Edinburgh Explorer Examining the frequency and nature of gambling marketing in televised broadcasts of professional sporting events in the United Kingdom

Objective : Gambling operators in the United Kingdom have introduced a voluntary ban on adverts broadcast during televised sport before 21:00 (the ‘whistle-to-whistle’ ban). To inform debates around the potential effectiveness of this ban, we examine the frequency and nature of gambling marketing in televised broadcasts across professional sporting events. Study Design : Frequency analysis of verbal and visual gambling marketing references during television broadcasts of football ( n =5), tennis, Formula 1, boxing and rugby union (each n =1) from 2018. Methods : For each gambling reference, we coded: whether it appeared in-play or out-of-play; location (e.g. pitch-side advertising); format (e.g. branded merchandise); duration (seconds); number of identical references visible simultaneously; brand; and presence of age restriction or harm-reduction messages. Results : Boxing contained the most gambling references, on average, per broadcast minute (4.70 references), followed by football (2.75), rugby union (0.55), and tennis (0.11). Formula 1 contained no gambling references. In boxing, references most frequently appeared within the area-of-play. For football and rugby union, references most frequently appeared around the pitch border or within the area-of-play (e.g. branded shirts). Only a small minority of references were for adverts during commercial breaks that would be subject to the whistle-to-whistle ban (e.g. 2% of references in football). Less than 1% of references in boxing, and only 3% of references in football, contained age restriction or harm-reduction messages. Conclusions : As gambling sponsorship extends much beyond adverts in commercial breaks, the ‘whistle-to-whistle’ ban will have limited effect on gambling exposure. Gambling sponsorship activities rarely contain harm reduction messages.


Introduction
In the United Kingdom (UK), it is estimated that there are two million problem gamblers or individuals at risk of problem gambling [1]. One factor suggested to drive gambling is exposure to marketing [2]. In the UK, the introduction of the Gambling Act (2005) permitted gambling marketing across all forms of media, resulting in a dramatic expansion of gambling marketing activity [3]. Sport sponsorship is a particularly effective form of marketing that allows brands to be associated with, and capitalise on, the emotional connections that consumers have with teams and events [4][5][6]. Gambling sport sponsorship has been highly prevalent in recent years.
In the 2018/19 football season, over half of the clubs in the English Premier League (EPL) and English Championship, and one third of teams in the Scottish Premier League (SPL), had gambling companies as shirt sponsors [7]. Moreover, all four professional leagues and both of the domestic cup competitions in Scotland, and three of the four professional leagues in England, were sponsored by a gambling operator [8].
Sponsorship of football teams and events provides a high profile platform for exposure to gambling marketing. A study of the British Broadcasting Corporation's (BBC) Match of the Day programmea popular free-to-air TV show providing highlights of EPL fixturesfound that viewers were exposed to over 250 instances of gambling marketing per-episode [9]. This was despite the fact the programme is broadcast on a non-commercial channel (i.e. does not feature explicit adverts in commercial breaks). Such exposure is also true for international football matches. For example, 17% of all advertising shown during coverage of the 2018 World Cup on a commercial television channel in the UK was for gambling operators [10].
The heavy involvement of gambling companies in football has been described as the 'gamblification of sport', with some young men reporting that they cannot watch football without placing multiple bets [11,12].
Research suggests that exposure to gambling sports sponsorship is associated with a variety of consumer reactions and outcomes. These include increasing knowledge of gambling brands, normalising gambling as an everyday activity, encouraging feelings of greater control over betting outcomes, and stimulating sign-up with more than one betting provider [13][14][15].
Exploratory research into gambling sponsorship during televised sport has also found that exposure to such marketing activities may encourage gambling among problem and recovering problem gamblers [16].
In addition, several studies have shown that many of the gambling products that are advertised during live football are for complex events that may mislead consumers about the likelihood of winning, resulting in high profit margins for the bookmakers [17,18].
There has been growing public concern and political pressure over the volume and frequency of gambling marketing in televised sports, and the potential impact it may have on younger viewers [19,20]. In response, the Industry Group for Responsible Gambling (IGRG) has amended its Gambling Industry Code for Socially Responsible Advertising to include a ban on all commercial gambling adverts during pre-watershed (21:00) televised sport [21]. The restriction will begin five minutes before the event and end five minutes after it finishes. The proposed restriction also prohibits betting adverts on highlights shows and replays of sporting events shown before 21:00, and an end to bookmaker sponsorship of pre-watershed sports programmes (albeit not sponsorship of teams participating in the events screened). The voluntary restriction was proposed by the IGRG after they reviewed the previous version of the Gambling Industry Code [22] and was adopted industry-wide from 1 st August 2019, coinciding with the start of the 2019/2020 football season in England and Scotland.
The IGRG's voluntary ban only applies to adverts shown during certain parts of the broadcast, for example during commercial breaks (e.g. at half-time). This means that other high-profile forms of gambling marketing (e.g. pitch-side advertising and shirt sponsorship) continue to be permitted throughout the broadcasts, including during the sporting action when audience attention is likely to be most focused. To examine to what extent (if at all) the proposed voluntary restriction will limit exposure to gambling marketing, and to inform debates around its potential effectiveness, we therefore examine the frequency and nature of gambling marketing in televised broadcasts from a range of professional sporting events.

Design
A frequency analysis of verbal and visual references to gambling observed during television broadcasts of football, tennis, Formula 1, boxing and rugby union in the UK. The design was informed by previous research into alcohol sponsorship of English club football [23], the UEFA EURO 2016 football tournament [24] and Formula 1 racing [25]. Similar methods have also been used to assess frequency of alcohol and gambling marketing references in Australian sport broadcasts [26,27].

Selection of broadcasts
A purposive sample of professional sporting events (n=9) were recorded as broadcast in the UK on either public service (e.g. BBC) or commercial broadcasters (e.g. Sky Sports or BT Sports) ( Table 1). The sample included five live football broadcasts (including top-flight league football in England and Scotland, continental club football, and international football) and broadcasts of four other sports: rugby union, tennis, boxing, and Formula 1. A larger sample of football matches was selected due to its position as the most popular sport in the UK, both in-person and on television.
[ Table 1] All the selected programmes were recorded in their entirety, using recordable DVD players or an on-demand TV service available for academic institutions (Box of Broadcasts). Where relevant, each recording included normal playing time (e.g. the full football match), added time, extra time, pre-and post-event interviews and discussion, analysis during sport breaks (e.g. half-time), and any commercial breaks. The recordings did not include any content that was not part of the main scheduled broadcast (e.g. content on on-demand television and content accessible through interactive television).

Defining gambling references
Consistent with previous research into alcohol marketing, a gambling reference was defined as any visual reference to gambling or to a gambling brand, lasting one second or more, during the broadcast programme or commercial break [24]. A reference was counted each time it appeared, irrespective of whether it had been previously seen (e.g. pitch-side advertising board seen in-play first and then again in a replay). A new reference was counted each time the camera angle changed, even if the reference source remained the same (e.g. pitch-side advertising first viewed from behind the goal and then again when the angle reverted to the wide angle). A new reference was also counted if a source went out of shot for more than a second (e.g. if the camera panned away from the pitch-side advertising and back again).
If multiple different references were presented at the same time, (e.g. static and electronic pitch advertising), each was recorded as a separate reference. If multiple identical references were visible at the same time (e.g. if the same brand name or slogan appeared multiple times around the border of a football pitch), they were recorded as 'identical references visible at the same time'. The only exception to this occurred when gambling marketing appeared as a shirt sponsor in a multiple-participation (team) sport, such as football. Coding each brand every time it appeared on players' shirts would have been prohibitively slow, therefore if multiple references to different gambling brands of the same format were visible (e.g. two different gambling sponsors visible on players' shirts at the same time) these were coded as 'multiple brands'.

Codebook variables
All gambling marketing references were captured using a codebook adapted from studies examining alcohol marketing in professional televised football [24]. Each reference was coded on a series of nominal and continuous criteria ( Table 2).

Procedure and inter-rater reliability
An initial codebook was piloted on one of the sampled broadcasts, and revised based on discussion between the researchers. All nine broadcasts were systematically coded in separate Microsoft Excel worksheets by AM and RP. To establish inter-rater reliability, AM and RP both independently coded part of one football broadcast. A football match was chosen for interrater reliability as this represented the majority of the broadcasts sampled. Only a section of the broadcast was coded due to the high levels of inter-rater reliability shown in previous applications of the codebook [24]. Both researchers coded 30 minutes of footage (15 minutes of pre-match and 15 minutes of in-play) and discussed any discrepancies. Following discussion, inter-rater reliability was established through percentage agreement on the number of references coded for each of the categorical variables (e.g. number of references in the prematch), consistent with previous applications [24]. When computed across each section of the codebook, there was high agreement for broadcast segment (100% agreement), reference type (100%), reference location (94%), reference format (97%), content of the reference (100%), and which brand was featured (94%). These estimates exceed the suggested 70% threshold for acceptable inter-rater agreement using the percentage measure [28] and are consistent with previous uses of this codebook [24]. We also used independent sample t-tests to examine any differences in average number of identical references identified and duration of references, and found no significant differences (t=0.17; p=0.87 and t=-0.41; p=0.97, respectively).

Ethics
As this study was based on publicly available television broadcasts, and did not include any research participants, no ethical approval was required.

Data analysis
Data were analysed using SPSS version 23 (SPSS Inc). All analyses were computed separately by sport (e.g. football or rugby union). For each broadcast, frequencies and percentages calculated for: the total number of gambling references; broadcast segment (in-play vs. out-ofplay); type of reference (verbal, visual or both); location of references (e.g. area of play); format of references (e.g. branded merchandise); gambling brands referenced, and the content of references (e.g. logos, signposts to gamble, harm-reduction messages). Range, medians, and sums were computed for the duration of references and number of identical references visible at the same time.
For each broadcast, we calculated the average number of gambling references per broadcast minute (total number of references divided by the length of the broadcast [in minutes]). This was then converted into the estimated frequency of gambling marketing references in seconds (number of seconds in a minute (60) divided by the average number of references per minute). As the length of each broadcast was dictated by the sport and scheduled time allocated to the programme, these metrics enabled us to compare across sport and broadcast types. For football, where multiple television broadcasts were captured, the average number of references per minute and estimated frequency of references (in seconds) was computed for each television broadcast separately and for a combined total. We did not conduct any inferential tests of these metrics due to the limited size, particularly for the non-football sports where only one broadcast was captured for each.

Football
We recorded 2,595 gambling marketing references across the five football matches, an average of 519 per broadcast (Table 3). Across the five matches, gambling marketing references appeared, on average, 2.75 times per broadcast minute; equivalent to once every 21 seconds.
The median duration of references was five seconds. The majority of references (77%) featured in-play and the most popular location of references was border of play (38%), for example static pitch-side adverts. The most popular format of references was branded merchandise (41%), principally due to sponsor logos on players' shirts ( Figure 1). Gambling brand logos appeared in 99% of the references and the most featured gambling brand was Betway (27%, sponsor of West Ham United football team). Harm-reduction and age restriction messages appeared in 3% of references, but were largely confined to commercial adverts. There were differences between football broadcasts depending on the competition and teams featured (Table 4) [ Table 3] [ Table 4] [ Figure 1]

Boxing
We recorded 358 gambling marketing references in the boxing broadcast, an average of 4.70 per broadcast minute or approximately once every 13 seconds ( Table 3). The most popular location of references was in the boxing ring (48%), as there was static gambling advertisements on the ring floor, the ropes around the boundary, and the corner covers ( Figure   2). William Hill was the official sponsor of the event and also the most popular brand featured (73%). Brand logos appeared in 99% of the references. The median duration indicated that at least half of the references lasted for 9 seconds or more, which is largely attributable to the limited change in camera angles during the match action (i.e. the camera mostly rotates around in a continuous motion rather than changing shot). There were no harm reduction messages and age restriction messages appeared in less than 1% of the references. None of the gambling marketing references in this broadcast were commercial advertisements which may have been due to the broadcast being a highlights programme.

Rugby Union
We recorded 101 gambling marketing references in the rugby union broadcast, an average of 0.55 references per broadcast minute or approximately once every 109 seconds ( Table 3). The majority of references featured in-play (92%). The most popular location of references was around the pitch border (98%) and the most popular format was static adverts (93%), for example pitch-side advertising hoardings. All the references were for the National Lottery, as their brand logo appeared on pitch-side adverts for Sport Scotland, a national organisation who receive funding from the lottery (Figure 3). The median duration of references was two seconds. There were no lower risk (0%) or age restriction (0%) messages featured. None of the gambling marketing references were commercial advertisements. This was expected as the programme was broadcast on a non-commercial television channel (BBC One).

Tennis
We recorded 26 gambling marketing references in the tennis broadcast, an average of 0.11 references per minute or once every 545 seconds ( Table 3). The majority of references appeared out-of-play (96%) and the most popular format and location of references was sponsor leadins (85%), all of which featured Bet365the official sponsors of the coverage (Figure 4). The median duration of references was 7 seconds. Gambling brand logos featured in 96% of the references. Harm-reduction messages and age restriction messages featured in 96% of the references. Only three of the references were commercial advertisements (12% of all references observed), all of which were 30 seconds in length. This included gambling adverts in periods that would not be permitted during the 'whistle to whistle' ban (e.g. during breaks in the match).

Formula 1
No references to gambling marketing were recorded across the entire Formula 1 broadcast.

Discussion
This is the first study to examine the frequency and nature of gambling marketing across popular sport broadcasts on TV in the UK. We found gambling marketing in all but one broadcast, although the frequency and extent of exposure varied. Boxing had the most frequent references, with gambling marketing appearing approximately every 13 seconds. We also found frequent gambling marketing references in football, particularly high-profile EPL and SPL matches in which teams had official gambling sponsors (e.g. on the shirts). In contrast, the tennis broadcast only featured gambling marketing approximately once every nine minutes, and no references were observed in the Formula 1 broadcast due to the organisation's ban on gambling sponsorship within the sport. This ban has recently been lifted for the upcoming 2020 F1 season with online bookmaker 188Bet announced as the sport's first ever gambling sponsor following the its takeover by Liberty Media in 2016 [29].
Across the sports where gambling marketing was observed, only a small proportion of references featured in the commercial breaks or sponsorship lead-ins. As the IGRG's voluntary ban on pre-watershed advertising only applies to this small proportion, the findings suggest that it will have limited impact on the volume of gambling marketing in sport broadcasts. For example, in EFL and SPL matches where both teams had gambling shirt sponsors, there were over 900 gambling marketing references. That half of the teams in EPL in the 2019/2020 season have a gambling operator as a shirt sponsor [30], suggests these levels of exposure will continue, irrespective of the voluntary ban. We identified a sophisticated array of opportunities to promote gambling companies through sporting broadcasts. These included, but were not restricted to, pitch-side advertising, match shirts, structures around the event (e.g. manager's dugout in football), and post-match interview boards. The majority of references featured in high-profile locations such as pitch-side advertising or shirt sponsorship and at the point when most people would be likely to be watching (i.e. during the match action as opposed to prematch build up or post-match discussion). This highlights the difficulties and complexity of regulating across sports broadcasts.
Research supports a link between marketing exposure and subsequent gambling behaviour [2], including suggestions that marketing may act as a trigger for current gamblers or relapse among those wanting to reduce gambling behaviour [31,32]. In response, governments in other countries have introduced statutory restrictions on gambling advertising during live sport. For example, Australia have introduced restrictions around gambling advertising during live sporting events on commercial television and radio, public-service broadcaster SBS, subscription television and online platforms [33], and Italy introduced a blanket ban on gambling sport sponsorship in July 2019 [34]. That our findings show a high frequency and visibility of gambling marketing in high-profile sporting broadcasts, including in programmes broadcast before the 21:00 watershed, suggests that such statutory and mandatory restrictions may also be required to reduce exposure, and provide greater protection to, young and vulnerable groups in the UK (e.g. problem gamblers). Any potential statutory ban on gambling sponsorship in the UK would need to consider the myriad of channels included in sponsorship deals. However, one potential option for policy-makers would be to introduce a ban on shirt sponsorship such as the one proposed by the Labour party in 2007 [35] or measures similar to those introduced in Ireland as part of the Public Health (Alcohol) Act 2018 which includes restrictions on advertising during a sporting event [36]. This brings with it various challenges as seen in France where a ban on sponsorship by alcohol brands is regularly flouted [24].
Although the primary purpose of sport sponsorship is to promote gambling brands and products, sports sponsorship also provides a high profile and high-reach platform to communicate harm reduction messages. Examples of harm reduction messages include warning that gambling is an age restricted activity, suggestions of controlled or reduced gambling, or signposts to sources of support (e.g. helplines or websites) [37]. In this study, however, we found that frequency of these messages was very limited across all sport broadcasts, appearing in no more than 3% of references in boxing, rugby, and football. In particular, there were no instances of harm reduction and age restriction warnings in highprofile locations or in the most frequent references in the broadcasts. For example, branded shirts worn by match participants only displayed the brand logos, but not any harm reduction message. Even when harm reduction messages were present, they were mostly confined to commercial advertising breaks or sponsorship lead-ins which, paradoxically, would actually be removed under the voluntary ban. If sport sponsorship is permitted, minimum standards of design may be required to ensure that adequate harm reduction messages are also provided to act as a buffer to the commercial message. However, it is important to note that present harm reduction messages might simply be ignored [38] or fail to have the intended effect on behaviour [39].
The study has a number of limitations. The results are based on a relatively small and purposively sampled selection of sport broadcasts. Both of the EFL and SPL football matches contained teams that had gambling sponsors, and therefore the results may not be generalisable to other fixtures where this is not the case. For boxing, a highlights programme of the main event was selected because the original full broadcast (including all the undercard fights) would have been too time consuming to code. As such, the results may not be representative of gambling marketing exposure across the entire live broadcast. The boxing highlights programme also does not provide insight into the number of commercial adverts that would have featured during the live broadcast, albeit the main event occurred after the voluntary 21:00 watershed. For Formula 1, the race was broadcast simultaneously on both a free-to-air (Channel 4) and subscription based commercial broadcaster (Sky Sports). The former was recorded for convenience and, although footage of the race would likely have been identical, the findings may not be representative of the commercial breaks on the respective broadcasters. Although we had good variety of teams and competitions for football, we only sampled one broadcast from other sports and therefore findings may not be reflective of exposure for other teams and competitions. For example, we did not include any women's sport as our intention was to include the most high-profile sporting events in the UK (i.e. largest audiences). Including broadcasts of Scottish Women's football in particular may have offered an interesting contrast to the men's game due to their public stance against sponsorship by gambling brands [40].

Conclusion
Televised sports broadcasts contain a high frequency of gambling marketing, only a minority of which were explicit advertisements appearing during commercial breaks and the majority of which do not include harm reduction messages. The voluntary 'whistle-to-whistle' agreement introduced by the IGRG therefore does not include much of the main elements of sponsorship activity including shirt sponsorship or pitch-side advertising, most of which appeared in-play and therefore will coincide with peak audience attention. It is doubtful that this proposal will do much to reduce viewers' exposure to gambling marketing, and that further mandatory restrictions may be required.