Last week the Gambling Commission released the first data findings from their new quarterly survey – The Gambling Survey for Great Britain.
That it was launched on the 29th February – a Leap Day – may be a coincidence but it does lend itself to puns about whether it does represent a leap forward in providing meaningful “Evidence” in a very contested area.
Here at GLEN we have been awaiting this “GSGB 1st wave” with great interest, and we have now pored over the data and have formed our own take away thoughts on what it now contains.
If you want to read these in full then scroll down, but before doing so you might be interested in hearing about our take in what is NOT included in this new Survey, and why we feel that is both significant and suspiciously unhelpful.
Undoubtedly, the biggest area of contention in the sector has been reliability and variation of statistics around harmful gambling, or what it often still insists on denoting as being “problem gambling”.
UKGC’s own quarterly telephone surveys, which the GSGB now replace, have consistently reported a “problem gambling” rate of between 0.2% and 0.3% of the adult population.
The NHS-led national Health Surveys ranged between 0.3% and 0.8%.
YouGov surveys commissioned by GambleAware suggested figures above 2%.
Naturally, while the industry “liked” the 0.2% for it being lower than the other sources and heralded these as being conclusive proof that gambling does not cause widespread harm, those with interest in public health and gambling harm reduction also unsurprisingly “liked” the higher values. Although “liked” should probably be taken as meaning “saw more value in accepting”.
But in a sector where any and all change is vociferously challenged by a “need to be EVIDENCE led” this marked variation causes massive problems.
So, what does the new GSGB say about harmful gambling?
Rather remarkably it currently says… NOTHING.
Nada, zip, or any other word you choose to use to describe the absence of anything at all.
This “1st wave” does not seek to report on rates of harm.
Nor will waves 2, 3 or 4. Only once a full year has been published will UKGC even hazard an estimate of harm.
This seems VERY UNHELPFUL when this current year will see all outstanding White Paper consultations end – without having had sight of this new data.
Even worse, this arguably could be seen as being a direct response by UKGC to outrage amongst industry figures that occurred when the pilot to the GSGB did give preliminary estimates of harm which came in at 1.3% – so 6 times the latest Quarterly Telephone Survey rates – and were included in the Government’s “High Stakes: Gambling Reform for a Digital Age” White Paper document.
After gleefully crowing over what the industry have flagged up as being reducing PG rates – seemingly now down to an easily dismissively small 0.2% – is this delay perhaps a response by UKGC to industry ire at the thought of having to acknowledge over 600,000 adults actively being severely harmed by their products.
Has this influenced UKGC not to report further on these until all new regulations are done and dusted?
A question, or even an allegation, which we feel needs to be posed.
Interestingly, this may mirror previous decisions made by UKGC in response to ill-received statistics by the industry lobby. Particularly looking back to their actions following rising rates of harm being reported in the British Gambling Prevalence Survey – the true ancestor of the GSGB.
BGPS rates reported severe harm rising steadily across its existence – 0.6% in 1999, 0.8% in 2007, and finally up to 1.2% in 2010 – whereupon UKGC promptly opted to scrap that longer form of survey.
Which in turn has led to increasing calls for something comparable to be reintroduced – hence GSGB (or BGPS mark 2.0)
However, putting aside conspiracy theories, let us instead look at the GSGB and what it potentially can tell us.
Our Thoughts on the difficult task of quantifying Harm caused through gambling
One of the first things to be said is that any survey which asks people to truthfully self-report their involvement in potentially stigmatising behaviours has a very hard task ahead of it.
Our members know only too well how gambling so often became something which they found themselves hiding, denying, and understating, if ever asked about it. Even when it was they themselves who were doing the asking.
For this reason GLEN has a high degree of skepticism about the ability of this or of any other survey to reveal the true extent of gambling participation and harm.
Our opinion is that if you want to see gambling prevalence and associated harm then purely objective and factual sources are far more likely to reveal real usable evidence. The obvious sources for both participation and spend are surely transactional databases where they exist.
For total spend and patterns of spend (deposits) we should be looking at data held by the big banks and financial institutes – similar to the Big Data research done using Lloyds Bank customer accounts a few years back [Muggleton N, 2021].
Analysis of transactional data held on this massive scale is capable of seeing past self-reported behaviors to what was actually spent on gambling products, as well as any conjunction with signs of financial stress in customer’s transactions and financial management.
For total spend AND other behavioural information, such as time spent playing, we should be looking to collated industry transaction (CRM) data.
Here at GLEN we would have loved seeing the UKGC collaborating with the big banks to access such large scale financial data, and also requiring operators to provide cross-sector transactional data which could be published in its entirety.
We do understand that this latter data would be skewed towards online rather than in-person gambling but it would still be about what did happen rather than what respondents said happened.
And while we would remain skeptical of the value of any harms quantified through the new survey process it is still bitterly disappointing that the Gambling Commission have chosen to delay even that much scrutiny.
That does not however totally disvalue the usefulness of what they have now published.
So, what is there in this “first wave” which can be picked out for greater exploration.
What does the GSGB tell us?
Demographic and Access Data
Where the GSGB does provide interesting insights is in the data insights it gives as to peoples’ access and use of technology, and in some of the subjective questions – the “reasons and reactions” that people express about their gambling.
We do urge caution on reading too much into the quantities here for the reasons already given, stigmatic aversion and reluctance to self-report harmful behaviours and consequences, but sometimes we can see unintended revelations or hints at such being the case.
The first thing we should touch upon is reach.
The survey was sent to 18,776 addresses despite it being estimated that only 17,086 (91%) of these were “residences” and this latter number then being used to calculate the “eligible” response rate.
As there was no seeming mechanism to tell if addresses were residences in advance then these would appear to be similar lack of differentiating as to whether responses which were returned came only from the assumed “eligible” 91%. As such we would be tempted to say that the response rate of 3229 “productive” addresses should mean a 17% return rate rather than the 19% they credit themselves with having achieved, although at the end of the day that is a very minor quibble.
Totaling up addresses where more than one person responded we get the total number of individual responses being 4801.
There was a consistent level of response (18-19%) across all three nations (England, Scotland, Wales).
Twice as many people did the survey online than sent in postal responses.
There were more women responding than men (55% v 45%) although this finding was reversed in those aged 65 and upwards.
One of the criticisms which many random address surveys have is that you have no idea whether the people responding are doing so because they are engaged by (interested in) the subject of gambling or are just a fixed proportion of the population who feel obliged/compelled to answer every survey which comes through their letter box. Statisticians would like it to be this latter group as it creates a “true” randomly leveled picture of engagement for the other questions, but normally no-one can really tell.
All of which makes the age and sex breakdown potentially interesting, but frustratingly vague.
However, we do know from national population statistics that there are proportionally more women than men in GB aged above 65 yet the fact that men in these age groups bucked the trend on responsiveness perhaps suggests that these individuals were responding because they are genuinely more engaged in the subject of gambling. And yet seemingly no men, or indeed women, aged over 75 reported gambling on non-lottery products, which does seem strange, and especially so when the drop of in responsiveness of over 75s compared to over 65s in lottery-only products would seem to suggest that older people were still willing to gamble into their older years but apparently hit a brick-wall in terms of non-lottery products before the 75 year mark.
One reason why we are spending time highlighting the age/sex differences is that in terms of estimating why people respond to a survey on gambling participation there is a massive question-mark in terms of how many people are engaged in the subject due to the presence of someone else close to them who gambles – the “Affected Other” impact.
One interpretation of women being more responsive to men in this survey when within working age bands may be that these include women who are more than averagely engaged in the subject. Something which could be argued from other data included should not seemingly be the case as women across all age groups self-report less participation in gambling than their male peers. So, the interesting question then becomes, are women below retirement ages feeling more inclined to respond because they feel that gambling is an issue of significance to them through their realisation that they are being impacted by another person’s gambling?
Perhaps potentially to a harmful level?
We simply cannot tell – the questions being reported on in this first wave do not ask about this. They do not ask if someone close to you – e.g. in your household – gambles, only if you yourself gamble.
Reading the full list of questions included in the Online survey does show that people are being asked to give feedback on whether they feel impacted by the presence of gambling in those around them but for some reason none of these responses, even just reporting how many people did identify that others around them participated in gambling, were deemed appropriate for this 1st Wave of data results.
Which is a shame, it would have been really interesting if there had been a question published which reported this as it would perhaps have added context to the differing response rates, including whether “selection bias” is present.
It also re-raises the question about why Harm is simply not being reported upon at all until after a whole year of data has been gathered and reported, and whether the UKGC are running scared of publishing harm figures which challenge the current industry narrative of almost “insignificant” harm existing because we all know that the one area where industry lobbies really feel unsure about is how to respond to harm being caused to people who are not their own customers and therefore are not “responsible” due to their own gambling involvement when harm is happening.
It is very easy to argue that only a tiny percentage of people are harmed by gambling and accordingly the industry shouldn’t be expected to bear the full burden of protecting those “unfortunate but statistically insignificant” people from the harm which, after all, they are simply causing to themselves through their own poor choices.
Why, the trope goes, introduce a Nanny state for all in response to harm being caused to so few when those few are in actuality actively causing their own suffering? A rather callous argument which echoes the “guns don’t kill people, people kill people” style industry defense we hear from across the Atlantic.
But just as in the USA the societal counter-argument to industry deflection is that this line of “individual rather than corporate” responsibility totally avoids having to acknowledge the extent of collateral harm which use of their products is inflicting on others, many of whom are women and children. And unlike with the USA gun lobby there are no fatuous responses available to the GB gambling industry along the lines of “its a constitutional right to gamble”, or “the best way to prevent harm is for affected others to also gamble.”
That is the reason why statistics on harms caused by gambling are so contentious, and why the last thing the industry really wants is a GB survey which all academics can look at and endorse as being a robust and accurate representation of the bigger harm picture. How many people in society are being harmed by gambling products.
Here at GLEN we are NOT convinced that UKGC is showing enough commitment to impartiality or that they are willing to upset the industry – who after all are their “partners/clients” – by being fully transparent and open.
As noted earlier the pilot produced record high figures of “Problem Gambling” harm – 1.3% – and if in twelve months time this figure is close to what is then produced it will be scandalous that they held this back until after the White Paper consultations were all wrapped up to the industry’s satisfaction.
Also, if the figures when they do appear look far closer to the 0.2% which the industry is currently “happy” with then we feel there should be a call for full disclosure as to how the methodology has changed from the original pilot.
With regards to this last point it is slightly concerning that UKGC has paired their release of this 1st Wave data report alongside assessments by the likes of Professor Sturgis into the methodology of the GSGB and the pilot version which identify that households where gambling harms exist are less likely to respond to surveys on that subject (negative selection bias) yet also suggest that online responses provide higher rates of PGSI scoring than those which are done by other methods of delivery/response (positive selection bias). We say “concerning” as these seem to be two competing “selection-bias” factors which seem to be quite difficult to reconcile or balance out, in which case we would probably prefer it that no attempt to do so was made, and that the statistics were presented on actual rather than adjusted basis with these possible qualifiers simply given as explanatory/cautionary notes.
Before we leave the demography data it probably shouldn’t come as any surprise that over 80% of respondents aged under 45 did so online rather than by post, and that this gradually dropped in older age groups. However, the point of parity occurred in the 65-74 year age which does suggest that rather than “silver surfers” – older people fully comfortable with online technologies – being the exception in older age groups that they will soon become the norm.
Why should this matter? Because familiarisation with technology also means easy access to online gambling apps.
There is a massive area of risk which we at GLEN are aware of but which very few others seem to be giving much attention.
People who gamble regularly but at relatively low stakes due to financial “prudence” or simply limited funds are at heightened risk of developing harmful gambling behaviours if they come into large sums of money. While this outcome has tended to be told using analogies such as someone winning a huge bingo jackpot then becoming severely addicted and losing not only all their winnings but also their life savings and family homes, there is another much more numerous high risk group – people who reach retirement age and suddenly have access to large lump sums (if on older “defined benefit” pensions) or have access to large pools of accessible money (if SIPPS).
This means that the conjunction of more tech savvy individuals at an age where they come into one-off large sums of money and suddenly also have a lot of free time to “kill” are a very dangerous combination and one which future iterations of the GSGB will perhaps shed alarming new light upon.
Reported Behaviours
Looking beyond response demographics to the actual behaviours being reported, what are the points worth picking out?
How about this one for starters.
4% of individuals who participated in non-lottery gambling in the last 12 months “Hated” doing so.
Hated? Really? That is a very strong negative reaction which dies suggest considerable likelihood of there being harm underlying this response.
The unanswered question here is did they hate the product/experience or the consequences it had for them. Many individuals who have formed addictive or unhealthy relationships with gambling do so reluctantly and regretfully out of compulsion. In effect they know they are doing something self-harming and feel bad because of that, but still cannot bring themselves to stop.
What would have been useful would be hearing how many of these respondents gambled more than once – as if they were hating the experience but went back and repeated it anyway then this would point more towards addictive compulsion being in play.
Perhaps unsurprisingly the highest percentage reasons given for gambling were to “win big money” (86% all participants) – which we are happy enough to presume relates in the main to National Lottery products (which were included for all these reason responses) and the perennial hope of unimaginable wealth – meaning a security legacy for future family generations – should it ever actually “be you”.
This was also the reason which scored most consistently across all age groups, which perhaps simply shows that everyone can continue to have aspirational dreams at any age.
What is perhaps more telling are the distributions of other reasons across age groups.
After winning big the next highest reason was “because its fun” (70% of all participants) – but there was a distinct pattern here where this concept peaked amongst those aged 25-34 then tailed away as people got older.
This was the same general distribution pattern as for those giving reasons for gambling as being:
- To Compete with others
- It’s exciting
- For the mental challenge
- Sense of achievement
All these reasons peaked with the 25-34 age group then fell away.
We are slightly confused as to what is meant by “sense of achievement”, or why the researchers felt this was an option worth including, but when grouping these other reasons alongside the “fun” response it does seem to point to social influences such as peer involvement as being significant drivers in this age group.
The same off-set bell shape also existed for other reasons but peaked slightly later in the 34-44 age group. These reasons being:
- Hobby or pastime
- To escape boredom
- To Relax
- To make Money
Collectively, apart from perhaps the hobby/pastime response, these reasons point towards gambling participation being a reaction to something else in ones’ life rather than being something which draws in players in that age group because of the merits of the product itself. This is concerning as these do appear to be “coping responses” and as such do point towards gambling for the “wrong” reasons, and the possible pathways towards addictive ideation forming.
The last reason listed above – to make money – did not drop away as markedly as the other reasons listed above – perhaps reflecting a correlation between winning big and winning anything on lottery products – but it was still notable that while roughly two-thirds of those aged under 45 gave this as reason for participating only just above half of those who were older than this did so.
This suggests that perhaps younger people tend to have exagerated expectations of gambling as a source of wealth creation whereas those older have such expectations tempered and see gambling as more of a time or leisure activity than one which generates financial gain.
What is perhaps interesting is that the fact that where there were obvious age related peaks for these reasons they were all in age groups above 25 years old – and not below – which brings into question the wisdom or effectiveness of the recent differentiation in maximum online slot limits being placed as that age. [£2 stakes if aged under 25, £5 if above that age]
It is however also true that responsiveness in this younger age group was the lowest of all.
In Summary – and a radical suggestion
All in all, the Gambling Survey for Great Britain is an interesting addition to the data being gathered, but it is questionable as to what purpose it is seen as serving, and whether the UKGC should be looking to other available sources of data in order to create large data insights into financial participation and associated risk.
Indeed, it may even bring into question the UKGC’s role in data gathering in the first place.
Perhaps we should be asking the financial institutes to monitor their own data and to send out their own surveys on gambling involvement to individuals who are seen to be gambling, and who do perhaps have associated harms showing in other financial data. That way we could perhaps gain a better understanding of what motivates individuals to continue gambling even when it is seen as being harmful to their financial wellbeing.
Perhaps by doing so we could shift the whole “safer gambling” process on its head and make it far closer to a Public Health driven model.
How about this for a radical idea – have the financial institutes apply their own affordability criteria to customer’s spending habits and have them decide whether or not to impose gambling blocks. After all, the whole point of introducing affordability checking into gambling is that it already exists within the financial and credit sectors – albeit for the exact reversal of reasons.
This brings up a very important observation.
Financial institutes have a far greater business interest in protecting their customers from harmful gambling than any gambling operator ever will have.
Does this mean that we are putting the wrong emphasis on where regulation for protective reasons should be happening?
Worth discussing, don’t you think?