Visibility of age restriction warnings, harm reduction messages and terms and conditions: a content analysis of paid-for gambling advertising in the United Kingdom

Objective: The inclusion and design of age restriction warnings, harm reduction messages and terms and conditions (T & Cs) in gambling advertising is self-regulated in the United Kingdom. Our study examines the visibility and nature of this information in a sample of paid-for gambling adverts. Study design: A content analysis of a strati ﬁ ed random sample of gambling adverts ( n ¼ 300) in the United Kingdom from eight paid-for advertising channels (March 2018). Methods: For each advert, we assessed whether any age restriction warnings, harm reduction messages and T & Cs were present. If so, visibility was scored on a ﬁ ve-point scale ranging from very poor ( (cid:2) 10% of advert space) to very good ( (cid:3) 26% of advert), which had high inter-rater reliability. Descriptive information on position, design and tone of language was recorded. Results: One in seven adverts (14%) did not feature an age restriction warning or harm reduction message. In adverts that did, 84% of age restriction warnings and 54% of harm reduction messages had very poor visibility. At least one in ten adverts did not contain T & Cs. In adverts that did, 73% had very poor visibility. For age restriction warnings, harm reduction messages and T & Cs, most appeared in small fonts and outside the main advert frame. Most harm reduction messages did not actually reference gambling-related harms. Conclusion: Age restriction warnings, harm reduction messages and T & Cs do not always appear in paid-for gambling advertising. When they do, visibility is often very poor and the messaging not clear. The ﬁ ndings do not support a self-regulatory approach to managing this information in gambling adverts. This


Introduction
While gambling is a popular recreational activity for many people, for some it can lead to significant problems that affect both the individual and others around them. 1e3 There is wide variation in past-year problem gambling rates in different countries across both Europe (0.12e3.4%) and the world (0.12e5.8%). 4 In the United Kingdom (UK), an estimated 430,000 adults (approximately 0.7% of population) are problem gamblers and a further two million (3% of population) are at risk of gambling-related problems. 5 Understanding the drivers of such behaviour, and opportunities for prevention, is important given that higher risk gambling is associated with individual, social and economic concomitant harms. 6 In the UK, and elsewhere, there are calls for a comprehensive harm reduction strategy that reflects the products, environment and marketing that shapes gambling behaviour. 1,7e9 Given the reported links between marketing and gambling behaviour, 10 including consumer protection information in advertising is a low-cost and high-reach intervention that may help to mitigate gambling-related harm. Such information can include age restriction warnings, harm reduction messages and terms and conditions (T&Cs). In the UK, there are legal age restrictions for gambling (16 years for lotteries and 18 years for other gambling), 11 and therefore, age restriction warnings may plausibly reduce the perceived relevance of marketing to young people. Age restriction messages are particularly important as earlier initiation and increased gambling behaviour among children and adolescents is associated with a variety of adverse outcomes. 12e14 Although there is no legal mandate to include age warnings in gambling adverts in the UK, self-regulatory marketing codes remind operators that it is an offence under the Gambling Act (2005) to invite, cause or permit an underage individual to gamble. 11 Harm reduction messages have the potential to moderate the promotional message by encouraging controlled or reduced gambling, highlighting negative consequences and signposting support. 15,16 In the UK, there is no legal requirement to include harm reduction messages in gambling advertising and no standardised requirements on design, language, format and placement, although harm reduction phrases are suggested by self-regulatory marketing codes (e.g., "When the FUN stops, Stop" or "Know your limits and play within it"). 17,18 Research suggests that harm reduction messages can promote moderate changes in gambling behaviour. 19,20 The efficacy and salience of messages, however, is dependent on the positioning design, content and context, with those that are brief, easy to read and direct being the most effective. 19 Research also indicates that harm reduction messages should vary or may need to be tailored to different audiences, as reactions differ by age and degree of gambling experience. 20,21 These differences relate to the language used, whether consumers must actively respond to remove the message, and message framing. 20,21 Advertised gambling products often display T&Cs, which may be related to the ability to withdraw funds, eligibility to obtain bonuses and inducements or time restrictions. 22,23 Although it is not a legal requirement in the UK for gambling marketing to contain T&Cs, complaints regarding transparency about the advertised product (e.g., a gamble or offer) are subject to laws on consumer rights and trading standards, which provide a de facto mandate for their inclusion. 24 The design, tone, positioning and information included in T&Cs is at the discretion of the gambling operator and varies depending on the level and complexity of gambling or inducement promoted. Hing et al. 23 found that T&Cs can influence the perceived attractiveness of an advertised gamble among sports bettors. It is suggested, however, that the manner in which such information is currently presented in marketing can lead consumers to overestimate the attractiveness of offers and underestimate the true cost of obtaining bonuses or inducements. 23 Research has found that gamblers are often exposed to persuasively presented incentives and inducements in advertising, which may encourage more frequent or higher spend betting behaviour or stimulate impulse bets. 25e28 Consequently, clear and easily comprehensible T&Cs, which accurately reflect the odds of winning or magnitude of potential returns, may be particularly important for discouraging higher risk gambling and may have a particularly pronounced impact on problem gamblers.
In this study, we examine the presence and visibility of age restriction warnings, harm reduction messages and T&Cs within paid-for of gambling advertising in the UK. We do so across a range of advertisement formats, including broadcast (e.g., television and radio), and non-broadcast (e-mail and outdoor), and across a range of gambling formats (e.g., sports betting and bookmakers to lotteries).

Design and sample
A content analysis was conducted on paid-for gambling advertising (n ¼ 300) from eight media channels in the UK (Table 1). Gambling adverts for print press, internet, television, radio and email were sampled during a two-week period in 2018 (5e11th March and 12the18th March), selected using six months of gambling advertising expenditure data. The weeks were chosen to reflect a week with high intensity of gambling advertising (i.e., the highest weekly expenditure in £GBP) and an adjacent week closest to the average weekly expenditure over that six months. Ebiquity, a marketing and media consultancy agency, 29 supplied the expenditure data, the advert 'creatives' and information on design and placement (e.g., day of week shown). The paid-for advert channels sampled were limited to those monitored by Ebiquity (i.e., not social media pages or smartphone apps). Spend data across the advertising channels sampled are reported elsewhere. 30 One-hundred and fifty adverts were randomly sampled from each week. These were chosen from a maximum of 370 adverts in week 1 (average-intensity week) and 666 adverts in week 2 (highintensity week). Within each week, the random sample was stratified by the proportion of adverts reported through each channel (Table 1). For stratification purposes, adverts only available at a monthly level (direct mail, door drops and outdoor) were divided across the two weeks and then randomly sampled. The type of Table 1 Sample of gambling adverts included, by week of selection and advertising channel. was the adjacent week that reflected average marketing expenditure over the six months. b The number of marketing creatives selected from each week was determined by calculating the weekly proportion of each channel as a function of 150 creatives (i.e., 10% of adverts were internet in week one, which translated into 15/150 in the sample selected that week). c The number of creatives for these channels was only available at a monthly level. Because there were only a small number of creatives in March 2018, for stratification purposes, the monthly total was divided across the two weeks. d This number rounded to zero in the percent calculation (see footnote b). To ensure at least one creative from each channel was included it was purposefully rounded up and one creative was removed from the media channel with highest level of representation (print press).

N. Critchlow et al. / Public Health xxx (xxxx) xxx
gambling and the brand promoted was recorded by the research team when reviewing adverts.

Visibility of age restriction warnings, harm reduction messages and T&Cs
A coding protocol was developed to examine the presence and visibility of (1) age restriction warnings, (2) harm reduction messages and (3) T&Cs about the gamble(s) or offer(s) promoted. If present, the visibility of each feature was rated on a five-point scale (1 ¼ very poor visibility to 5 ¼ very good visibility). For T&Cs, a 'not applicable' option was included for adverts that did not promote a specific gamble or offer, and therefore, T&Cs were not required.
Ratings were based on the relative space taken up by each feature within the advert (%) ( Table 2). The visibility thresholds were developed, piloted and revised by the research team by analysing gambling adverts from the advertising channels sampled. The concept of proportional thresholds was derived from other conceptually similar public health contexts, such as the size of warnings used on cigarette packs. 31,32 In this study, the upper threshold (26% of advert space) was determined by purposively identifying adverts considered to have good visibility of harm reduction features during initial codebook development. In development and piloting, descending increments of 5% were considered to adequately reflect the different sizes of age warnings, harm reduction messages and T&Cs observed across adverts. We decided to use the same thresholds for all three features to ensure meaningful comparison. Although colour, font and positioning also influence the degree of visibility for each feature, these were prohibitively complex, subjective and time-consuming to code. Visibility, in terms of relative space within the advert (%), provided a single metric that was comparable across advertising formats. For radio advertising, visibility was based on the duration of each feature (in seconds).
In addition to quantitative coding, a free-text response was provided for each item to enable the coders to record a rationale for their visibility score (e.g., how much relative space they estimated was taken up by each feature). The free-text response also captured descriptive information on content and language used (e.g., whether websites or telephone helplines were provided in harm reduction messages) and illustrative detail on positioning (e.g., whether situated inside the main advert frame) and design (e.g., fonts and colours used).

Inter-rater reliability
To establish inter-rater reliability, the two researchers who completed the main coding (N.C. and C.M.) independently rated two adverts from each advertising channel sampled (n ¼ 16 adverts; 5% of sample). Discrepancies in ratings, and the level of detail recorded in the free-text responses, were resolved through discussion. There was agreement for 94% of adverts for the visibility of age restriction warnings (Cohen's k ¼ 0.78, classed as moderate), 94% agreement for visibility of harm reduction messages (k ¼ 0.78, moderate) and 100% agreement for visibility of the T&Cs (k ¼ 1.00, absolute agreement).

Data analysis
Data were analysed using SPSS, version 23 (Chicago, IL). Frequencies and proportions (%) were computed for advertising format, whether the advert appeared midweek (defined as Monday to Thursday) or at the weekend (Friday to Sunday), type of gambling advertised and gambling brands referenced. Frequencies and proportions (%) examined the visibility of age restriction warnings, harm reduction messages and T&Cs. For each feature, the main themes regarding size, positioning, font, positioning and text were summarised across all adverts, based on narrative information reported in the free-text responses.

Feature
Definition Codes and visibility threshold Age restriction warning How much marketing space is afforded to information or warnings that the content is only appropriate for those above the minimum legal purchasing age for that form of gambling (16 years old for lotteries and 18 years old for other gambling)?
No age restriction warning Very poor visibility (10% of space) Poor visibility (11e15% of space) Acceptable visibility (16e20% of space) Good visibility (21e25% of space) Very good (26% of space) Harm reduction How much marketing space is afforded to information about controlled or reduced gambling (e.g., gamble responsibly) or signposting to help for higher risk gambling (e.g., helpline)?
No harm reduction message. Very poor visibility (10% of space) Poor visibility (11e15% of space) Acceptable visibility (16e20% of space) Good visibility (21e25% of space) Very good (26% of space) Terms and conditions (T&Cs) How much marketing space is afforded to terms and conditions about the bet, gamble, offer, or inducements promoted (e.g., time limits on free bets, eligibility criteria, or restrictions on any cash withdrawals)?
No terms and conditions. Not applicable e No gamble promoted Very poor visibility (10% of space) Poor visibility (11e15% of space) Acceptable visibility (16e20% of space) Good visibility (21e25% of space) Very good (26% of space) price matching with named competitors and references to sporting events sponsored by another gambling brand.

Visibility of age restriction messages
One in seven adverts (14%) did not feature an age restriction warning (Table 3) (Fig. 1). Of the adverts which did, 84% of age warnings were rated as 'very poor visibility' (10% of advert space), with the free-text responses indicating that most only took up 1e5% of space (Fig. 2). The limited visibility of age restriction warnings was consistent across advertising channels.
The free-text responses indicated that most age restriction messages were small logos stating '18þ' (or '16þ' for lotteries) in the same banner as the harm reduction messages or short phrases reported as part of the T&Cs (e.g., 'over 18s only'). Most age restriction messages appeared in small font sizes (relative to the advertising content) were positioned outside of the main frame of the advert (i.e., banners at the extremities) and juxtaposed in terms of colour or size against visually stimulating advert content (Fig. 2).

Visibility of harm reduction messages
One in seven adverts (14%) did not contain a harm reduction message (Table 3) (Fig. 3). Of those that did, approximately half (54%) of harm reduction messages were rated as 'very poor visibility' (10% of advert space) (Fig. 4) and almost a third (30%) as 'poor visibility' (<15% of advert space).
The free-text responses indicated that most harm reduction messages were in small font sizes (relative to the advert content), positioned outside the main frame of the advert (i.e., banners at the extremities) or juxtaposed against stimulating content (e.g., grey colours against more striking graphics) (Fig. 4). Of the four adverts rated as 'acceptable visibility' (16-20% of advert space), three were radio adverts during which the harm reduction messages were clearly narrated at the end. One Internet banner advert had a harm reduction message rated as 'very good visibility' (26% of space), as the message took up the entire final frame of the advert (Fig. 5). There were no substantive differences across the other advertising channels for harm reduction messages, with most having consistently poor or very poor visibility.
Phrases used in harm reduction messages included 'BeGam-bleAware.org', 'Play it safe', 'Play responsibly', 'Please bet responsibly', 'Enjoy gambling responsibly', 'Gamble responsibly' and 'When the FUN stops, stop'. Some adverts also signposted support, e.g., 'For advice and information visit www.begambleaware.org' and 'Need help? Call the National Gambling Helpline on 0808 8020 133', mostly within the T&Cs. No adverts communicated negative consequences associated with higher risk or problem gambling, or specific guidance on controlled gambling (e.g., limit setting or taking breaks).

Visibility of T&Cs
Five percent of adverts promoted a brand only (i.e., no gamble), and therefore, T&Cs were not required. For the remainder of adverts that did require such information, 11% did not contain any T&Cs (Fig. 6). Almost three-quarters of adverts (73%) had T&Cs  rated as 'very poor visibility' (10% of advert space) (Fig. 7), and 11% were rated as 'poor visibility' (15% of space). The free-text responses indicated that T&Cs were almost always presented in small fonts and neutral colours (e.g., black text on a white background), featured outside the main frame of the advert (i.e., banner at the bottom), contained information deemed complex or technical (e.g., stipulations on withdrawing free bets or bonuses) and were juxtaposed against stimulating advert content (Fig. 7). Only a few adverts had T&Cs rated as 'acceptable' (2%), 'good' (3%) or 'very good' visibility (2%). In some radio adverts, for example, clear narration outlined the T&Cs for approximately 10 s at the end (typically a third of the advert), while some e-mail adverts contained an extensive list of T&Cs at the bottom of the message. There was little difference among other advertising channels (e.g., print press and television). Comparatively high visibility did not, however, always translate into ease of readability or comprehension. In one print advert, for example, although the T&Cs occupied around 20% of advert space, this contained a lot of  complex information and was displayed in small font, outside the main advert frame, and negatively juxtaposed against the marketing visuals (Fig. 8). This pattern of presentation was also true for most e-mail advertising.

Discussion
To our knowledge, this is the first study to examine the presence and visibility of age restriction warnings, harm reduction messages and T&Cs within paid-for gambling advertising in the UK. We found that these features were not present in at least one in ten gambling adverts. Currently, there is no legal mandate that gambling advertising in the UK must contain such consumer protection features. This means that inclusion is at the discretion of gambling operators, who are free to decide on design, tone, language, format and positioning, with no explicit and harmonised guidance.
The design and positioning of consumer protection messages influences their efficacy, with those that are brief, easy to read and direct most likely to be effective. 19e23 Even when such features were present in this study, approximately three-quarters of adverts had age warnings and T&Cs with very poor visibility, and more than half had harm reduction messages with very poor visibility. This included messages positioned outside the main advert frame, designs juxtaposed against stimulating content (i.e., neutral colours versus brighter evocative graphics) and information in small fonts. It has been suggested that gambling platforms and marketing are carefully designed to include subtle cues that initiate or increase gambling behaviour, something which literature has termed 'dark nudges'. 33 Our findings are consistent with this perspective, as most consumer protection features were subordinate to the advertising message and strategically placed where they will likely receive minimal attention in comparison with the marketing. The results therefore suggest that minimum standards of design may be necessary to increase the visibility of consumer protection messages in gambling advertising. Any attempts to revise or standardise such messages should be based on research that has tested optimal designs with consumers, as opposed to self-regulatory industry-led designs.
We also identified issues with the language used in consumer protection messages. For example, consistent with previous research, most adverts provided extensive and complex T&Cs concerning eligibility to participate or receive offers and often did so in a way that limited visibility. 22 Research has shown that consumers find it challenging to interpret the likelihood of winning on complex gambles (e.g., player to score and team to win), even before taking into account the emotional and attentional commitment needed to process the marketing content and stipulations of participation. 34 There is also evidence that how T&Cs are currently presented in marketing can lead consumers to overestimate the attractiveness of an offer and underestimate the true cost of obtaining bonuses or inducements. 23 Further research exploring consumers' comprehension and recall of T&Cs in gambling marketing, particularly in addition to processing the marketing content and gamble promoted, is a key research priority to improve the efficacy of current practice.
We found that no harm reduction messages explicitly discussed possible negative consequences of gambling (e.g., loss of time or money) or provided objective advice on controlled gambling (e.g., limit setting). Instead, most used phrases from the current selfregulatory guidance, such as 'When the FUN stops, stop', 'Enjoy gambling responsibly' and 'Play it safe' or simply stated telephone or Web addresses to signpost sources of support. 17,18 These messages have been criticised for failing to provide objective guidance on controlling gambling, relying heavily on an individual's interpretation of responsibility and encouraging gambling. 35e38 For example, the word 'fun' was almost always displayed more prominently than the rest of the message, thus acting as a promotional cue (Fig. 5). Research has also shown that the 'When the FUN Stops, Stop' message e which appeared in approximately two-fifths of the adverts analysed e may be associated with increased gambling compared with when no harm reduction message is shown at all. 36,38 In addition, concerning reasons for past month gambling, a 2018 report by the Gambling Commission 39 found that in 2018, only 29% did so for fun or enjoyment. For the benefit of consumers, appropriate harm reduction messages need to be developed and tested without the involvement of vested interests of the operators.
This study has several strengths. The findings are from a large stratified random sample of paid-for gambling advertising and are based on a range of broadcast and non-broadcast advertising formats, gambling types and gambling brands, thus increasing generalisability. The visibility thresholds were piloted on gambling advertising, and the study had good inter-rater reliability. There are, however, some limitations and avenues for future research. First, we only considered the visibility of consumer protection features. Future research is required to examine the attention allocated to such messages by consumers and perceived salience and behavioural impact. This could eye-tracking studies to examine attention to paid to consumer protection messages 40 or research comparing industry self-regulatory messages to alternative   designs. 41 There is already some evidence about how to more effectively present T&Cs 23 and limitations around current harm reduction messages, 38 albeit more evidence is needed to improve visibility and effectiveness across different media and gambling formats. Second, all advertising came from two weeks in March 2018, and the findings may not be representative of advertising at other points of the year. For example, as the weeks selected included ongoing sports seasons (e.g., football and horse racing), this understandably led to a high proportion of adverts for sports betting. It is plausible there may be a greater proportion of adverts for other forms of gambling at other stages of the year (e.g., lotteries and casinos), so comparative research would be beneficial. The sample was restricted to paid-for advertising, and the findings may not be generalised to other forms of marketing, particularly emergent activities such as social media. 30 Sampling was only based on advertising spend, but data were not available on audience targeting or reach. Finally, adverts were only stratified by the number of adverts reported through each channel, but not proportional spend.

Conclusion
This is the first study to examine the visibility, design, content and positioning of age restriction warnings, harm reduction messages and T&Cs in gambling advertising in the UK. The findings show that these features are not always present in gambling advertising, and even when they are, there are issues with respect to size, positioning, content and design. Given these limitations, additional regulatory steps should be considered to ensure that gambling advertising is accompanied by mandatory consumer protection messages. These messages should subscribe to minimum standards of design which ensure they are always easy to read, clearly visible, informative and reflect the actual harms that can result from gambling rather than subjective messages. These messages should be objectively developed and tested through research and consultation with consumers and those involved in gambling harm reduction and free from the vested interests of operators.

Acknowledgements
The authors thank Professor Gerda Reith (University of Glasgow), Dr. Andy MacGregor, Hannah Biggs, Jessica Shields (All ScotCen Social Research), Dr Richard Purves (University of Stirling) and Professor Agnes Nairn (University of Bristol) for their advice as part of the wider project team. The authors thank Ebiquity for supplying the adverts and information and pay particular and special thanks to Steven Ginnis, Holly Kitson and Yasmin White (all Ipsos MORI) who made the arrangements to access the advert sample.

Ethical approval
The study was approved by the University of Stirling's General University Ethics Panel (GUEP413).

Funding
The study upon which this article is based was funded by GambleAware, a national charity instructed by Government to commission research into gambling in Great Britain. GambleAware is funded through contributions from the gambling industry, but decisions about what research to fund are made by the Advisory Board for Safer Gambling (ABSG, formerly the Responsible Gambling Strategy Board [RGSB]), an independent group that provides advice on gambling policy and research to Government. In September 2016, the ABSG (then RGSB) and GambleAware published a Research Commissioning and Governance Procedure that describes how research priorities are set and commissioned, in isolation from the gambling industry.