by saxitoxin on Thu Dec 27, 2012 1:22 am
Metsfanmax wrote:saxitoxin wrote:Metsfanmax wrote:These aren't reasons to reject the study, they're just reasons to not be confident in the specific result unless you live in Philadelphia.
1. It's reason to reject the conclusion of the study, which states "the probability of success may be low" without the necessary qualification "in the singularly most crime-ridden city in the United States."
That fact doesn't substantially affect the result. The question being asked was not "how likely are you to get shot when possessing a gun," but rather "how much
more likely are you to get shot when possessing a gun." Only the former depends on the absolute crime rate.
As I stated, it doesn't affect the data result, it affects the conclusion. The conclusion being "the probability of success may be low" with the implication of the discussion being this conclusion is valid outside the city limits of Philadelphia. Unless Philadelphia represents the demographic mean of all U.S. urban, suburban and rural areas - which it does not in at least one category (absolute crime rate), and most likely in many others as well - the study is of marginal value. This point is aside, however, from the more important fact that ...
Metsfanmax wrote:Metsfanmax wrote:The 4.5 times more likely result was given after already having taken into account general characteristics that made a person at higher risk to be involved in gun violence. It does not seem to have taken into account the difference between legal and illegally owned guns, but I would argue that this is subsumed in the general "high-risk" category they describe in the paper.
2. I can't comment on this as I don't have access to the full paper and "taken into account general characteristics that made a person at higher risk to be involved in gun violence" isn't explained - or as far as I can tell, even contained - in the abstract.
http://ajph.aphapublications.org/doi/pd ... 008.143099
This confirms the suspicion I originally stated -
However, compared with control participants, shooting case participants were significantly more often Hispanic, more frequently working in high-risk occupations, less educated, and had a greater frequency of prior arrest. At the time of shooting, case participants were also significantly more often involved with alcohol and drugs, outdoors, and closer to areas where more Blacks, Hispanics, and unemployed individuals resided.
You can't do a study of gang-on-gang violence and then imply the mortality figures for street soldiers in a turf war between the Latin Kings and the Surenos will also be representative for a mercantile exchange broker who has a 9MM in his house safe.
next study
