Laelia wrote:
Jubbergun wrote:
I'm inclined to agree with that, mainly because it's hard to randomly select a sample large enough to get an accurate representation while still using accepted methods for accounting for 'unreported' incidents...which, if I understood correctly when I read it, is what accounted for the figures of 2 to 2.5 million instances of successful defense.
Your Pal,
Jubber
The main problem is that you have so many negative responses that even a very tiny false positive rate has major effects on your data. If your rare event affects 0.1% of respondents, and you have a 0.5% false positive rate, you would get 0.5% * 99.9% = 0.4995% false positives. This is 5 times higher than your rate of true positives. Even if you have a very high false negative rate (eg. 50%), you would be overestimating the true rate of whatever you're measuring by several times. Controlling a survey to have few enough false positives is very difficult.
None of this really affects the argument you're having, I just think it's an interesting statistical quirk of this type of survey. This is part of the reason regular prostate exams and mammograms for younger men and women are a bad idea - the actual rates of these cancers in younger people are very low, and both tests have some level of false positive results, so you end up incorrectly detecting cancer in many people who don't have it.
I think that if the rebuttal to the study you posted is accurate, they further compounded the possibility of error with their method of survey. While I disagree with the author that having a majority of samples from certain areas skews the result, since areas he may have thought required representation would reflect areas with legally restricted firearm access (California and many major metropolitan areas), preferring that the respondent be the male head of household lacked the necessary randomness and impartiality to which the study should have adhered.
I decided against buying a handgun even after reading the study...but I did get an extra aluminum bat.
Dvergar wrote:
I like how you edited out the names those targets belong to. Just because the Democrats did something wrong doesn't make it suddenly ok for you to do it to. I think most of us learned that when we were 4 with a story involving a bridge.
I didn't edit anything out. All three of those are hotlinked from another site. It's not like I could run over to a Palin website and grab a copy since she had the good sense to pull it. If you know where there's a copy with the names attached, post it.
Your Pal,
Jubber
