How valid are the findings?Experts debate scientific value of Internet poll
By Charlene Laino
MSNBC
July 18 — The planned publication of an analysis of the MSNBC “click-and-tell” poll in a peer-reviewed scientific journal points to the validity of the findings, says study author Al Cooper of the San Jose Marital Services and Sexuality Centre in San Jose, Calif. But questions about its value as science remain.
THE SURVEY is a major step forward in the understanding of important new areas of online sexual activities, Cooper says. Several publications are planned based on the data, he adds, including one accepted for a fall issue of the journal Sexual Addiction & Compulsivity: The Journal of Treatment and Prevention.
But by its very nature, the sample is self-selected. In order to take the survey, one must have access to a computer and be online. Then, the participant must have visited MSNBC.com at least once to find the survey. And even then, probably only people with a fairly strong interest in cybersex are going to take the time to complete all 76 questions — which is required or the whole questionnaire discounted.
Vanderbilt University’s Donna Hoffman has told MSNBC.com in the past to get a true snapshot, one would have to sample a random cross-section of the entire Internet population, she said. “Cooper’s final analysis was [predominantly] male,” she noted, “while men only make up about half of all rs.”
Such polls “are entertaining,” Hoffman, a specialist in Internet marketing and electronic commerce, said of the 1998 MSNBC.com cybersex survey. “But as science, it is silly.”
(MSNBC.com notes that by their very nature, surveys posted on its Web site are self-selected and unscientific, not the random samples utilized by polling organizations.)
statistician Robin Mathy, a graduate student at the University of Oxford in Oxford, England, who analyzed the results, said the design of the new survey is a vast improvement over the ’98 poll. As a result, she says, the answer are much more reliable.
“All surveys involve self-selection,” she explains. “The real question to pout to the test is two-fold: Is the survey heterogeneous and it it representative.
“We can’t claim that people who came to MSNBC.com and took the poll are a representative sample of all people,” Mathy acknowledges. But we did take steps to overcome the limitations.
For example, she says, eight of the 76 survey questions were designed simply to weed out people who were unreliable: One asked “I am the most intelligent person I know, True or false,” for example, while another asked whether the user spends 5,000 hours a year online, which is impossible.
People who answered four “suspect’ question in an unreliable manner were e excluded form the final survey population, she says.
Los Angeles therapist Rob Weiss says that while the survey might have over-estimated the number of cyber sex addicts, it nonetheless provides useful insights, many of which are on target.
It makes sense that nearly 10 percent of the survey population indicated an addiction, he says, when you consider that people who filled it out not only were already on the Web but also had enough interest in the topic of online sexuality to take the time to answer 76 questions.
“So the numbers obtained from the survey are absolutely believable for that population,” he says, “just might be a little high for the population at large.”
Mathy notes that online survey have been shown, in some cases, to be more reliable than traditional phone polling. For example, studies on HIV counseling have showed that people are less likely to lie on sexual questions when anonymously giving answers to a computer than when talking to someone.
SUGGESTS RESEARCH AREAS
Weiss and Cooper’s team agree that the survey serves a valuable public health service, offering insight into a potentially destructive sexual disorder.
Told of some of the findings, Weiss aid, “these are absolutely believable. This work and Cooper’s former poll provide information that may be of use in understanding cyber-addiction.”
Noting that the survey is the most comprehensive survey of online sexual pursuits to date, Cooper agrees tit serves as a wake-up call to consumers and health care professionals that cybersex addiction is a real problem, even if for only a minority of Netizens.
“It’s not the perfect randomized controlled trial, but in my opinion it does a service,” he says. “Now that we have begun to define the scope of the problem, we can work together to help those in danger,” he said.
Click to expand
Posted - Jul 17 2002 : 12:03AM