If the study linked by JAK is, indeed, the study referenced there are a couple items of note:
1. The abstract from the article says that BRFSS data from eight states was reviewed. Twenty-nine states, or more, have been participating in this study since its inception in 1981. Why was data from only eight states reviewed? Excluding participating states creates a bias. It may not have been an intentional bias but one was created, nonetheless.
2. Self-reported data is notoriously flawed. Cook and Campbell (1979) pointed out that subjects (a) tend to report what they believe the researcher expects to see, or (b) report what reflects positively on their own abilities, knowledge, beliefs, or opinions. Another concern about such data centers on whether subjects are able to accurately recall past behaviors.
3. A sample size of ~16,000 is relatively small. Recent estimates suggest that somewhere between 43 and 55 million households own guns. Let's be conservative and say that equates to ~43,000,000 gun owners. It would be difficult to achieve a confidence interval (ability to accurately predict an outcome or result) with such a small sample size.
In summary, even if Ozaki had cited the study in his report, in fairness to law-abiding gun owners he should also have mentioned that the study is based on relatively old, unreliable data with an exceptionally small and biased sample.
