Evidence of harm

The impact of food marketing on dietary behaviour and its antecedents is a function of both exposure to the marketing (the reach and frequency of the marketing message) and its persuasive power (the creative content of the marketing)⁽¹⁸⁾. There is considerable evidence that children are exposed to vast quantities of food marketing across multiple media and settings including television^{(Reference Kelly, Vandevijvere and Ng19)}, digital media⁽²⁰⁾, cinemas^{(Reference Wong, Pauzé and Hatoum21)}, sports^{(Reference Dixon, Lee and Scully22)}, outdoors^{(Reference Finlay, Robinson and Jones23)} and even in schools^{(Reference Kelly, Clerkin and Nic Gabhainn24)}. This marketing predominantly promotes foods that contribute to unhealthy diets, with proportions of unhealthy food advertisements typically ranging from 50 to greater than 90 % of all food advertising^{(Reference Boyland and McGale10)}. Globally, the most frequently marketed food categories include fast-food, sugar-sweetened beverages, chocolate and confectionery, sugary breakfast cereals and snack foods^{(Reference Boyland and McGale10,Reference Kelly, Vandevijvere and Ng19)} . There is some evidence of ethnic and socio-economic inequalities in levels of exposure to food marketing and the nature of the foods promoted, such that children from minority or disadvantaged backgrounds are disproportionately exposed to unhealthy food advertising^{(Reference Finlay, Robinson and Jones23,Reference Backholer, Gupta and Zorbas25)} .

Studies also demonstrate that the food marketing to which children are exposed uses a variety of creative strategies that are likely to appeal to, and resonate, with child audiences. These include promotional characters, celebrity endorsements, branding, visual imagery, novel designs, animation, dynamic elements and special effects^{(Reference Boyland and McGale10)} as well as digital engagement tactics, events, games and incentives^{(Reference Truman and Elliott26)}. Other features unique to digital media such as ‘geo-tags’ (the addition of geographical identification metadata to media files such as photos) and interactivity are also increasingly commonplace in marketing targeted at children^{(Reference Bragg, Pageot and Amico27)}.

The first peer-reviewed evidence of the effects of food advertising on eating behaviour in children emerged from the US in the late 1970s and early 1980s^{(Reference Goldberg, Gorn and Gibson28,Reference Gorn and Goldberg29)} . Since then, these findings have been corroborated and effects further characterised in studies captured within multiple large-scale systematic reviews and meta-analyses^{(Reference Boyland, Nolan and Kelly30–Reference Cairns, Angus and Hastings35)}. The most recent global evidence review, conducted to inform updated WHO global guidelines on restrictive food marketing policies, found that food marketing was associated with significant increases in children's food intake, choice and preference towards advertised/unhealthy foods as well as purchase requests by children to caregivers^{(Reference Boyland, McGale and Maden36)}. Specific marketing techniques, such as promotional characters, have also been found to have significant effects on taste preference for unhealthy products, food liking and snack choice^{(Reference Packer, Russell and McLaren37)}. Some analyses have also found a moderating effect of BMI, such that children living with overweight or obesity consumed significantly more kilocalories following food advertisement exposure relative to those of a healthy weight^{(Reference Russell, Croker and Viner33)}. Evidence for socio-economic differences in the impact of food marketing is more equivocal^{(Reference Backholer, Gupta and Zorbas25)}.

Based on a systematic review of the evidence, largely for the effects of television advertising, Kelly et al.^{(Reference Kelly, King and Chapman38)} proposed a hierarchical model to indicate the ways in which food marketing affects a range of outcomes pertinent to diet and health. The model posits that there is a sequential (although recurrent and reinforcing) set of effects of food marketing exposure on intermediate outcomes (such as purchase and consumption) and, later, weight-related outcomes. Longer term outcomes can potentially be mediated by both the direct (brand awareness) and indirect (brand attitudes) effects of marketing^{(Reference Kelly, King and Chapman38)}. The validity of the model has been supported by analyses of cross-sectional data which have shown indirect associations between commercial television exposure and children's BMI via purchasing and with consumption via purchase requests^{(Reference Boyland, Muc and Kelly39)}. Importantly, the model discredits the notion that there is a simple, direct link between exposure to food marketing and the development of obesity. However, subsequent studies have demonstrated that children's additional energy consumption following food marketing (relative to non-food marketing) is not compensated for at subsequent eating occasions and, as such, would logically contribute to weight gain over time^{(Reference Norman, Kelly and McMahon40)}. The evidence for a relationship between food marketing and obesity also meets established epidemiological criteria for causality^{(Reference Norman, Kelly and Boyland11)}.

Based on an effect size for the impact of food marketing on short-term food consumption in children^{(Reference Boyland, Nolan and Kelly30)}, Brown et al.^{(Reference Brown, Ananthapavan and Veerman41)} were able to demonstrate, through modelling, that introducing restrictions on unhealthy food advertising on television in Australia (a 21⋅30 h ‘watershed’) would reduce children's energy intake by an average of 115 kJ/d (approximately 115⋅06 kJ (27⋅5 kcal)) and BMI by approximately 0⋅35 kg/m² with greater benefits from the most disadvantaged children. Similarly, a 21⋅00 h watershed for unhealthy food advertising on television in the UK is predicted to decrease energy intake by an average of 38.07 kJ (9⋅1 kcal) daily which could reduce the number of UK children with overweight (including obesity) by 3⋅6 %^{(Reference Mytton, Boyland and Adams42)}.

Unique vulnerability of children

Children are thought to be particularly vulnerable to the effects of food marketing. This vulnerability stems from several different factors, outlined below.

First, young people live in a media-saturated environment of which most channels are commercial (i.e. carry marketing)^{(Reference Story and French17)}. Children are avid users of media, and rapidly uptake new media forms. In 2020, nearly all UK children (5–15 years) went online (mostly via laptops, tablets and mobile phones), 56 % watched live TV and almost all (96 %) watched TV programmes via video-on-demand services⁽⁴³⁾. In the UK, use of social media apps, video-sharing platforms and live streaming platforms is already established in many children aged 5–7 years and grows rapidly between the ages of 5 and 15 years⁽⁴³⁾.

Secondly, while using these media, children are heavily targeted by food marketers because they spend substantial amounts of their own money annually, influence the spending of even more via household food purchases and are adult consumers of the future^{(Reference Story and French17)}. Investment in marketing that reaches children is an investment in brand awareness, recognition, preference and loyalty which are all believed to precede or drive purchase behaviour^{(Reference McNeal44)}.

Thirdly, children may be more readily deceived and exploited by food marketing based on their neurocognitive and developmental immaturity. Communications research in the 1970s focused on children's ability to recognise the selling and persuasive intent of marketing messages. That is, studies sought to understand the age at which children are able to recognise that the ultimate goal of advertising (from the perspective of the marketer) is to achieve product sales, and that the marketing message may enhance positive characteristics or attributes of a product and withhold negative ones in order to be persuasive^{(Reference Harris, Brownell and Bargh45)}. This research posited that possessing such cognitive abilities would confer an ability to resist the marketing and be therefore inoculated from any harmful effects. However, this rationale has several limitations for understanding food marketing in the modern environment. Much of the evidence on advertisement recognition stems from research using television advertisements that can be far more readily distinguished from other television content (i.e. programming) even by young children than even simple digital marketing can^{(Reference Ali, Blades and Oates46)}. Social media influencer marketing further blurs the boundaries between food marketing and entertainment content^{(Reference Coates, Hardman and Halford47)} and exploits children's trust and the parasocial relationships they develop with online personalities^{(Reference Coates, Hardman and Halford48)}. Also, it is increasingly clear that food marketing does not operate through conscious, rational arguments and persuasion but rather by subconscious routes and implying (exaggerated) emotional or social benefits from consumption^{(Reference Folkvord, Anschütz and Boyland49)}. A more recent framework for understanding children's vulnerability, the food marketing defense model, incorporates the emotional component of marketing and dispels the idea that recognition and understanding alone are sufficient for protection, positing that cognitive resources and a motivation to resist are also required^{(Reference Harris, Brownell and Bargh45)}. This resistance would therefore be challenging for all children (including adolescents) given the levels of exposure and the use of powerful, salient and persuasive techniques they experience (outlined in the previous section), particularly at a time when their neurological development is incomplete (so they are vulnerable to ‘risky’ decisions) and they are seeking to establish their social identities and desiring conformity and acceptance from peers^{(Reference Blakemore and Robbins50,Reference Knoll, Magis-Weinberg and Speekenbrink51)} .

Limiting children's screen time to reduce risk of harm to this vulnerable group may appear to be a pragmatic solution for concerned parents, but is it ethical to restrict children's access to the opportunities and benefits the internet offers? The internet offers unrivalled opportunities for learning, civic participation, creativity and communication⁽⁵²⁾ so facilitating children's access and engagement with this media could be considered essential for their growth and development in the digital era. However, there is a responsibility of States to ensure that this participation is possible without harming their health.

Child rights

High in fats, salt and/or sugar marketing affects numerous child rights enshrined within the UN CRC including the right to the enjoyment of the highest attainable standard of health, the right to adequate food and the right to privacy⁽⁵³⁾. The CRC recognises the unique sensitivities of children and that they have a need for special protection. Importantly, governments that ratify the convention (such as the UK) are legally bound to uphold the commitments enshrined in this international human rights treaty. In addition, the UN Guiding Principles on Business and Human Rights assert that particular attention should be paid to the rights and needs of groups or populations that may be at elevated risk of becoming vulnerable or marginalised⁽⁵⁴⁾. Given their developing neurocognitive and social capacities (as described earlier), it is logical to interpret that children might be such a group⁽⁵³⁾.

The ways in which unhealthy food marketing negatively affects rights to health and food have been fully explored elsewhere^(13,20) . In brief, Article 24 of the CRC⁽⁵³⁾ requires that ‘States Parties recognize the right of the child to the enjoyment of the highest attainable standard of health’ and that they ‘take appropriate measures to diminish infant and child mortality [and] to combat disease and malnutrition, through, inter alia, the provision of adequate and nutritious foods’. In light of the substantial evidence linking unhealthy food marketing to poorer diet and health outcomes (summarised in a previous section of this paper), it can be argued that States, as part of their duty to respect, protect and fulfil children's rights under the CRC, should restrict such marketing in order to reduce these negative impacts on children's health⁽¹³⁾.

It has also been asserted that food marketing through digital media has substantially amplified the harms to children's wellbeing and raised new issues of rights violations surrounding privacy and freedom from exploitation^{(Reference Tatlow-Golden and Garde55)}. Article 16 of the CRC⁽⁵³⁾ states that ‘no child shall we subjected to arbitrary or unlawful interference with his or her privacy’ and that ‘the child has the right to the protection of law against such interference’. Human rights apply online as offline, and therefore in the modern era this article can be understood to include a right to the protection of personal data^{(Reference Tatlow-Golden and Garde55)}. The digital advertising ecosystem is built on the extraction of personal data and the use of these data to deliver personalised and micro-targeted marketing^{(20,Reference Tatlow-Golden and Garde55)} that is highly effective and impactful^{(20,Reference Buchanan, Kelly and Yeatman56)} . Article 36 of the CRC⁽⁵³⁾ requires that States ‘protect children from ‘all other forms of exploitation’’ (i.e. those not covered by other CRC articles). It has been argued that food marketing is inherently exploitative and manipulative^{(Reference Harris, Brownell and Bargh45)} as it seeks to manipulate emotion and bypass rational decision-making^{(Reference Folkvord, Anschütz and Boyland49)}. Digital food marketing, in particular, can be said to be exploitative of young people's incredulity due to its use of highly salient creative content^{(Reference Boyland and McGale10)}, and the embedded, immersive and interactive ways in which the marketing is delivered⁽²⁰⁾ that often are ‘under the radar’ or beyond consumers' awareness^{(Reference Tatlow-Golden and Garde55)}.

Because of the multiple ways in which food marketing and child rights intersect, it has been proposed that a ‘child rights-based' approach to the regulation of unhealthy food marketing offers a powerful and universally applicable way to consider children as rights holders who are central to any policy discourse’ (p. 30)⁽⁵³⁾.

The ‘responsible’ food industry and the self-regulation of advertising

WHO and other leading health organisations^{(Reference Kraak, Zhou and Rincón-Gallardo Patiño57)} have long asserted that, based on the afore-mentioned evidence of harm and of child rights violations, there is a clear need for robust and effective restriction of the unhealthy food marketing to which young people are exposed⁽⁵⁸⁾. The original ‘set of recommendations on the marketing of food and non-alcoholic beverages to children’ noted that States should ‘identify the most suitable policy approach’ (p. 14)⁽⁵⁸⁾. Evaluation of progress in implementing these recommendations noted that as of 2018, just 54 % of the countries in the WHO European region had taken steps to limit the marketing of unhealthy foods and a majority of those steps reflected industry self-regulatory approaches⁽¹³⁾. Substantial evidence has accrued in recent years to demonstrate that mandatory or ‘government-led’ policies are more effective than industry self-regulation^{(Reference Boyland, McGale and Maden8,Reference Chambers, Freeman and Anderson59,Reference Kovic, Noel and Ungemack60)} . For example, in a recent global evidence review, mandatory policies (v. no policy) were associated with a greater proportion of desirable (for public health) than undesirable effects on exposure, power, purchasing and unintended consequences whereas the opposite was found for industry voluntary measures^{(Reference Boyland, McGale and Maden8)}. Even the enforcing agencies have sometimes identified the lack of effectiveness of these measures. Evaluations of the UK's Advertising Standards Authority's advertising code that covers online food advertising, conducted by the Advertising Standards Authority themselves using automated ‘avatars’ to trawl the web using data profiles mimicking children, identified 947 code breaches in 2 weeks on child-focused YouTube channels in 2018⁽⁶¹⁾ with further breaches found in a 2020 follow-up⁽⁶²⁾.

Industry-derived or co-regulatory codes often place emphasis on ensuring marketing communications are ‘legal, decent, honest and truthful’⁽⁶³⁾ or do not ‘mislead consumers’⁽⁶⁴⁾ suggesting a belief that it is possible for food marketing to children to be decent and to not mislead if it is constructed in a particular way. Related to this, an analysis of a public–private partnership launched in 2011 between the UK Government, industry and other organisations in the areas of food, alcohol, physical activity and health at work known as the ‘public health responsibility deal’ is illuminating in what it reflects about the integrity of industry commitment to meaningful change. The analysis by Knai et al. found that the most effective strategies to improve diet (including restrictions on marketing) were not reflected in the pledges made by the industries involved in the deal. Rather, most of the pledges were offers to provide information, raise awareness and communicate with consumers – actions that are known to have limited to no effect on positive behaviour change^{(Reference Knai, Petticrew and Douglas65)}. It was noted that ‘both the production and uptake of pledges by Responsibility Deal partners were largely driven by the interests of partners themselves, enabling these wider [alcohol, food, etc] systems to resist change’ (p. 1)^{(Reference Knai, Petticrew and Douglas65)}.

The underpinning reasons for the preference from food and advertising industries for self-regulation above mandatory policies are therefore apparent in terms of the likely (lack of) impact on business models and revenue. There are also additional benefits to being involved in such schemes including avoiding or delaying government intervention and reaping the reputational benefits of involvement^{(Reference Ronit and Jensen66)}.

Ethics of monitoring children's exposure to digital marketing for public health research

Monitoring children's exposure to digital marketing is challenging⁽²⁰⁾, but multiple methodologies have emerged in recent years^{(Reference Tatlow-Golden, Jewell and Zhiteneva14,Reference Bica, Wickramasinghe and Zhiteneva67)} that are beginning to yield hugely informative data on the extent and nature of this exposure^{(Reference Kelly, Bosward and Freeman68–70)}. There are also ethical considerations inherent in conducting this monitoring for the purposes of informing policy design. These have been discussed extensively elsewhere^{(Reference Tatlow-Golden, Jewell and Zhiteneva14,Reference Tatlow-Golden, Verdoodt and Oates71)} and are summarised briefly here.

Under the EU General Data Protection Regulations, participants have certain rights in relation to the processing of their personal data for research purposes that researchers should be aware of, including a right to be informed and notice requirements. Whether or not internet data should be considered public or private, and therefore the extent to which it is reasonable for a researcher to process it without users' consent or awareness, is a challenging and contentious issue^{(Reference Tatlow-Golden, Verdoodt and Oates71)}. For example, even if social media data are legally public according to the terms of the platform it relates to, users may have reasonable expectations of privacy that may reflect the settings they have applied to their account and the requirement for sign-in. Even if one user consents to research participation, there are also ethical issues inherent in the concurrent collection of data from others in their networks who did not consent^{(Reference Tatlow-Golden, Verdoodt and Oates71)}. Monitoring digital food marketing exposure may also yield data on sensitive topics, e.g. young people's engagement with content on eating disorders or alcohol. Further, it is possible that researchers could uncover evidence of criminal activity or other safeguarding concerns that may require them to breach confidentiality conditions by informing the relevant authorities. Security procedures in data processing (including storage, management and sharing) must be fundamental to digital monitoring involving children, as must the information participants (and their gatekeepers) are provided with, their understanding of that information (age-appropriate provision should be made) and the processes used to determine if they have a genuine willingness to participate that has been indicated free from coercion^{(Reference Tatlow-Golden, Verdoodt and Oates71)}.