That's the first thing I thought when I read Philly. From all I've heard and read there's large parts of the city that are a wasteland. Manipulated (cherry picked) data is just G.I.G.O.
| Twin Cities Carry Forum Archive http://www.twincitiescarry.com/forum/ |
|
| Oh jeez, here we go again... http://www.twincitiescarry.com/forum/viewtopic.php?f=24&t=14297 |
Page 2 of 2 |
| Author: | havegunjoe [ Tue Oct 06, 2009 10:03 am ] |
| Post subject: | Re: Philidelphia study disembowled. |
Just as I tell people, 87% of all statistics are made up on the spot. |
|
| Author: | Judge [ Tue Oct 06, 2009 11:16 am ] |
| Post subject: | Re: Philidelphia study disembowled. |
[quote="havegunjoe"]Just as I tell people, 87% of all statistics are made up on the spot.[/quote ] 74% of people know that. |
|
| Author: | Scott Hughes [ Tue Oct 06, 2009 12:28 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
chunkstyle wrote: Quote: 40 cases is the minimum number considered to produce a statistically significant sample, depending on what shape the results take. The size of the sample after that doesn't really matter except in reducing the margin of error or increasing the confidence level, and size in relation to the population doesn't matter. But that assumes no errors in sample selection, and more significantly, no bias in sample selection deliberately induced to skew results to a desired end. This is when science becomes propaganda. Perfectly said chunk.....In a lot of instances this type of propaganda is targeted at that segment of the population that couldn't sort out the number(s) even with their shoes off counting their toes. The propagandists rely on people being scared by scary numbers. |
|
| Author: | kimberman [ Tue Oct 06, 2009 1:17 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
Not only did they NOT screen for unlawful possession, they did NOT screen for drugs/alcohol in blood at time of assault, NOR for gang membership. Those are the three variables that crop up most often in violent crimes (just ask the Criminologists). The MD's always cherry pick their data base in order to secure the desired result. This seems to be just another version of the original Kellerman study (43 times) with all the same flaws. Good refutation is contained here: http://reason.com/blog/2009/10/05/why-s ... -be-better |
|
| Author: | Scott Hughes [ Tue Oct 06, 2009 1:42 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
That's the first thing I thought when I read Philly. From all I've heard and read there's large parts of the city that are a wasteland. Manipulated (cherry picked) data is just G.I.G.O.
|
|
| Author: | AGoodDay [ Tue Oct 06, 2009 4:21 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
chunkstyle wrote: Quote: 40 cases is the minimum number considered to produce a statistically significant sample, depending on what shape the results take. The size of the sample after that doesn't really matter except in reducing the margin of error or increasing the confidence level, and size in relation to the population doesn't matter. But that assumes no errors in sample selection, and more significantly, no bias in sample selection deliberately induced to skew results to a desired end. This is when science becomes propaganda. Yup. I also forwarded this stuff to my statistics professor for his thoughts. He talked to me today and basically said he can see where they were trying to go with this, but their method still looked like crap, and they're drawing a conclusion from the information that is not appropriate to draw from the information that they have. Andrew, any chance you'll have a version of the article I would be able to get my hands on? |
|
| Author: | Andrew Rothman [ Tue Oct 06, 2009 6:39 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
Eventually -- looks like a couple weeks. |
|
| Author: | SethB [ Wed Oct 07, 2009 3:42 pm ] |
| Post subject: | Re: Oh jeez, here we go again... |
The statistic may be perfectly correct. It's still garbage. (See Simpson's Paradox.) Consider the following hypothetical society: 10% of people are gangsters. 90% of people are civilians. 90% of gangsters (9% of population) have guns. 10% of civilians (9% of population) have guns. 80% of gangsters without guns (0.8% of population) get shot. 40% of gangsters with guns (3.6% of population) get shot. 1% of civilians without guns (0.9% of population) get shot. 0% of civilians with guns (0.0% of population) get shot. Now, let's do some statistics. People with guns: 18% of population. People with guns who get shot: 3.6% of population (3.6%+0%) People without guns: 82% of population People without guns who get shot: 1.7% of population. So, overall, percentage of people with guns who get shot: 3.6/18= 20% Percentage of people without guns who get shot: 1.7/82= 2.07% So if you look at the overall percentages, it looks like you're 10 times more likely to get shot if you have a gun. But it's easy to see that if you're a gangster, you're less likely to get shot if you have a gun. And if you're a civilian, you're less likely to get shot if you have a gun. And if you're a doctor who doesn't understand statistics or cause and effect, having a gun makes you 40 times more likely to be a gangster. |
|
| Page 2 of 2 | All times are UTC - 6 hours |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|