New study finds connection between gambling addiction and lootbox purchases in adolescents

    
24

The research and science continues in the case of lockbox/lootbox purchasing habits and gambling addiction. The latest piece of scientific literature on the subject has been published today in the form of a new study that seems to have found habitual connections between lootbox purchasing habits and gambling addiction specifically in adolescents.

The paper comes by way of York St. John University and University of York researchers. A synopsis shared by one of the journal’s authors, media effects researcher David Zendle, breaks the findings down further, explaining that the more money older adolescents spent on lootboxes, the greater their problem gambling severity appeared to be. The study also found several motivations for lootbox purchases, including a desire to get skins to fit in and the enjoyable feeling of simply opening a box.

“A large-scale survey of 16- to 18-year-olds (n = 1155) found evidence for such a link (η2 = 0.120). The link between loot box spending and problem gambling among these older adolescents was of moderate to large magnitude. It was stronger than relationships previously observed in adults. Qualitative analysis of text data showed that gamers bought loot boxes for a variety of reasons. Several of these motivations were similar to common reasons for engaging in gambling. Overall, these results suggest that loot boxes either cause problem gambling among older adolescents, allow game companies to profit from adolescents with gambling problems for massive monetary rewards, or both of the above. Possible strategies for regulation and restriction are given.”

A similar connection between amount of money spent on lootboxes and an increase in problem gambling behavior in adults was found in a UK study in September, which should be no surprise, as it shares its first author with this new paper. Another study from the UK Gambling Commission found nearly a third of the adolescents it surveyed bought lootboxes. The German Youth Protection Commission also raised some alarm over lootbox practices in February.

With all of that said, Zendle still argues that further research is needed. “Why is it stronger in this group as opposed to adults? We’re not sure,” he writes. “This population is known to be especially vulnerable to developing problem gambling. It may be to do with this – but significant further work is needed to investigate.”

24
LEAVE A COMMENT

Please Login to comment
  Subscribe  
newest oldest most liked
Subscribe to:
Reader
rafael12104

Heh. Another log for the pyre. Yes, I mean pyre.

Yes, one day hell will freeze over and all that seems trite will be prophetic.

Ye shall reap what you sow!

Reader
Danny Smith

you mean the garbage thats designed with such mindsets as “show one frame of the highest rarity reward for a minor endorphin rush to help form the habit” is predatory?

But the addicts i know said they were just kinder eggs/trading cards/shut up i can quit any time i want?

Reader
Axetwin .

There is a difference between common knowledge and empirical data. This study turns that knowledge into the data. It’s something concrete and something people can cite as a source the next time someone wants to try to regulate lootboxes. I’m only explaining this because it seems like many here in the comments don’t understand that.

Reader
Bryan Correll

I understand that. I also believe that Duh! Monthly would have been the appropriate journal for publication.

Reader
Arktouros

Empirical evidence is information acquired by observation or experimentation.

Nothing was observed.

Nothing was experimented.

Throwing up a reddit survey and making a research paper is about as scientific as Facebook telling you that you’re a Samantha and not a Carrie.

Reader
Ashley Bau

Ehh… there is such a thing as scientific surveys and I believe that falls under observational research (it has been a minute though). I don’t know how THIS survey was conducted but in theory this is viable ‘supporting’ evidence (it supports correlation so a more robust argument would have to be formed and or experimentation would need to be conducted for a more solid case of causation).

Reader
Arktouros

There is such a thing as scientific surveys, but as the National Science Foundation points out you have to be pretty careful with them as there’s room for doubt when they aren’t conducted scientifically. I found this to be a pretty good read for guidelines when all these survey based “research” papers denouncing lock boxes started cropping up all over the place.

A lot of people are looking for a win on this topic because it frustrates them so they’re quick to latch onto anything that might lead that direction. That includes bad arguments that aren’t based on good science. If we want to use those arguments, we should get the real science to back them up and not just have some biased survey with an unrepresentative sample as the core of the argument. It’s weak and it will be dismantled by people who come up with terms like “surprise mechanics.”

Reader
Bryan Correll

I think I saw that study in Duh! Monthly.

Reader
Schmidt.Capela

“It is just so blindingly obvious that it almost doesn’t need to be stated, but apparently, it does and therefore I have done.” – John Bercow

This was said in a very different context (it was about how no Prime Minister would be allowed to just suspend the UK Parliament and push no-deal Brexit without parliamentary oversight), but it’s appropriate here too; this kind of research shouldn’t be necessary because the expected results are blindingly obvious, but since the corporate lobby disregards even common sense to push for their interests, research such as the one covered here is nonetheless needed.

Reader
Arktouros

Straight from the research article:

No research has yet examined why adolescents buy loot boxes.

Absolute waste of time as usual. The entire research is founded upon the assumption that people are only buying lock boxes because they want go gamble. This is not why most people buy lock boxes. Most people who buy lock boxes do so because the company is shoving exclusive rewards into them that can only be gained by buying the lock boxes.

If an actual research study was done rather than some shitty reddit survey (albeit I did laugh at them mentioning the person who “incorporated an abusive message to the researchers” or them disqualifying the people who said they dropped $1 million as if that’s any less believable than the rest of the results they got) you’d see that if you gave people an option to buy something they wanted vs cramming it in some gamble box they would buy the thing 9:10 rather than buy the boxes. It is entirely a problem with the availability of the product.

Doing any kind of research paper about lock boxes and gambling and not even doing a study on why people buy lock boxes is absurdly negligent.

Bree Royce
Staff
Bree Royce

Ark. The part you’re quoting, 1.4, is pointing out that nobody has done this research until this paper.

The section you’re looking for is “3.3. Qualitative analysis of motivation data,” which specifically examines the reasons the subjects gave for buying loot boxes, including cosmetic appeal, supporting devs, peer-pressure and fitting in, pressure to compete, compulsion to get everything, compulsion to get limited edition stuff, joy of opening boxes and getting unknown stuff, fear of missing out, value, skipping grinds, faster progress, profit, and so on.

Reader
Arktouros

Taking a random reddit survey isn’t research nor is it a study. The data is entirely questionable. I could tell you that I took that survey they put and lied and said I was a 17 year old Fortnite player and you or the survey has no way of actually determining if that is true. None of those people even could have actually paid for loot boxes before and were just trolling answers. You’re talking about a thousand responses in a gaming market of hundreds of millions of people. Internet Online surveys aren’t research, how many Boaty McBoatfaces do people really need to understand the concept?

Get gamers. Put them into test scenarios. Record the data. Use control groups and non-control groups. Get actual testing data from actual people doing the actual things you’re reporting as research. It’s not hard. It’s just expensive, and more importantly it’s most likely not to prove the results that people want.

Reader
traja

You don’t seem to understand the first thing about statistics or research. It makes no difference how many gamers there are in the world. All your criticism should be focused on the sampling method because the sample size is absolutely fine here. Also bringing up a public vote for a name is entirely irrelevant.

Now I agree that Reddit has issues as a sampling method. Just not for the reasons that you think it has issues.

smuggler-in-a-yt
Reader
Kickstarter Donor
Patreon Donor
Loyal Patron
smuggler-in-a-yt

Yeah, this. If it made it into the RS journal and went through peer review I’m satisfied someone did the due diligence on the collection method.

People will lie about everything, good PIs take that into account.

The problem is that a digital world has skewed a lot of what people used to consider a “full and valid dataset”. Hell, I just read something out of Harvard that made some pretty interesting assertions about policy based on an analysis of GitHub of all places.

All good science should raise more questions than it answers. I’d hope the same is true of this study.

Reader
Schmidt.Capela

You just have to figure how to map the population you are able to survey into the population you want to study. If you can find a close enough mapping between the populations then you can use your survey results to draw reliable conclusions about the wider population regardless of the source of the surveys.

Of course, determining that mapping usually is neither simple nor easy.

Reader
Arktouros

I know enough that throwing out a reddit survey doesn’t and shouldn’t qualify as research.

If this were actual research, as in an actual empirical data study was being done, then scale wouldn’t matter. However when you’re taking a survey it’s important that you get enough participants to reflect the actual population. 1000 people in a market of tens of millions isn’t even remotely close to an accurate picture. Even if we just assume a nice even number like only 1 million gamers are children under 18 that’s still 0.001%.

Just absurd.

Reader
traja

You really need to take a course in basic statistics because that is not how this works. For the size of the population you are sampling to be a significant factor your sample needs to be a few percentage points of it. Basically absolutely ridiculous sample if you are sampling millions.

1000 is 0.1% of 1 million.

Reader
Arktouros

You don’t need to sample millions that would be equally absurd however your sample size can’t too small either.

If I threw up a survey on Reddit asking Americans if they like Anime or not and 60% of 1100 people said yes making the claim that 60% of all Americans like Anime would be just laughed at.

A few percentage points seems like it would be a more reasonable amount, but 0.1% (sorry about the 0.001% it was super late when I wrote that) isn’t even remotely close to where it needs to be by at least a factor of 10.

But as usual people are so hungry for there to be any kind of progress on this topic they’re willing to latch onto anything reaffirming their assertions without questioning it.

Reader
traja

No you are just wrong about this. It is incredibly obvious that you haven’t ever taken a full course in statistics because you are constantly wrong about it. It’s ok, we can’t all be experts on everything. All I have is a single university course in statistics from years ago, but it enough that I can immediately spot that you don’t know this stuff.

Countless studies are done using only a few dozen people and the population being studied can be 7 billion large.

The real problems with these studies is almost always that the sample is not necessarily representative. That is the problem in your example study too and not the sample size of 1100, which is actually really good. It is the problem with this lootbox study as well although it’s not quite as bad as your example. When using Reddit polls it will always be impossible to determine if your sample is representative. So it’s not that this sample is necessarily not representative but that it’s not clear how representative it is.

All that means is that this study on its own is not enough to draw definite conclusions. What it does though is add to the body of science on the topic. Once there are several studies and there is a clear direction that the majority of studies on the topic seem to point towards it becomes very difficult to deny.

Reader
Bruno Brito

Take the loss, Ark. You got owned here.

Reader
rosieposie

Congrats on the most predictable post of the year, btw. I saw it before my eyes even before I had clicked on the article.

Reader
Arktouros

Cool.

PurpleCopper
Reader
PurpleCopper

In other news, water is wet.

Reader
Randy Savage

.

you-dont-say.jpg