Sunday, December 28, 2014

Statistics

A bit dated from last summer...
Statistics can be like nailing jello to the wall in that the interpretation of Statistics, and quoted out of context Statistics leads to erroneous results.

We have all heard that if you have a gun in the home you are more-than-likely to be killed by a gun in the house ( sometime quoted, by that gun). This was based on reports from statistical data, but wait that is not all, there are questions to be asked. Is the murder rate of a home invasion higher for gun owners versus non gun owners. The environment where the gun owner lives...was home invasion murder with a gun in a higher risk than those who were not. What is the ratio of home invasion murders with a gun between gun owners and non gun owners in those same high risk areas...was there a member of in home involved with gang violence.

Back to the quote you heard often, if you have a gun in the house you are more likely to be killed by the gun is false. The data was isolating gun owners and leaving out an important fact, according to John Lott, that the guns are being brought into the home, it was not the gun owner's personal weapon gun that was involved in their home invasion murder.
Statistics can be like nailing jello to the wall in that the interpretation of Statistics, and quoted out of context Statistics leads to erroneous results.

You have to be careful evaluating reports from all sources. The integrity of reporting data was lost along time ago. One has to consider who is sponsoring the research and what are they mining for. Then how was the data collected and who sponsored the data collection, and what are the controls pertaining to the data collection. Of course, there are no endless funds for wholesome research, so you need to make the most of what you have without bias, and you do this with a critical analysis...still knowing what you are working with is less than perfect--who is going to do that? Many times that is left to us.

Anyway, I do not totally agree with everything that John Lott reports, but he is one of the few that pushes back and says, hey, let's take a look at that data, how did you come up with those results, and is there another way to interpret that data. I ran across this letter to the editor conversation below.

==========

First, Stephen DunLop, MD is the president of Hoosier Concerned about Gun Violence and he recently wrote an article in the "IndyStar" in Indianapolis which was refuted by John Lott with a letter to the Editor, quoted below:

"Stephen Dunlop, with Hoosiers Concerned about Gun Violence, makes a number of mistakes in his Aug. 15 letter, “Gun violence remains a public health issue.” In talking about the risks of guns in the home, Dunlop ignored that the research he cited assumes that if a person was killed and a gun was owned in the home, it was the gun in the home that was responsible for the death. In fact, virtually all of those deaths were due to guns being brought in by criminals getting into the home. For one of the papers in the meta-analysis, in only eight of the 444 homicide cases was a “gun involved (that) had been kept in the home.” Nor do the studies separate homes of gang members from those of law-abiding citizens.

As to Dunlop’s claim that my research is “discredited,” if he had looked at the literature, he would have discovered that about two-thirds of peer-reviewed research by economists and criminologists find that right-to-carry laws reduce violent crime. And no one finds higher murder, rape or robbery from concealed handgun laws.

John R. Lott Jr.

President
Crime Prevention Research Center"