NON-STANDARD REPORTING METHODS COVER UP FLAWS IN HARVARD STUDY
Look at the explanatory note that accompanies Table 3:
[From the Table explanatory notes]
"Note. For ease of comparison, similar populations were obtained by comparing the 4 states with the lowest gun ownership rates ("low gun states") and the 6 states with the highest gun ownership rates ("high gun states"). The 6 states with the highest average gun ownership rates for 1988 to 1997 were Louisiana, Alabama, Mississippi, Wyoming, West Virginia, and Arkansas. The 4 states with the lowest average gun ownership rates for 1988 to 1997 were Hawaii, Massachusetts, Rhode Island, and New Jersey."
What is misleading about that approach? It uses a non-standard and unnecessary reporting method so that they can conflate the data from somewhat disparate states that have very different homicide rates. The standard method (used in sociology, criminology, medicine and other fields) is the use of rates per 100,000 residents, in order to compensate for different population sizes in different jurisdictions. Using the "rate per 100,000" standard method would allow you to examine each state's homicide rate separately, with no need to aggregate. If you did this, you would find that "high gun ownership" Wyoming and West Virginia have pretty low homicide rates. That might cause people to ask why the authors' paradigm doesn't seem to work with all the states--and they don't want people to ask that sort of question. For example, it is somewhat embarassing for Harvard to explain why the U.S. homicide rate (5.6 per 100,000) is 255% of the homicide rate (2.2 per 100,000) in heavily armed West Virginia. Harvard finds it much more convenient if they can just hide the West Virginia stats within a larger grouping where it doesn't need to be explained. Heck, it won't even be noticed by the average incompetent media hack or politician or voter.
AN ADDITIONAL DECEPTIVE APPROACH:
Note that Table 3 uses homicides aggregated over a 10 year span, which has the effect of impressing the casual observers with totals 10x higher than the typical year. If deception were not their intent, they could have simply used a 10 year average, to approximate the yearly losses. This aggregation of many years to yield a single number is a common tactic used by the anti-self-defense fanatics in order to inflate the numbers so that they will look far more terrifying than the real numbers. A similar tactic was used in the Violence Policy Center's "study" ("A License To Kill") of disqualified Texas CCW/Self-Defense-Permit holders.
The aggregation and conflation tactics of the "researchers" also fails to note that the 4 lowest gun ownership states are Northeastern states with one laid-back island-state thrown in, while the 6 highest gun ownership states are all Southeastern states with the exception of Wyoming. Innumerable researchers have found that the Southeastern culture is one of macho and honor, where people are culturally disposed to violence of all types. It is the region that gave our language the phrase, "Some people just need killing". I say this not to harass or insult Southeasterners, but simply to make an observation of a fact that is relevant to this study. I also observed, in a different thread, that the homicide rate in America is generally highest in the southern tier of states, with less in the middle, and the lowest homicide rates in the northern tier. That low homicide rate even occurs in northern states that are heavily armed--like Wyoming, Washington, North Dakota, Minnesota, South Dakota, Maine, Vermont, etc. You will note that this model provides a better fit of the data than the Harvard "study".
Also, as mentioned on the other thread, the entire study is almost useless because the use of proxies to determine the rate of firearms ownership is highly deficient. See my message 166 on the other thread for details.
http://www.freerepublic.com/focus/news/800686/posts?q=1&&page=151