such as the Department of Health’s Obesity and Health Eating policy [10]. Based on the current study findings, we suggest that obesity stigmatisation and discrimination is incorporated into the policy as an action. This appears to be particularly relevant with previous research suggesting that obesity stigmatisation and discrimination may be a barrier to engaging in some of the actions that are already present such as physical activity [28, 40].