Online gamblers are both more likely to have problematic habits and to answer online surveys

GambleAware Research Claims Online Problem Gambling Surveys Over-Report The Problem Due To Selection Bias

Problem gambling is over-reported, contends a recent survey. What’s more, online gamblers who have risky behavior are more likely to take Internet surveys, recent research found.

This flies in the face of conventional wisdom. Most problem gambling advocates say that rates are underestimated due to inherent problems with self-reporting. One thing, at least, is clear, which is that the apparent rate varies widely depending on what methodology one uses.

The survey in question, by UK’s GambleAware, concludes that problem gambling is over-reported, and that online surveys are to blame. According to the report, released May 14:

“The key finding of the research is that surveys using predominantly or entirely online self-completion produce consistently higher estimates of gambling harm compared to surveys using a paper self-completion questionnaire as part of a face-to-face interview.”

Problem gambling advocate Jamie Salsburg sees a bias in the opposite direction, he told Online Poker Report on Monday. Problem gamblers are more likely to lie one-on-one, he says. Those face-to-face interviews thereby skew the numbers towards indicating less of an issue than really exists. Salsburg is himself a recovering gambling addict who last wagered on July 15, 2010.

Salsburg, who writes a blog and hosts the After Gambling Podcast, said embarrassment leads problem gamblers to downplay their habits when asked about them face-to-face. Lying about gambling is in fact one of the diagnostic criteria for the disorder.

Keith S. Whyte, executive director of the National Council on Problem Gambling, declined to comment on this study.

Online surveys are more likely to self-select online gamblers

Infrequent gamblers and those who use retail channels exclusively are less likely to take an online survey, GambleAware found. Online bettors are more likely to take Internet surveys, and also more likely to show indicators of problem gambling.

That means the results of those online surveys will appear to show that the crisis is more widespread than it actually is, according to GambleAware. This is what’s known as selection bias.

To quote the study:

“Disproportionately high numbers of online and frequent gamblers mean the online surveys over-estimate gambling harm because online and frequent gambling are independently associated with a higher risk of gambling harm.”

Being an online gambler creates the highest risk factor for problem gambling at 5.4%. That’s a substantial jump from the No. 2 indicator, which is wagering twice or more a week, the research found. The latter corresponds to a 4% increase in probability of problem gambling.

Even so, GambleAware recommended that problem gambling surveys should migrate to being conducted mostly online. That’s because of the need for large sample sizes, made difficult by the cost of in-person surveys. It stresses, however, that researchers should work diligently to find ways to correct for selection bias and arrive at rates that reflect the prevalence in general society.

- Heather Fletcher is the lead writer with OnlinePokerReport. She's a career journalist, with bylines in The New York Times, Adweek and other publications. Reach her at [email protected]
Privacy Policy