Sensitivity to reward and punishment: Associations with fat and sugar intake among South African students

Abstract

This study examined whether higher sensitivity to reward predicted higher fat and sugar intake among a sample of South African students at a university in the Western Cape and explored whether this relationship was mediated by food cue responsivity. It also examined whether sensitivity to punishment predicted higher fat and sugar intake among those who eat in response to anxiety. University students (n = 320) completed a series of questionnaires that measured sensitivity to reward and punishment, diet, their tendency towards hedonic eating, and their tendency towards eating in response to anxiety. Results showed that higher sensitivity to reward predicted higher fat intake. This relationship was partially mediated by eating in response to food-rich environments (hedonic eating). Sensitivity to punishment failed to predict diet. The results of this study add to the growing body of evidence showing a relationship between sensitivity to reward and eating behaviours, and how this relationship might play out in a university environment.

Keywords

Behavioural approach system behavioural inhibition system diet emotional eating personality power of food reinforcement sensitivity theory

South Africa has the highest rates of overweight and obesity in sub-Saharan Africa, with 7 out of 10 women and 4 out of 10 men being considered overweight or obese (Lopez, Murray, & Gakidou, 2014). Overweight and obesity are associated with a wide range of health conditions, such as type 2 diabetes, hypertension, coronary heart disease, stroke, cancer, metabolic syndrome, and osteoarthritis (Kopelman, 2007). As such, levels of obesity within a society have a significant impact, not only on quality of life but also on associated healthcare costs. Finding food and eating it are behaviours which are essential for survival and reproduction (Rui, 2013). Over the years, a plethora of neural circuits have developed in humans, as in all species, to ensure that feeding behaviours are maintained (Rui, 2013). Sensory pathways which feed into the corticolimbic reward system in the brain are responsible for our processing of environmental food signals, while internal physiological cues indicate satiety. However, when the neural systems and their personality and behavioural correlates generate an over-abundance of psychological desire for palatable foods in response to environmental food cues, excessive eating can occur (Berthoud, 2011, 2012). The brain’s reward system, long linked to addiction, plays a key role in promoting eating in response to the hedonic properties of food (Holsen et al., 2012; Pecina, 2008). One of the main causes of overweight and obesity is a diet that is high in fat and sugar, two properties of food which are highly correlated with hedonic eating. This article explores how Jeffery Gray’s (1982, 1987) Reinforcement Sensitivity Theory (RST) – one theory through which this desire can be understood – accounts for variance in fat and sugar intake among a sample of urban university students in South Africa. Excessive desire for food-related reward, in an environment characterised by high availability of unhealthy foods, we suggest, accounts for some of the unhealthy dietary patterns of this population.

Weight and diet among the South African population

Recent national surveys into the eating habits of South African learners (National Youth Risk Behaviour Survey, NYRBS-2008) (Reddy et al., 2008) and the general population (South African National Health and Nutrition Examination Survey, SANHANES-2012) (Shisana et al., 2014) revealed that the diets of both adults and youth are high in sugar and fat. Weight and diet also differ among individuals according to gender, location, and age (Shisana et al., 2014). In both the SANHANES-2012 (Shisana et al., 2014) and the NYRBS-2008 (Reddy et al., 2008), the prevalence of overweight and obesity was significantly higher in females than in males across all age categories. Furthermore, body mass index (BMI) increased with age in both sexes during the period between the two studies.

As far as diet was concerned, the results from the NYRBS-2008 foreshadowed those of the SANHANES-2012, with both showing that, among the youth, consumption of sugar and fat was high (Reddy et al., 2008). The SANHANES-2012, for instance, reported that 18.3% of South Africans consumed high amounts of fat (Shisana et al., 2014). Overall, this survey showed that young, White South Africans living in urban areas were most likely to consume a diet high in fat than other members of the population (Shisana et al., 2014). Sugar consumption followed a similar pattern, showing young, White, urban youth to be at most risk for a high-sugar diet.

Aside from the racial specificity of these patterns in South Africa, such findings – in the realm of age and locale – are in keeping with research from elsewhere in the world, including the United States. Such research has shown that younger people tend to eat higher proportions of fatty and sugary foods than older people (Wakimoto & Block, 2001). It has been proposed that the relative freedom of food choice that characterises young adulthood, as well as perceived peer group norms, plays a role in the eating behaviour of youths (Louis, Davies, Smith, & Terry, 2007).

Shisana et al. (2014) and Vorster (2002) note that the distribution of the fat and sugar scores in the SANHANES-2012 and NYRBS-2008 also reflects the pattern of nutritional transitions documented in middle-income countries, with the highest fat and sugar scores being found in the youngest age groups in formal urban areas. Thus, normal developmental processes which render the youth likely to eat unhealthily (such as peer norms and striving for independence from the home) might be interacting with macro-level economic and societal changes to put South African youth at especially high risk for unhealthy diet.

However, the above-mentioned biological variables, and their personality and behavioural correlates, may play a role above and beyond that of sex, location, and age in determining diet (Tapper, Baker, Jiga-Boy, Haddock, & Maio, 2015; Tapper, Pothos, & Lawrence, 2010). An emerging body of work has drawn attention to the role of brain reward mechanisms and associated personality traits in the field of eating behaviours (Davis & Fox, 2008; Provencher et al., 2008). One theory that has presented a paradigm for exploring the relationship between personality and eating behaviour is Jeffery Gray’s (1982, 1987) RST.

RST conceives of personality as consisting of three systems, namely, the Behavioural Approach System (BAS), the Behavioural Inhibition System (BIS), and the Flight, Fright, Freeze System. The scope of our study is concerned with the BAS and BIS (Gray, 1982, 1987). As derived from RST, the BAS controls an individual’s tendency to approach reward-related stimuli (Gray, 1982, 1987). The BAS also meditates sensitivity to conditioned signals of reward and mediates sensitivity to conditioned signals of non-punishment (Cooper, Gomez, & Aucote, 2006). Its converse, the BIS, controls an individual’s sensitivity to punishment and encourages avoidance of signals related to punishment.

Of concern to this research is the fact that people with a more active BAS are more sensitive to reward and have a greater tendency to approach reward-related stimuli than individuals with a more active BIS. Individuals with a more active BIS are more sensitive to punishment and more likely to avoid signals and stimuli related to punishment than those with a less active BIS.

A number of studies have shown that individuals who display higher sensitivity to reward are more responsive to appetising food cues (Beaver et al., 2006; Tapper et al., 2010), have a greater preference for foods high in fat and sugar (Davis et al., 2007), and consume a diet that is higher in fat (Tapper et al., 2015) than those who have a lower sensitivity to reward.

In this study, we examine whether sensitivity to reward predicts fat and sugar intake in a South African student population, and the extent to which this relationship is mediated by the individual’s eating in response to food-rich environments (hedonic eating). As noted in the introduction to this article, obesity is associated with hyperactivity of the subcortical reward circuitry (Holsen et al., 2012). And, while several neural pathways are involved in the hedonic and incentive aspects of eating, these hedonic circuits in the brain are ‘likely to be impaired by genetic and/or environmental factors, resulting in energy imbalance, obesity, and obesity-associated metabolic diseases’ (Rui, 2013, p. 16). Under certain environmental, and in the present case, personality conditions, individuals are more likely to eat in response to food cues in the environment, and are more likely to be sensitive to the reward-related properties of food.

A second personality trait that may influence diet is sensitivity to punishment. An individual who is more sensitive to punishment may experience higher levels of anxiety (Hasking, 2006) than someone less so. Where an individual has a tendency to eat in response to anxiety, this may lead to a diet that is higher in fatty and sugary foods. Some support for this hypothesis comes from Voigt et al. (2009) who found that undergraduate students who were more sensitive to punishment had a poorer diet. Likewise, Tapper et al. (2015) found that higher sensitivity to punishment predicted a higher sugar intake. We therefore also examined the relationship between sensitivity to punishment and intake of fat and sugar, and the extent to which this effect is moderated by the individual’s tendency to eat in response to anxiety. We predicted that sensitivity to punishment would be associated with higher intakes of fat and sugar in persons who eat in response to anxiety, but not among those who do not.

It must be noted at the outset that this research was conducted on a university campus which is not representative of the broader South African context. However, it is not the purpose of this research to provide broad, generalisable findings concerning the determinants of eating behaviour among all South Africans. What this research does set out to do is provide evidence of the potential utility of using personality constructs to predict eating behaviour among urban youth. Additionally, it seeks to explore which, if any, of a set of personality-related factors, under conditions of food availability – and not food scarcity – might drive individuals to follow an unhealthy dietary path.

Method

Participants

Participants were 434 undergraduate students at a large residential university in South Africa. All undergraduate psychology students on the university campus were invited to participate in this study. However, those participants who were currently dieting to lose weight were not eligible for the study, and were excluded. After excluding these individuals, there were 320 participants (60 males, 260 females). Of these, 20 identified themselves as Black African, 232 as White, 55 as Coloured, 6 as Indian, and 7 of other ethnicity. The majority (97%) were aged between 18 and 25 years.

Instruments

Demographics

This information was obtained by administering a demographics questionnaire. The participants were asked to indicate their racial group, gender, and age.

Fat and sugar intake

These were assessed using a validated Food Frequency Questionnaire (FFQ) (Margetts, Cade, & Osmond, 1989), in which participants were asked to record the frequency with which they ate 63 different food items over the previous month. Among a British sample, the FFQ has shown good test–retest reliability (Armitage & Conner, 1999) and convergent validity (Margetts et al., 1989; Thompson & Margetts, 1993). To compute daily intake of fat and sugar, we first calculated average levels per gram for each of the 63 foods based on data provided by the British Foods Standards Agency (2002, 2008). Total daily food intake was calculated by multiplying frequency of consumption by average portion sizes based on Bingham and Day (1987) and the British Foods Standards Agency (2008). Finally, for each food item, the amount of fat and sugar per gram was multiplied by average daily portion size for each participant. The sum across the 63 foods provided daily total intakes, in grams, of fat and sugar, for each participant.

Sensitivity to reward and punishment

These were assessed using the Reinforcement Sensitivity Theory Personality Questionnaire (RST-PQ) (Corr & Cooper, 2015). This 84-item questionnaire includes subscales assessing sensitivity to reward (behavioural approach sensitivity; BAS) and sensitivity to punishment (behavioural inhibition sensitivity; BIS). For instance, a participant would be asked to rate how much the statement, ‘I put in a big effort to accomplish important goals in my life’, on a 4-point Likert scale from ‘not at all’, to ‘highly’ (a BAS subscale item). The RST-PQ shows good discriminant validity (Corr & Cooper, 2015) and internal consistency (Smederevac, Mitrovis, Colovic, & Nikolasevic, 2014). In this study, Cronbach’s alphas for the BAS and BIS subscales of the RST-PQ were .85 and .93, respectively.

Eating in response to food-rich environments (hedonic eating)

This was assessed using the 15-item Power of Food Scale (PFS) (Cappelleri et al., 2009). Participants were requested to state how much statements such as ‘If I see or smell a food I like, I get a powerful urge to have some’ applied to them, on a 4-point Likert scale, from ‘I don’t agree’ to ‘I strongly agree’ (Cappelleri et al., 2009). The PFS shows good construct and internal reliability (Cappelleri et al., 2009) and had a Cronbach’s alpha of .92 in this study.

Eating in response to anxiety

The Emotional Eating Scale (EES) (Arnow, Kenardy, & Agras, 1995) assesses the extent to which an individual eats in response to a range of different emotions. For instance, participants must rate how likely they are to eat when they feel sad, on a 5-point Likert scale, ‘no desire to eat’ to ‘an overwhelming urge to eat’ (Arnow et al., 1995). Across 25 items, participants rate the degree to which certain feelings elicit the urge to eat. The scale shows good construct validity and adequate reliability and validity (Schneider et al., 2012). We used the nine items that related to anxiety; they produced a Cronbach’s alpha of .84 in this study.

Procedure

Students received an emailed invitation to take part in the study in return for being entered into a prize draw for a R500 (approximately US$42) shopping voucher. The questionnaires were presented in the form of an online survey. After the designated data collection period, the survey was closed, and the responses exported. The data were then analysed by one of the researchers, using SPSS.

Ethical considerations

Ethical approval was obtained from the Research Ethics Committee: Human Research (Humanities) at Stellenbosch University. Prior to beginning the online questionnaires, the participants were asked to give informed consent to participate. As the questionnaires were administered online, their consent was given in the form of clicking ‘yes’ to the question ‘Do you understand and wish to continue with the questionnaires?’ If the participant selected ‘no’, then the site closed and they could not proceed.

Respondents could also opt out at any point in the process of answering. Following completion of the study, participants were presented with a debrief information page, explaining the purposes of the study to them. Upon debriefing, participants were also directed to contact either researcher if they had further questions. The data collected were kept confidential and only the researchers had access to them.

Data analysis

All analyses were conducted using SPSS. We excluded outliers >3.5 standard deviations (SDs) from the mean. We calculated mean intake of fat and sugar in grams per day for each participant. Independent t-tests and Pearson’s correlations were used to explore gender differences in predictor and criterion variables, and associations with age, respectively.

We used hierarchical regression models to test the main hypotheses. Given associations with age and gender (see above), these variables were entered into Step 1 of all models. Where interaction terms were being computed, variables were mean-centred. The first model (H1) examined whether higher sensitivity to reward (a high BAS score) accounted for a significant amount of variance in fat and sugar intake. To examine whether the relationship between sensitivity to reward and fat intake was mediated by individuals’ tendency to eat in response to food cues in the environment (Power of Food, POF), three additional regression models were employed. To establish mediation effects, following Baron and Kenny (1986), we examined the following relationships as to whether

BAS predicted POF;

POF explained variance in total fat intake when BAS was controlled for;

BAS effected total fat intake less when POF was controlled for.

If these three conditions are satisfied, Baron and Kenny (1986) state that a mediation effect can be established.

To examine whether high sensitivity to punishment (as measured by BIS score) predicted higher intakes of fat and sugar among those with a greater tendency to eat in response to anxiety (as measured by the EES), the second model (H2) included sensitivity to punishment at Step 2 and an interaction term between BIS and anxiety eating was entered at Step 3.

Results

Two outliers (>3.5 SDs from the mean) were removed from the fat intake variable (one male and one female) and one outlier (female) from the sugar intake variable. Mean intake of fat was 83 g/day (SD = 39) and mean intake of sugar was 201 g/day (SD = 102). Independent t-tests and Pearson’s correlations showed a significantly higher intake of fat among men compared to women, t(317) = 3.06, p = .002, and a greater tendency for women, compared to men, to report eating in response to anxiety, t(318) = 2.60, p = .011. Age showed a significant negative correlation with sugar intake (r = −.13, p = .20). Hierarchical regression models were used to test the main hypotheses. Given associations with age and gender (see above), these were entered into Step 1 of all models. Where interaction terms were being computed, variables were mean-centred. The results for H1 – that higher sensitivity to reward (a high BAS score) accounted for a significant amount of variance in fat and sugar intake – showed that, as predicted, higher sensitivity to reward was associated with higher fat intake (accounting for 2% of variance), though not with a higher sugar intake (see Table 1). To examine whether the relationship between sensitivity to reward and fat intake was mediated by POF, three additional regression models were employed. These showed that sensitivity to reward significantly predicted an individual’s sensitivity to food-rich environments (measure on the POF scale) (ΔR² = .032, p = .001, β = .18, p = .001), that POF still significantly predicted fat intake when BAS was controlled for (ΔR² = .022, p = .007, β = .15, p = .007), but that less variance in fat intake was accounted for when sensitivity to food-rich environments was controlled and sensitivity to reward entered as a predictor (ΔR² = .016, p = .02, β = .13, p = .02). The fact that sensitivity to reward still significantly predicted fat intake in this latter equation suggests that sensitivity to food-rich environments partially, but not fully, mediates the relationship between sensitivity to reward and fat intake.

Table 1.

Hierarchical linear regression models predicting fat and sugar intake from BAS.

	Fat			Sugar
	B	SE B	β	B	SE B	β
Step 1
Constant	77.85	5.56		232.28	14.77
Gender	16.50	5.54	.17**	4.74	14.74	.02
Age	0.34	1.10	.02	−6.79	2.88	−.13
R ²		.03*			.02^a
Step 2
Constant	33.14	16.87		219.51	45.17
BAS	0.50	0.18	.15**	0.14	0.48	.02
R ²		.05			.02
ΔR²		.02**			.00

BAS: Behavioural Approach System; SE: standard error.

p < .10; *p < .05; **p < .01; 1 = men, 0 = women.

The results for H2 – that high sensitivity to punishment (as measured by BIS score) predicted higher intakes of fat and sugar among those with a greater tendency to eat in response to anxiety – showed no significant main effects or interactions between sensitivity to punishment and anxiety eating on sugar intake (Table 2).

Table 2.

Hierarchical linear regression models predicting fat and sugar intake from BIS and anxiety eating.

	Fat			Sugar
	B	SE B	β	B	SE B	β
Step 2
Constant	77.90	5.58		232.70	14.73
BIS	0.06	0.16	.02	0.74	0.42	.10^a
Anxiety eating	4.30	3.12	.08	2.71	8.23	.02
R ²		.04			.03
ΔR²		.01			.01^a
Step 3
Constant	78.24	5.61		234.94	14.74
BIS × anxiety eating	−0.14	0.21	−.04	−0.98	0.55	−.11^a
R ²		.04			.04
ΔR²		.00			.01^a

BIS: Behavioural Inhibition System; SE: standard error.

p < .10.

Discussion

The results showed that higher sensitivity to reward among South African students predicted a higher intake of fat, but did not significantly predict sugar intake. These findings replicate results from a British sample (Tapper et al., 2015) and are also consistent with research conducted in Canada showing an association between reward sensitivity and a preference for fatty foods (Davis et al., 2007). The findings have potential implications for health promotion, particularly in light of recent advances in digital technologies that make it increasingly possible to tailor health advice. For example, a person with a high sensitivity to reward may find it effective to substitute problematic eating behaviours with behaviours that result in alternative forms of reward. It is possible that health interventions that provide virtual rewards for reaching certain milestones, or encourage users to compete against peers, may be more effective among individuals with a higher sensitivity to reward. The utility of matching effects has been shown in the attitude change literature (Maio & Haddock, 2010), although further research would be needed to confirm such possibilities in the present context. Our results also extend the findings of Tapper et al. (2015) by showing that the relationship between sensitivity to reward and fat intake was partially mediated by eating in response to food-rich environments (hedonic eating). This finding is also consistent with a review of the literature concerning the properties of the neural circuitry involved in the regulation of food intake and energy expenditure, in which Rui (2013) highlights the manner in which sensitivity to the hedonic properties of foods can lead to excessive eating, and obesity. This finding also draws attention to the fact that the sensory experience of eating is an important determinant of food intake control (McCrickerd & Forde, 2016), suggesting that research into how the positive hedonic response associated with certain sensory cues might be mitigated in order to minimise risk to those individuals who are most sensitive to them.

This finding is also consistent with experimental work that has shown that individuals with higher sensitivity to reward are more responsive to appetising food cues (Beaver et al., 2006; Tapper et al., 2010). However, in contrast to Tapper et al. (2015), we failed to find a significant relationship between sensitivity to punishment and intake of sugar, although the data indicated a trend in the predicted direction. This discrepancy may be linked to differences in the diets of the two samples. While mean daily sugar intake was 86 g/day in the British sample (Tapper et al., 2015), it was 201 g among our sample. It is possible that this discrepancy could reflect the different ages of the participants; while most participants in our study were students between the ages of 18 and 25 years, the mean age of participants in the British sample was 33 years (Tapper et al., 2015). The fact that sugar intake was negatively correlated with age in both samples supports such an interpretation.

As noted, research into associations between age and diet has revealed that, in the main, younger people tend to eat higher proportions of fatty and sugary foods. The proposed reasons for this include the freedom from parental constraints on eating, as well as susceptibility to peer influence, which characterise young adulthood (Brown, McIlveen, & Strugnell, 2000; Louis et al., 2007; Perry & Murray, 1982; Wakimoto & Block, 2001). It is possible that peer group is more influential in the diets of younger versus older adults, and thus among the participants in this study (Mostazir, Jeffery, Voss, & Wilkin, 2015). Furthermore, our findings support the pattern of nutritional transitions documented in middle-income countries, where the highest sugar intake is found among youth in formal urban areas (Shisana et al., 2014; Vorster, 2002). Thus, normal developmental processes, such as peer influence and striving for independence from the home, might interact with macro-level economic and societal changes to explain more of the variance in sugar intake than personality variables. While past research has indicated that reward hyposensitivity (low BAS) is associated with compensatory overeating (to stimulate brain rewards in the absence of adequate levels of activation), we found no such association (Kenny, 2011).

This study also found that the tendency to eat in response to anxiety did not moderate the relationship between sensitivity to punishment and sugar intake or sensitivity to punishment and fat intake. Our data therefore did not support the hypothesis that higher sensitivity to punishment would be associated with a high fat/sugar diet among those with a tendency to eat in response to anxiety. Other research has suggested that questionnaires designed to assess eating in response to negative emotions may better reflect concerns over eating rather than the eating behaviour per se (e.g., Evers, de Ridder, & Adriaanse, 2009). Thus, it is possible that our questionnaire failed to accurately capture this behaviour. As such, future work would benefit from including an observational measure of the extent to which anxiety elicits eating.

In terms of limitations of the study, the FFQ used to assess diet has not been validated in a South African population, so the extent to which it accurately captures differences in intake is unclear. It is possible that we may have found significant effects for sugar intake had the questionnaire specifically been developed for the South African population. Furthermore, the FFQ used in the study was standardised in the United Kingdom. Authors such as Cullen et al. (2002) and Trichopolou, Naska, and Costacou (2002) have shown that portion sizes, as well as composition of diet, may vary significantly by culture. Thus, further research needs to be done to develop a context-specific food questionnaire with validity and reliability among South African samples. The results of this study may reflect normal cultural differences in food intake among South Africans rather than BIS- or BAS-related differences.

Additionally, the sample was restricted to undergraduate students. Approximately 72% of our sample self-identified as ‘White’. This is not representative of the South African context, where Whites make up less than 10% of the population (Statistics South Africa, 2011). Compared to the general population, participants in this study were younger, better educated, and with a greater proportion of females. Given age, gender, and socioeconomic differences in diet, it is unclear whether the findings would generalise to the rest of the population. Further research, with a more representative sample, would help address this question.

Conclusion

The findings in this study support the pattern of nutritional transitions documented in middle-income countries, in that in our sample – young people in a formal urban area – sugar intake was higher than the national average (Shisana et al., 2014; Vorster, 2002).

Furthermore, our results showed that higher sensitivity to reward among these students predicted a higher intake of fat, but did not significantly predict sugar intake. These findings replicate results from a British sample (Tapper et al., 2015) and provide further evidence for the association between reward sensitivity and a preference for fatty foods (Davis et al., 2007).

Our results also extend the findings of Tapper et al. (2015) by showing that the relationship between sensitivity to reward and fat intake was partially mediated by eating in response to food-rich environments (hedonic eating). This finding is consistent with experimental work (Beaver et al., 2006; Tapper et al., 2010), and indicates that individuals who are sensitive to their food environment might be at greater risk of consuming unhealthy foods, if such foods are available.

We note that research into associations between age and diet have revealed that younger people tend to eat higher proportions of fatty and sugary foods due to susceptibility to peer influence, and more freedom to eat what they want to (Brown et al., 2000; Louis et al., 2007; Perry & Murray, 1982; Wakimoto & Block, 2001). It is possible that peer group is more influential in the diets of younger versus older adults, and thus among the participants in this study (Mostazir et al., 2015). Thus, normal developmental processes, such as peer influence, and exposure to a new, unconstrained eating environment, might interact with macro-level economic and societal changes (the nutritional transitions noted above), to explain more of the variance in sugar intake than personality variables.

Finally, as this study found that sensitivity to reward is of limited power in predicting fat intake, it is also possible that health interventions should target context in influencing students’ eating behaviour. As noted, past research has shown that students are influenced by context, peers, socialisation, and convenience when it comes to their eating behaviour, and not merely intrinsic factors such as personality (Lien, Jacobs, & Klepp, 2002). On a university campus, there are multiple sources of inexpensive, high-fat, high-sugar foods that are readily available. Thus, students’ food choices could be influenced if healthier alternatives were made available, which is supported by past studies and policy recommendations (Story, Kaphingst, Robinson-O’Brien, & Glanz, 2007). As this study found, there is a significant negative correlation between age and sugar intake, and intervention that makes healthier food options available could be implemented in schools and targeted at those who are youngest and most at risk for unhealthy eating behaviour.

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Armitage

C. J.

Conner

(1999). The theory of planned behaviour: Assessment of predictive validity and ‘perceived control’. British Journal of Social Psychology, 38, 35–54.

Arnow

Kenardy

Agras

W. S.

(1995). The Emotional Eating Scale: The development of a measure to assess coping with negative affect by eating. International Journal of Eating Disorders, 18, 79–90.

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Beaver

J. D.

Lawrence

A. D.

van Ditzhuijzen

Davis

M. H.

Woods

Calder

A. J.

(2006). Individual differences in reward drive predict neural responses to images of food. The Journal of Neuroscience, 26, 5160–5166.

Berthoud

H. R.

(2011). Metabolic and hedonic drives in the neural control of appetite: Who is the boss? Current Opinion in Neurobiology, 21, 888–896.

Berthoud

H. R.

(2012). The neurobiology of food intake in an obesogenic environment. Proceedings of the Nutrition Society, 71, 478–487.

Bingham

Day

(1987). Average portion weights of foods consumed by a randomly selected British population sample. Human Nutrition: Applied Nutrition, 41, 258–264.

British Foods Standards Agency. (2002). McCance and Widdowson’s the composition of foods (6th summary ed.). Cambridge, UK: Royal Society of Chemistry.

British Foods Standards Agency. (2008). Manual of nutrition (11th ed.). London, UK: The Stationery Office.

10.

Brown

McIlveen

Strugnell

(2000). Nutritional awareness and food preferences of young consumers. Nutrition & Food Science, 30, 230–235.

11.

Cappelleri

J. C.

Bushmakin

A. G.

Gerber

R. A.

Leidy

N. K.

Sexton

C. C.

Karlsson

Lowe

M. R.

(2009). Evaluating the Power of Food Scale in obese subjects and a general sample of individuals: Development and measurement properties. International Journal of Obesity, 33, 913–922.

12.

Cooper

Gomez

Aucote

(2006). The Behavioural Inhibition System and Behavioural Approach System (BIS/BAS) Scales: Measurement and structural invariance across adults and adolescents. Personality and Individual Differences, 3, 295–305.

13.

Corr

P. J.

Cooper

(2015). The Reinforcement Sensitivity Theory of Personality Questionnaire (RST-PQ): Development and validation. Psychological Assessment, 1-14.

14.

Cullen

K. W.

Baranowski

Owens

de Moor

Rittenberry

Olvera

Resnicow

(2002). Ethnic differences in social correlates of diet. Health Education Research, 17, 7–18.

15.

Davis

Fox

(2008). Sensitivity to reward and body mass index (BMI): Evidence for a non-linear relationship. Appetite, 50, 43–49.

16.

Davis

Patte

Levitan

Reid

Tweed

Curis

(2007). From motivation to behaviour: A model of reward sensitivity, overeating, and food preferences in the risk profile for obesity. Appetite, 48, 12–19.

17.

Evers

de Ridder

D. T. D.

Adriaanse

M. A.

(2009). Assessing yourself as an emotional eater: Mission impossible? Health Psychology, 28, 717–725.

18.

Gray

J. A.

(1982). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampual system (1st ed.). Oxford, UK: Oxford University Press.

19.

Gray

J. A.

(1987). The psychology of fear and stress. Cambridge, UK: Cambridge University Press.

20.

Hasking

P. A.

(2006). Reinforcement sensitivity, coping, disordered eating and drinking behaviour in adolescents. Personality and Individual Differences, 40, 677–688.

21.

Holsen

L. M.

Savage

C. R.

Martin

L. E.

Bruce

A. S.

Lepping

R. J.

. . .Goldstein

J. M.

(2012). Importance of reward and prefrontal circuitry in hunger and satiety: Prader-Willi syndrome vs simple obesity. International Journal of Obesity, 36, 638–647.

22.

Kenny

P. J.

(2011). Common cellular and molecular mechanisms in obesity and drug addiction. Nature Reviews Neuroscience, 12, 638–651.

23.

Kopelman

(2007). Health risks associated with overweight and obesity. Obesity Reviews, 8, 13–17.

24.

Lien

Jacobs

D. R.

Klepp

(2002). Exploring predictors of eating behaviour among adolescents by gender and socio-economic status. Public Health Nutrition, 5, 671–681.

25.

Lopez

A. D.

Murray

C. J. L.

Gakidou

(2014). Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: A systematic analysis for the Global Burden of Disease Study 2013. The Lancet, 384, 766–781.

26.

Louis

Davies

Smith

Terry

(2007). Pizza and pop and the student identity: The role of referent group norms in healthy and unhealthy eating. Journal of Social Psychology, 147, 57–74.

27.

Maio

G. R.

Haddock

(2010). The psychology of attitudes and attitude change. London, UK: SAGE.

28.

Margetts

Cade

Osmond

(1989). Comparison of a Food Frequency Questionnaire with a diet records. International Journal of Epidemiology, 18, 868–873.

29.

McCrickerd

Forde

C. G.

(2016). Sensory influences on food intake control: Moving beyond palatability. Obesity Reviews, 17, 18–29.

30.

Mostazir

Jeffery

Voss

Wilkin

(2015). Childhood obesity: Evidence for distinct early and late environmental determinants: A 12-year longitudinal cohort study (EarlyBird 62). International Journal of Obesity, 39: 1057–1062.

31.

Pecina

(2008). Opioid reward ‘liking’ and ‘wanting’ in the nucleus accumbens. Physiology and Behavior, 94, 675–680.

32.

Perry

C. L.

Murray

D. M.

(1982). Enhancing the transition years: The challenge of adolescent health promotion. Journal of School Health, 52, 307–311.

33.

Provencher

Begin

Gagnon-Girouard

Tremblay

Boivin

Lemieux

(2008). Personality traits in overweight and obese women: Associations with BMI and eating behaviours. Eating Behaviour, 9, 294–302.

34.

Reddy

S. P.

James

Sewpaul

Koopman

Funani

N. I.

Sifunda

. . .Omardien

R. G.

(2008). Umthente Uhlaba Usamila – The South African Youth Risk Behaviour Survey 2008. Cape Town: South African Medical Research Council. Retrieved from http://hivhealthclearinghouse.unesco.org/library/documents/umthente-uhlaba-usamila-second-south-african-youth-risk-behaviour-survey-2008

35.

Rui

(2013). Brain regulation of energy balance and body weight. Reviews in Endocrine and Metabolic Disorders, 14, 387–407.

36.

Schneider

K. L.

Panza

Appelhans

B. M.

Whited

M. C.

Oleski

J. L.

Pagoto

S. L.

(2012). The Emotional Eating Scale: Can a self-report measure predict observed emotional eating? Appetite, 58, 563–566. doi:10.1016/j.appet.2012.01.012

37.

Shisana

Labadarios

Rehle

Simbayi

Zuma

Dhansay

. . .Faber

(2014). South African National Health and Nutrition Examination Survey (SANHANES-1): 2014 edition. Cape Town, South Africa: HSRC Press. Retrieved from http://www.hsrc.ac.za/en/research-outputs/view/6493

38.

Smederevac

Mitrovis

Colovic

Nikolasevic

(2014). Validation of the measure of revised reinforcement sensitivity theory constructs. Journal of Individual Differences, 35, 12–21.

39.

Statistics South Africa. (2011). Census in brief. Retrieved from http://www.statssa.gov.za/census/census_2011/census_products/Census_2011_Census_in_brief.pdf

40.

Story

Kaphingst

K. M.

Robinson-O’Brien

Glanz

(2007). Creating healthy food and eating environments: Policy and environmental approaches. Annual Review of Public Health, 29, 253–272.

41.

Tapper

Baker

Jiga-Boy

Haddock

Maio

G. R.

(2015). Sensitivity to reward and punishment: Associations with diet, alcohol consumption, and smoking. Personality and Individual Differences, 72, 79–84.

42.

Tapper

Pothos

E. M.

Lawrence

A. D.

(2010). Feast your eyes: Hunger and trait reward drive predict attentional bias for food cues. Emotion, 10, 949–954.

43.

Thompson

R. L.

Margetts

B. M.

(1993). Comparison of a Food Frequency Questionnaire with a 10-day weighed record in cigarette smokers. International Journal of Epidemiology, 22, 824–833.

44.

Trichopolou

Naska

Costacou

(2002). Disparities in food habits across Europe. Proceeding of the Nutrition Society, 61, 553–558.

45.

Voigt

D. C.

Dillard

J. P.

Braddock

K. H.

Anderson

J. W.

Sopory

Stephenson

M. T.

(2009). Carver and White’s (1994) BIS/BAS scales and their relationship to risky health behaviours. Personality and Individual Differences, 47, 89–93.

46.

Vorster

H. H.

(2002). The emergence of cardiovascular disease during urbanisation of Africans. Public Health Nutrition, 5, 239–243.

47.

Wakimoto

Block

(2001). Dietary intake, dietary patterns, and changes with age: An epidemiological perspective. The Journals of Gerontology, Series A: Biological Sciences & Medical Sciences, 56(Suppl. 2), 65–80. Retrieved from http://biomedgerontology.oxfordjournals.org/content/56/suppl_2/65.full.pdf