Over the past two decades, young adults’ consumption of sugar-sweetened beverages (SSBs) has decreased in the US, with more than a third of young adults reporting no regular SSB consumption, according to a report from Kit et al (2013). This same report revealed, however, that young adults still consume an average of 155 kcal a day from SSBs, energy just about equal to that of a 12 fl oz. can of soda (Kit et al., 2013). More recently in 2015, 20% of US high schoolers reported drinking soda daily according to National Youth Risk Behavior Surveys (Miller et. al., 2017). With so many young consumers, possible health complications caused by SSBs can impact large parts of the population, thus it is important to understand the effects SSB consumption has on our bodies.

In terms of their impact on dental health, the commercial carbonated SSBs Fanta and Coca Cola have been shown to have similar cariogenic (cavity-inducing) effects on teeth as table sugar (Giacaman et al., 2018). Another review of several soft drink consumption studies also found a small association between soft drink consumption and presence of cavities in teens. Nutritionally, this same review also found negative associations between SSB consumption and intake of protein (r = -0.14), fruit, fruit juice, and riboflavin (Vartanian et. al, 2007). The relationship between SSB consumption and obesity has also been explored. A study by Hu et al. argues that there is convincing evidence that decreasing SSB consumption lowers obesity risk. This study also observes a 55% higher risk of obesity in children with higher SSB consumption (2013). In addition, a systematic review and meta-analysis of 32 unique articles testing the relationship between SSB consumption and weight gain found that adding one serving of an SSB to one’s daily diet led to a 0.06 increase in BMI in children over the course of one year. This suggests a positive association between SSB consumption and weight gain in children (Malik et al., 2013). These health implications raise the question: Why are beverages with such negative effects still being consumed so frequently by teenagers?

One potential cause is another negative health complication of SSB consumption: its addictive nature. SSBs and other high-fat, high-sugar, highly processed foods are incredibly stimulating to reward systems and can be addictive (Schiestl, 2021). A study of dopamine release in obese participants and recreational drug addicts found a similar conditioning to stimulus—energy-rich food and drugs, respectively—and decreased dopamine release in both groups. In this way, the author’s results support a clear similarity between the neurochemistry associated with overeating and drug addiction (Volkow, 2017). Besides altering one’s neurochemistry, highly processed foods have been shown to compromise gut-brain signals, leading to increased craving of high-fat high-sugar foods (Small & DiFeliceantonio, 2019).

Another reason why children consume SSBs regardless of their negative health effects is advertising, specifically television advertising. Advertisements for SSBs, alongside other popular high-fat, high-sugar children’s foods, like sugar-coated cereals, make up 75% of food advertising on television (Calvert, 2008). Chances are that many people are frequently exposed to SSB advertising. According to a report by the global market measurement firm, Nielsen, television is still the most viewed type of digital media, with all users watching an average of 30.5 hours weekly (Nielsen, 2021). In fact, children are frequently exposed to food advertising, with children ages 6-11 being exposed to an average of 1313 food advertisements on TV in 2015 alone (Harris & Kalnova, 2018). A content analysis of children’s television channels showed that about half of advertising displayed during children’s programming is food advertising (Lewis & Hill, 1998). These statistics make one wonder how children are being exposed to this much advertising, and why do SSB brands advertise so heavily to children?

The answer to both questions is spending motivated by profit. The campaign to keep children consistently engaged with television advertisements that promote sugared beverages is driven by children’s increased spending power, and is the result of a large monetary investment by SSB companies. In 2002, 4-12-year-olds spent $30 billion while 12-17-year olds spent $112.5 billion. Aside from what they spend, children and teens also influence regular family purchases like meals, and 4-12 year olds have control over an estimated $500 billion in household costs (Calvert, 2008). Soft drink companies allotted 84% of their total advertising spending on TV ads in 2018, according to the 2020 Sugary Drink FACTS report, which uses Nielsen data to track the spending of 48 different soft drink brands, which each spent over $100,000 on SSB advertising in 2018. All together, these companies spent a total of $1.04 billion on SSB advertising in 2018, a 23% increase from their spending in 2013 (Sugary Drink FACTS, 2020). This advertising is often aimed at children and teens due to their increasing spending power. . With such clear incentive to advertise, despite the negative health effects of their products, what is done to regulate soft drink companies?

Such aggressive attempts to advertise unhealthy products are met with government restrictions such as beverage taxes. A study by Harris et al. reveals that beverage taxes are the main form of regulating SSB distribution and consumption as SSB companies avoid government advertising restrictions through First Amendment protections (2012). Their research reveals that taxing beverages based on sugar content does not affect consumption, but it does incentivize SSB brands to lower the sugar content of their drinks. Finally, Harris et al. argues that current counteradvertising and education campaigns are not able to combat current SSB advertising, which are supported by billion-dollar spending campaigns (Harris et al. 2012). Another review of advertising’s impact and the value of potential restrictions argues that more research is required, both inside and outside a laboratory setting, to better understand the relationship between advertising and SSB consumption in teens (Harris et al., 2009). This is one of many knowledge gaps that, if answered, could supplement efforts to combat massive SSB marketing campaigns. It calls into question, however, what research has already been done and what knowledge gaps still exist about advertising’s effects on young consumer behavior.

Another type of advertising investigated by researchers is alcohol advertising. Although alcohol and SSB advertising market different products, research in this field contributes a lot to our general understanding of the psychological mechanisms of advertising. For example, a study testing advertising’s effects on alcohol consumption has displayed a causal link between exposure to alcohol portrayal and alcohol consumption. The participants who were exposed to alcohol consumption in advertisements consumed, on average, 1.5 more glasses than those who did not. The authors concluded that this was the result of advertising’s ability to increase brand salience, or the degree to which a brand is thought of or noticed (Engels et al., 2009). Increased brand salience is believed to improve consumer attitudes toward a product, encouraging increased probability of purchase and consumption (Miller & Berry, 1998). However, the authors also suggested that the increased consumption could have been the product of the concept of unconscious mimicry, the idea that people view others consuming a substance might unknowingly repeat that behavior. (Engels et al., 2009). This mimicry, known as the “chameleon effect,” suggests that humans are likely to repeat or mimic behavior they witness in other humans, more so than those who do not observe that behavior. An article investigating the “chameleon effect” finds that it is the result of an unconscious perception-behavior link and is inherent in people (Chartrand & Bargh, 1999).

A study of the association of exposure to SSB advertising and SSB consumption also suggested mimicry as a mechanism for increased consumption. The study, which exposed participants to either SSB or neutral advertising, tracked participants’ consumption and found that more SSBs were consumed after exposure to SSB advertising than neutral advertising. This study, however, was performed on adult Dutch test subjects, not on US teens, who the authors suggested as a possible future test sample (Koordeman et al., 2010). A similar study conducted in the UK by Halford et al. used child participants, but instead exposed them to snack advertisements, provided participants with snack options, and then tracked consumption (2004). This study focused on the role of advertisement recognition in the relationship between exposure and consumption and tested recognition by having participants identify the ads they were exposed to. While the study found that all participants ate more after watching the food advertising and that there was a positive association between advertising recognition and consumption, the authors stated that the results were not enough to establish a causal link. While this may be due to the geographically limited sample of participants or the small sample size, no reasons were explicitly stated (Halford et al., 2004).

In my study, I intend to use a similar design to that of Kemps et al. by exposing participants to advertisements for either SSBs or neutral beverages. After watching these advertisements, they will then sample four different beverages (two of which will include SSBs featured in the advertisements), and the amount of each beverage consumed will be measured. I will use ads for the same four SSB companies used in the study (eg. Coca-Cola, Fanta, Pepsi, and Sprite), and I will provide the same beverages to participants in the taste test (orange juice, mineral water, Coca-cola, and Sprite). Why the specific brands were chosen to be advertised and presented in the taste test was not indicated in the study, aside from the fact that two of the taste test beverages are SSBs and two are not (2021). The participants sampled the beverages under the impression that they were conducting a taste test. This “bogus taste-test” method has been analyzed and verified as a valid measure of food-intake. Previously it has been successfully used to investigate the association between food consumption and traits such as liking of taste and baseline hunger (Robinson et al. 2017). Kemps et al. also had participants complete tasks to measure other factors that might also be associated with consumption (2021). However I do not intend to use those tasks in my study, due to technological limitations.

The three advertising-consumption studies I cited (Chartrand & Bragh, Koordeman et al., & Halford et al.) all display higher levels of consumption after exposure to advertising, as a result of participants imitating the consumption displayed in soft drink advertisements. Therefore, teens exposed to soft drink advertising will consume more of the SSBs than those exposed to the neutral advertising. However, none of the aforementioned studies used US teens as participants, which my study would. Establishing a causal link between television advertising and SSB consumption in US teens could further establish the underlying causes of frequent SSB consumption, and better understand how high levels of advertisement exposure affect young consumers. Defining mimicry as a mechanism could allow for research of advertising without drinking behaviors and restriction of drinking imagery in advertising to mitigate the effects of advertisement exposure and decrease soft drink consumption.

Subject Recruitment and IRB Approval:

This study received IRB approval from The Berkeley Carroll Institutional Review Board (see IRB application here). This study involves the consumption of four different kinds of soft drinks, some of which contain sugar and caffeine. Subjects may have adverse reactions to these substances, however they were informed about this multiple times, including prior to agreeing to participate in the study, so those with allergies or other reactions were screened and are able to cease participation. The subjects were also exposed to soft drink advertising during the testing session, linked here. Participants were informed about the ad exposure multiple times prior to the testing sessions, so those with serious negative relationships with the advertising were not forced to participate and are allowed to cease participation.

This study recruited teenage students from the Berkeley Carroll Upper School as participants. I hope to recruit at least 60 students, with a maximum of 100, due to the amount of materials. The sample frame of American young adults differs from that of the study this research is modeled after, which uses Australian young adults (Kemps et al, 2021). While there is no age difference, the difference in nationality may affect the results. The US is the leading consumer of soft drinks, with more than 20% of young adults consuming soda daily (Euromonitor International, 2019, Miller et al., 2017). Thus, this study may show greater total consumption of soft drinks. However, this hopefully does not affect the effects of advertising. Students were informed about the research through a schoolwide email. In the email, they were notified that some of the beverages consumed in the study contain sugar and caffeine. They were also asked to indicate interest in participating, after which their emails were recorded to contact them with the date and time of their testing session. This email list was randomly divided in half to determine the experimental and control groups. The two groups were split into smaller groups of equal number, with no more than 15 people per group for testing. After each smaller group was emailed with the time for their session, the email list was no longer used for anonymity purposes, save for possible rescheduling.

Experimental Design

The experimental design for this study replicates that of Kemps et al.’s study on soft drink exposure and soft drink consumption, attentional bias, and inhibitory control (2021).

Informed Consent, Pre-Survey, and Viewing Sessions:

Participants were seated in a classroom environment and provided with packets containing all of the surveys they were to fill out over the course of the experiment separated by a blank page. While participants may have been seated at the same tables, dividers were employed to make sure participants are not influenced by one another’s survey responses. After giving informed consent, participants completed a pre-survey, where they reported their current thirst, their general enjoyment of soft drinks, which soft drinks they normally consume, and when and why they consume soft drinks. These questions are intended to provide the researcher with information regarding other possible causes of increased soft drink consumption, like thirst and habitual consumption. Subjects that report a high level of either characteristic had their results noted in data collection. They received an anonymous alias number at their seat to record on their survey packet. This number matched the number on their beverage tray later in the experiment and ensured anonymity of all respondent data.

After completing the pre-survey, participants viewed a premade 2.5 minute block of advertisements for various beverages. Exposure to soft drink advertising, the independent variable, is not quantified in this experiment, rather it is manipulated by either being presented, as in the experimental sessions, or not presented, as in the control sessions. The experimental groups viewed an advertisement each for Sprite, Coca-Cola, Fanta, and Pepsi. These are all widely consumed and popular soft drink brands, according to a report by Brand Finance (2021). This report lists these brands as part of the top 15 of 2021, and although this statistic comes from a brand awareness firm, their website states they are respected and cited by many annual reports and collect data on thousands of consumer brands. The control groups viewed advertisements for four neutral, non-sugar-sweetened beverages: Fairlife milk, Hint flavored water, Ocean Spray fruit juice, and Vita Coco coconut water. All groups also viewed an advertisement for Bubly non-sugar sweetened sparkling water. A common advertisement for sparkling water was used in the original study to limit “demand effects.” Demand effects, or demand characteristics, are biases in results that occur when participants become aware of an experiment’s purpose, and thus behave in a way that confirms the researcher(s)’s hypothesis (Mummolo, 2019). In this case, adding an advertisement for a non-sugar-sweetened beverage placed among a group of advertisements for sugar-sweetened beverages reduces demand effects by making it less obvious to participants that they are viewing advertisements for a very specific type of beverage. With one exception, the types of beverages advertised were consistent with those used in Kemps et al., as well as the common sparkling water advertisement. The brands advertised in the experimental branch matched those used in the study exactly. However, the non-sugar-sweetened beverages advertised in the study were Australian beverages that are not widely sold in the US, so American brands were used instead. In addition, an advertisement for coconut water was used rather than one for cordial. Cordial is a concentrated beverage used as a mixer for nonalcoholic and alcoholic drinks, and the brand advertised in the original study is an Australian, unsweetened variety (Australian Beverage Council, 2021). Because the brand advertised was Australian and this study uses US teens as a sample, coconut water was chosen as a replacement due to its increasing popularity as an unsweetened and healthy beverage (Cappelletti et al., 2015). I do not anticipate this change to have any major effects, as both advertisements serve the same purpose, promoting an unsweetened beverage. After the viewing session, participants were asked to complete a short survey on the likeability of the advertisements. This survey was originally used in a pilot study to identify which ads viewers might enjoy watching, but it has been adapted in this study to further reduce demand effects. It does so by drawing attention to the advertisements that the participants are being exposed to, but giving them the impression that their ratings of the ads are being tracked, rather than other behaviors measured later in the study.

Mock Taste-Test

After completing the likeability survey the participants were presented with transparent, 250 mL glasses—sealed with transparent lids to prevent spillage—of four different drinks: mineral water, orange juice, Sprite, and Coca-Cola. These four beverages are the same ones presented to participants in Kemps et al. (2021). Each tray had an alias number on it corresponding to the number at the participant’s seat. The drinks were presented in a row, the order of which was randomized from tray to tray using a 4x4 Latin square (See procedure for details) (Dean, 2001). Participants were given 10 minutes to taste-test the beverages. This is a mock taste test intended to measure participants’ consumption of the beverages they are tasting, rather than their opinions of the beverages. They were provided with a rating sheet for each of the beverages in their survey packet and were encouraged to fill it out. The validity of this mock taste test as a tool to accurately measure consumption when manipulated by another variable has been confirmed by multiple studies (Robinson, 2017). In place of the attentional bias task in the study, which cannot be used due to technological restrictions, participants were asked to report the order in which they first taste each drink by numbering the first drink they taste 1, and so on with a sharpie. An inhibitory control task was not used for the same reason. After the taste test, the participants were dismissed and their data collected. Their data was collected based on the corresponding alias number they received at the beginning of the study. Drink consumption was measured per beverage by measuring the remaining beverage left in a participant’s cup after the taste test and calculating the difference between that and the initial amount using a 250 mL graduated cylinder (recorded in milliliters).

The control group serves as a negative control in this experiment, providing data for what an average teenager might consume in the absence of any soft-drink advertising. There is no positive control for this experiment as there is currently no available condition that produces an expected result.

Data Analysis

Data was grouped into experimental and control, not considering specific testing session date.

For the pre survey, an average of responses to the 5-point Likert scale was taken for each condition. The total number of responses to each answer on the multiple choice questions was taken and converted to a percentage of total responses per condition. Each of the written responses were recorded, and any overlap was noted. The pre survey data was collected to check for possible correlations.

The responses from the Likert scales on the likeability survey were not recorded as the purpose of the likeability survey is mainly to decrease demand effects during the mock taste test.

The amount of beverage consumed and the order of beverages first tested were collected and associated with the alias number of the tray and survey packet they were presented with. Then, average beverage consumption per condition was calculated in mL, first by individual beverage, then by general beverage type (flat: orange juice and water; soft: Sprite and Coca Cola). In addition, the average total beverage consumption was calculated per experimental condition. This allowed for examination of if either group drank significantly more of any beverage on average. The order in which beverages were chosen was inversely translated to a 4-1 point scale (1st choice = 4pts). Each beverage’s average “choice score” was calculated per experimental condition and compared to see if either group displayed significant bias in beverage choice.

After each trial, participants’ consumption of each beverage (in mL) was recorded using a graduated cylinder. Participants’ beverage choice was determined and recorded based on the number order they indicated on their cups.

To avoid bias for beverage choice, the order of beverages on the tray was randomized using a 4x4 Latin Square (Figure 1). To quantify beverage choice, each participant received a choice score for each beverage, based on the order they consumed them. The beverage chosen first received 4 points, the second received 3 points, the third received 2 points, and the last received 1 point. The sum of the scores for the two flat drinks and the two soft drinks were calculated for each participant as well. The average choice scores for flat and soft drinks were compared across experimental conditions. The average choice score for flat and soft drinks, respectively, was 4.55 and 5.45 in the experimental group, and 4.94 and 5.06 in the control group. In both conditions, a 2-tailed correlated t-test found no statistical significance between the average choice score for soft drinks and for flat drinks. This does not show that subjects chose randomly, however, in fact these results are attributed to the high variability and the small sample size of the study, thus the data was not graphed. It is also possible that the 1-4 scale used to quantify choice preference was ineffective. Future studies could be conducted using this method, but with a much larger sample size to confirm whether it is viable for quantifying choice. However, it should be noted that this study was unique in its 1-4 scale rating, as well as how choice was tested during the mock taste test. This method was chosen due to an inability to access the software used to test and quantify beverage choice in Kemps et al. 2021. Studies with access to that software and sufficient participants should replicate the choice test results reported by Kemps et al. (2021)

WORKS CITED:

Australian Beverage Council (2021). Cordial. Australian Beverages. https://www.australianbeverages.org/beverages/cordial/

Calvert, S (2008).The Future of Children , Spring, 2008, Vol. 18, No. 1, Children and Electronic Media (Spring, 2008), pp. 205-234. https://www.jstor.org/stable/20053125

Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception–behavior link and social interaction. Journal of Personality and Social Psychology, 76(6), 893–910. https://doi.org/10.1037/0022-3514.76.6.893

Dean, A.M. (2001). Experimental Design: Overview, International Encyclopedia of the Social & Behavioral Sciences, Pergamon, 5090-5096, ISBN 9780080430768, https://doi.org/10.1016/B0-08-043076-7/00417-4. (https://www.sciencedirect.com/science/article/pii/B0080430767004174)

Engels, R. C., Hermans, R., van Baaren, R. B., Hollenstein, T., & Bot, S. M. (2009). Alcohol portrayal on television affects actual drinking behaviour. Alcohol and Alcoholism, 44(3), 244–249. https://doi.org/10.1093/alcalc/agp003

Euromonitor International (2019). Passport global market information database. https://go.euromonitor.com/factsheet-soft-drinks.html

Giacaman, R. A., Pailahual, V., & Díaz-Garrido, N. (2018). Cariogenicity induced by commercial carbonated beverages in an experimental biofilm-caries model. European Journal of Dentistry, 12(01), 027–035. https://doi.org/10.4103/ejd.ejd_188_17

Halford, J. C. G., Gillespie, J., Brown, V., Pontin, E. E., & Dovey, T. M. (2004). Effect of television advertisements for foods on food consumption in children. Appetite, 42(2), 221–225. https://doi.org/10.1016/j.appet.2003.11.006

Harris, J. L., & Graff, S. K. (2012). Protecting young people from junk food advertising: Implications of psychological research for first amendment law. American Journal of Public Health, 102(2), 214–222. https://doi.org/10.2105/ajph.2011.300328

Harris, J. L., & Kalnova, S. S. (2018). Food and beverage TV advertising to young children: Measuring exposure and potential impact. Appetite, 123, 49–55. https://doi.org/10.1016/j.appet.2017.11.110

Harris, J. L., Pomeranz, J. L., Lobstein, T., & Brownell, K. D. (2009). A crisis in the marketplace: How Food Marketing contributes to childhood obesity and what can be done. Annual Review of Public Health, 30(1), 211–225. https://doi.org/10.1146/annurev.publhealth.031308.100304

Hu, F. B. (2013). Resolved: There is sufficient scientific evidence that decreasing sugar-sweetened beverage consumption will reduce the prevalence of obesity and obesity-related diseases. Obesity Reviews, 14(8), 606–619. https://doi.org/10.1111/obr.12040

Kemps, E., Tiggemann, M., & Tuscharski, A. (2021). The effect of television advertising on soft drink consumption: Individual vulnerabilities in approach bias and inhibitory control. Appetite, 165, 105300. https://doi.org/10.1016/j.appet.2021.105300

Kit, B. K., Fakhouri, T. H. I., Park, S., Nielsen, S. J., & Ogden, C. L. (2013). Trends in sugar-sweetened beverage consumption among youth and adults in the United States: 1999–2010. The American Journal of Clinical Nutrition, 98(1), 180–188. https://doi.org/10.3945/ajcn.112.057943

Koordeman, R., Anschutz, D. J., van Baaren, R., & Engels, R. C. M. E. (2010). Exposure to soda commercials affects sugar-sweetened soda consumption in young women. an observational experimental study. Appetite, 55(1), 166. https://doi.org/10.1016/j.appet.2010.05.056

Lewis, M. K., & Hill, A. J. (1998). Food Advertising on British children’s Television: A content analysis and experimental study with nine-year olds. International Journal of Obesity, 22(3), 206–214. https://doi.org/10.1038/sj.ijo.0800568

Malik, V. S., Pan, A., Willett, W. C., & Hu, F. B. (2013). Sugar-sweetened beverages and weight gain in children and adults: A systematic review and meta-analysis. The American Journal of Clinical Nutrition, 98(4), 1084–1102. https://doi.org/10.3945/ajcn.113.058362

Martina Cappelletti, Giovanna Ferrentino, Isabella Endrizzi, Eugenio Aprea, Emanuela Betta, Maria Laura Corollaro, Mathilde Charles, Flavia Gasperi, Sara Spilimbergo (2015). High Pressure Carbon Dioxide pasteurization of coconut water: A sport drink with high nutritional and sensory quality. Journal of Food Engineering, 145, 73-81, ISSN 0260-8774, https://doi.org/10.1016/j.jfoodeng.2014.08.012. https://www.australianbeverages.org/beverages/cordial/

Miller, S., & Berry, L. (1998, September-October). Brand Salience versus Brand Image: Two Theories of Advertising Effectiveness. Journal of Advertising Research, 38(5), 77. https://link.gale.com/apps/doc/A54069376/AONE?u=nysl_oweb&sid=googleScholar&xid=cb5d8123

Miller, G., Merlo, C., Demissie, Z., Sliwa, S., & Park, S. (2017). Trends in beverage consumption among high school students — United States, 2007–2015. MMWR. Morbidity and Mortality Weekly Report, 66(4), 112–116. https://doi.org/10.15585/mmwr.mm6604a5

MUMMOLO, J., & PETERSON, E. (2019). Demand Effects in Survey Experiments: An Empirical Assessment. American Political Science Review, 113(2), 517-529. doi:10.1017/S0003055418000837

Nielsen. (2021). The Nielsen Total Audience Report: Advertising Across Today's media. Nielsen. (n.d.). Retrieved November 12, 2021, from https://www.nielsen.com/us/en/insights/report/2021/total-audience-advertising-across-todays-media/.

Robinson, E., Haynes, A., Hardman, C. A., Kemps, E., Higgs, S., & Jones, A. (2017). The bogus taste test: Validity as a measure of laboratory food intake. Appetite, 116, 223–231. https://doi.org/10.1016/j.appet.2017.05.002

Schiestl, E. T., Rios, J. M., Parnarouskis, L., Cummings, J. R., & Gearhardt, A. N. (2021). A narrative review of highly processed food addiction across the lifespan. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 106, 110152. https://doi.org/10.1016/j.pnpbp.2020.110152

Small, D. M., & DiFeliceantonio, A. G. (2019). Processed Foods and food reward. Science, 363(6425), 346–347. https://doi.org/10.1126/science.aav0556

Sugary Drink F.A.C.T.S. (2020). Sugary drink advertising to youth: Continued barrier to public health progress. UConn Rudd Center for Food Policy & Obesity . https://media.ruddcenter.uconn.edu/PDFs/Sugary_Drink_FACTS_Full%20Report.pdf

Vartanian, L. R., Schwartz, M. B., & Brownell, K. D. (2007). Effects of soft drink consumption on nutrition and health: A systematic review and meta-analysis. American Journal of Public Health, 97(4), 667–675. https://doi.org/10.2105/ajph.2005.083782

Volkow, N. D., Wise, R. A., & Baler, R. (2017). The dopamine motive system: Implications for drug and food addiction. Nature Reviews Neuroscience, 18(12), 741–752. https://doi.org/10.1038/nrn.2017.130