Research out of the University at Buffalo by John Welte and colleagues suggests that gambling problems — pathological gambling, to be specific — are more problematic than alcohol dependence in older adults. Some of the findings are interesting.
But one finding stood out for me as being a bit sensationalistic. That finding was that pathological gambling — something other studies have consistently pegged in the 0.8% to 2.0% range of adults (see Stucki & Rihs-Middel, 2007) — is more common than alcohol dependence (which studies put in the 3.8% range, see Keyes et al., 2009). Past research has shown that alcohol dependence (also known as alcoholism) is something that occurs in the adult population at twice the rate of pathological gambling.
In Welte’s (2011) study, however, the researchers found something different altogether. They found that from age 22 onwards, pathological gambling is more prevalent than alcoholism. And in the age 31 to 40 group, they found it nearly 3 times as prevalent (at over 5 percent of that age group)!
So what’s going on here? What could account for this significant discrepancy between this new study and much of the previous research?
How Researchers Define Something is Key
One of social science researchers’ dirty little secrets is that they can make statistical data pretty much show anything they want, as long as they design the study correctly from the start. And there’s no better way to do this than to create definitions that favor your hypotheses.
For instance, if you’re a researcher who studies a lot of gambling, you’re going to start off making sure that the definition of gambling you use is as broad as possible. While this is great for inclusiveness, it also means that the “problem” of gambling is going to cast a much wider net, including things that most of wouldn’t ordinarily even consider gambling.
How did the current study’s researchers define gambling?
Both surveys included questions on the frequency of past-year gambling on specific types of gambling. These were: (1) raffles, office pools, and charitable gambling, (2) pulltabs, (3) bingo, (4) cards, not in a casino, (5) games of skill, e.g., pool, golf, (6) dice, not in a casino, (7) sports betting, (8) horse or dog track, (9) horses, dogs off-track, (10) gambling machines, not in a casino, (11) casino, (12) lottery, (13) lottery video-keno, (14) internet gambling, and (15) other gambling.
I’m not sure the researchers understand the meaning of the word “specific” when they then go ahead and list every possible type of activity that involves money and chance. As well as that very specific category we’re all used to playing, “other gambling.”
Would most of us consider playing the lottery “gambling,” in the same league as playing craps at the casino? It’s no wonder the researchers found that more than 70 percent of people “gambled” in the past year in all adult age groups. And that, when we reach our 30s, 25 percent or more of us gamble more than 52 times a year. To put you in that category, all you need do is buy one lottery ticket a week. Or play the college basketball brackets.
The other issue is how researchers use a phrase — “problem gambling” — that isn’t actually defined in the DSM-IV (the psychiatric reference manual that defines mental disorders). Researchers have long used a largely arbitrary cutoff of three symptoms, out of a possible 10, to define this “problem gambling.” (A person needs five symptoms to qualify for a pathological gambling diagnosis.)
Here’s what the researchers in this study did:
Endorsement of five or more criteria is considered DIS (Diagnostic Interview Schedule, Robins et al. 1996) pathological gambling (APA 1994), and endorsement of three or more criteria was considered to be DIS problem or pathological gambling.
This has been done so often now, the researchers didn’t even bother with a rationale for using the number three.
Chasing down when and how exactly researchers began accepting three as the magic number to define “problem gambling” is like trying to research your family genealogy. Researchers keep referring back to other research, which refers back to other research, and so on. If you follow this path long enough, ostensibly you’ll get to a study that demonstrates this is a reliable and valid method for defining “problem gambling.”
The problem here is that, as far as I can tell, that study was never done.
For instance, Toce-Gerstein et al. (2003) states:
For our analysis, we labeled gamblers meeting three or four criteria “problem gamblers,” consistent with much of the existing literature (Lesieur & Blume, 1987; Shaffer et al., 1997; Abbott & Volberg, 1999).
Then you go and look all of those studies up and read them. The Shaffer meta-analysis divided gambling into three levels (Level 1 = no problem, Level 2 = subclinical problem, Level 3 = pathological gambling), but only because they found lots of studies with varying criteria that didn’t always match up quite properly. This was their way of modeling the problem in order to properly categorize and analyze the research. It’s also important to note that Shaffer et al. considered Level 2 to be “bi-directional” — that is, people could be moving toward becoming a pathological gambler, or they could be moving away from pathological gambling.
Abbott & Volberg refined this criteria in a 159 page report (not a peer-reviewed journal article) prepared for the New Zealand government. They suggested the DSM-IV criteria could range in the 3 to 4 symptoms for a present gambling problem. Somehow, over time, we see the criteria moving from “3 to 4” to just three. Three will naturally include more people.
The Lesieur & Blume study refers to the South Oaks Gambling Screen, a 20-item test that was later shown to greatly overestimate pathological and problem gambling (see, for example, Thompson et al., 2005). So that doesn’t exactly help us here.
None of these citations actually conducted a study to demonstrate that three criteria defined a clinically and statistically significant “problem gambling” category. They were just researchers who theorized these numbers “felt right.”
I suspect the reasoning went something like this: “If five is the current diagnostic cutoff for pathological gambling, and four was the old cutoff (prior to the DSM-IV), then three feels like a good number to consider something a ‘problem gambling’ behavior.”
Ostensibly, you could do this with any disorder in the DSM-IV. You could create a whole new set of “subclinical disorders” that suggest a person is “at-risk” and therefore would benefit from treatment. From super-mild depression (since mild depression is already a diagnosis) and “problem attention disorder”, to “problem bipolar disorder” and “problem anxiety disorder.” The list is endless! Better yet, you now categorize a whole class of people that previously didn’t have a disorder to one that does.
Put the two issues together — liberal “gambling” definition + arbitrary definition of “problem gambling” — and what you have is — ta da! — researchers finding a non-diagnosis (“problem gambling”) is far more common than an actual diagnosis (“alcohol dependence”).
Now, don’t get me wrong. I’m not saying that there might not be a continuum of gambling disorders and their severity. But in the current diagnostic schema, you either have pathological gambling or you don’t. You might be “at risk” for pathological gambling if you find you meet 2 or 3 or even 4 of those criteria. But you also may not be; the research can’t say.
Indeed, three might be the “magic number” that separates problem gambling from just people who enjoy gambling, but the research — to me — is unconvincing.
Gambling can turn into a serious problem in societies that offer it to their citizens. We benefit from studies that provide us an accurate understanding of the extent of the problem, and its comparison to other serious mental health issues and concerns. I’m not convinced the current study does this.
Read the full news release: Gambling Problems are More Common than Drinking Problems, Study Finds
Keyes, Katherine M.; Geier, Timothy; Grant, Bridget F.; Hasin, Deborah S.; (2009). Influence of a drinking quantity and frequency measure on the prevalence and demographic correlates of DSM-IV alcohol dependence. Alcoholism: Clinical and Experimental Research, 33(5), 761-771.
Stucki, S. & Rihs-Middel, M. (2007). Prevalence of adult problem and pathological gambling between 2000 and 2005: An update. Journal of Gambling Studies, 23(3), 245-257.
Toce-Gerstein, M., Gerstein, D.R., & Volberg, R.A. (2003). A hierarchy of gambling disorders in the community. Addiction, 98, 1661-1672.
Welte, J.W., Barnes, G.M., Tidwell, M-C., Hoffman, J.H. (2011). Gambling and Problem Gambling Across the Lifespan. Journal of Gambling Studies, 27(1), 49-61. DOI: 10.1007/s10899-010-9195-z