Dr. Robert V. Kail
Professor of Psychological Sciences
Purdue University PSYC 1248
West Lafayette, Indiana 47906
Dear Dr. Kail:
It has come to my attention, through a published report, that the findings and conclusions of the study prepared by Dr. Douglas Gentile, and published in an upcoming issue of Psychological Science, were based on flawed methodology. This raises serious doubts about the validity of findings. I am writing to make sure you are aware of this error so that appropriate action can be taken.
The concern -- acknowledged by Dr. Gentile to Mr. Gary Langer, director of polling for ABC News -- arises from the fact that the sample group for the study was not randomly chosen (see attached). It was a “convenience” sample of individuals who agreed to participate in the survey.
As you are likely aware, such a sample is not truly representative of a national population group. Thus the results cannot be projected onto the broader population of children in this country. And the sampling error of plus or minus 3% that Dr. Gentile cited in the study is also meaningless.
In his article, Dr. Gentile wrote that his findings are “nationally representative within 3%.” This claim is not verifiable given the nature of the sample population. Mr. Langer also spoke with a representative of Harris Interactive, the highly reputable polling firm that provided the data on which Dr. Gentile based his study. The representative, Dana Markow, reportedly told Mr. Langer that in order to compute a sampling error, a survey requires a probability sample. And the survey on which Dr. Gentile based his findings was “not a probability sample.”
Dr. Gentile’s study generated considerable news coverage around the nation, virtually all of which centered on his claimed finding that 8% of youth between 8 and 18 years old exhibited “pathological” patterns of game playing.
A press release from the National Institute on Media and the Family, with which Dr. Gentile is affiliated, contained this headline: “New Study Finds Nearly One out of 10 Young Gamers May Be ‘Addicted.’” The dramatic assertion generated headlines in many newspapers, including the Washington Post (see attached).
It is safe to say that the sole reason the study generated the kind of media attention it did was due to the inclusion of specific numbers that would appear to have been based on scientific research. In fact, the numbers reflected no such thing. Because of the composition of the group studied, neither the overall figure, nor the cited sampling error is supported by the data Dr. Gentile presented.
ABC News reported that Dr. Gentile said:
“he was unaware the data in his study came from a convenience sample – ‘I guess I’d assumed [Harris Interactive] had gathered the population initially as a part of a random probability sample.’”
We accept Dr. Gentile’s admission of error at face value, although it is hard to understand how a researcher would base a scientific study upon an assumption about the nature of the group he was studying. It is not that Dr. Gentile did not have time to make sure that the group was a truly national representative sample: the data was collected in January, 2007.
The admission is especially ironic considering that the first words of the abstract in the article went out of its way to note the shortcomings of previous convenience studies:
“Researchers have studied whether some youth are ‘addicted’ to video games, but previous studies have been based on regional convenience samples. Using a national sample, this study gathered information about video gaming habits and parental involvement in gaming…”
Based on the public comments of both Dr. Gentile and Harris Interactive, we are requesting that any references to the study in your publication and on your Website, clarify the methodological flaws in Dr. Gentile’s study and inform your readers how those flaws affect the accuracy of the study. I understand that the issue in which the study will appear may not yet have been published. If that is the case, I would ask you to note the deficiencies in the methodology in the upcoming issue and in any press materials you may make available with publication. Failure to do so will inevitably lead your readers to believe information that is not accurate.
I have no doubt that you value your publication’s credibility and reputation. Therefore, I hope this clarification is made quickly so that future readers of your publication are informed that the claims made by Dr. Gentile are not supported by the survey he has based them on. It would be unfair and misleading for a respected publication to leave on the record such knowingly mistaken information.
I hope to hear from you about this matter shortly, and I look forward to your publication’s response.
Very truly yours,
Michael D. Gallagher
President and CEO
Entertainment Software Association
Cc: Professor Douglas Gentile
National Institute on Media and the Family
Media