Drunken Research?


Friday, June 17, 2011
For most people, it’s just common sense that it is never a good idea to consume alcoholic beverages to the point that one’s judgment and physical capabilities are impaired, particularly if one is going to operate a vehicle or other machine, such as a firearm.

But for Garen Wintemute, the longtime gun control-advocacy junk researcher at the University of California at Davis, who promotes himself as “one of the world’s foremost experts on gun-related violence,” talking about alcohol and guns in the same breath is just another way to do what he most likes to do–expressing his feeling that guns and gun owners are inherently unsafe–and getting paid for doing it.

His new “study,” published by Injury Prevention, pushes the idea that gun owners, particularly those who keep and carry firearms for protection, are more likely to drink heavily and often. Not surprisingly, it was paid for by the anti-gun Joyce Foundation, which has given $26 million to anti-gun groups and causes between 2003 and the present, and the California Wellness Foundation. And to make sure the money keeps coming in, Wintemute says “new and more comprehensive [funded] research is needed, since legislation authorizing the public carrying of loaded and concealed firearms has become almost universal in the United States.”

There aren’t too many jobs in which you can get paid hundreds of thousands of dollars for doing not much more than punching the time clock and making sure the boss knows about it. But Wintemute’s sugar daddies and mommas at the anti-gun foundations don’t seem to care. Wintemute even admits the limitations of his latest diatribe.

For starters, Wintemute’s poll depends on the truthfulness of the poll’s respondents. Undoubtedly and understandably, habitual drunkenness is viewed with disrespect in our society, and some people won’t admit such behavior to a stranger conducting a telephone poll. Similarly, people may be unwilling to admit gun ownership to strangers, to protect themselves against being burglarized or placed on gun owner registration lists. It’s not mere coincidence that national polls began showing a significant decline in gun ownership in the early 1990s, as the Clinton Administration’s multi-faceted campaign against the right to arms took shape.

Also, the eight states Wintemute used for his study were self-selecting on a basis he does not disclose. Polls suffering from this limitation are referred to as “SLOP,” for “self-selecting opinion poll,” because the self-selection process can inject bias into poll question responses and severely skew the poll’s findings.

Wintemute’s study doesn’t even try to determine whether there is any greater incidence of mishaps or misconduct, firearm-related or not, among gun owners who consume alcohol. But that might be for good reason. After all, as Right-to-Carry laws have become “almost universal,” the nation’s murder rate has fallen to a 47-year low. Maybe Wintemute understands the meaning of the old axiom, “if you don’t want the answer, don’t ask the question.”

Copyright 2011, National Rifle Association of America, Institute for Legislative Action.
This may be reproduced. It may not be reproduced for commercial purposes.
11250 Waples Mill Road, Fairfax, VA 22030    800-392-8683

A little more on SLOPpy polling: Some of you may be wondering what makes a poll self-selecting or SLOP. If someone calls you on the phone and starts asking poll questions, you have several options as to how you answer. You can choose to hang up on the pollster. If you do this, you’ve just self-selected yourself out of the polling sample. You can answer the questions, but lie through your teeth when it suits you. Once again, you’re self-selecting. In this case, you’ve decided to give false data; however, the pollster has no way of knowing that you’re lying. Finally, you can answer the questions truthfully. This is also an act of self-selection. In the first two cases. the pollster either gets false data or loses good data. These false and missing data points bias the study; the sample set is no longer random.

To be valid, a study like this must rely upon data that respondents cannot hide or lie about from a truly random sample. In this case, Wintemute could have found and used just such a data set. One can easily postulate that arrests for drunk driving or public intoxication are an indication that a particular individual drinks heavily. Such an arrest record is something that cannot be subject to self-selection. Furthermore, having a CCW permit on file is also something that cannot be hidden. Those two record sets, persons with arrest records and persons with CCW permits, can be checked against one another to see if there is any correlation between those two groups. And, in fact, they have. Earlier studies have found that CCW holders are more law abiding than their unarmed neighbors. This is old news published long ago by John Lott and others.

But of course, one cannot expect a check from the Joyce Foundation for merely restating old and undesirable data.