A large fundamental problem with attempting to capture the effect of an increase in gun ownership to crime rates is the data available on guns. I read through the posts quickly, since it's finals week so I may have missed similar responses, but I saw a few mentions of Lott's More Guns, Less Crime, which I recall generating lots of fanfare when it came out a few years ago. I am not entirely sure how Lott got his measurements on actual gun ownership, since it is not readily available or accurate, and whether or not he attempted to separate handguns (involved in more crimes) from other types of firearms. An economist, Mark Duggan, has a working paper called "More Guns, More Crime" where he uses a proxy variable of subscriptions to Guns N Ammo magazine to represent increases in gun ownership, and since Guns N Ammo mainly showcases handguns, it seems like a good proxy. He spends some time showing results from FBI reports and General Health Statistics reports to show that increases in the magazine subscription have a direct relationship in increases in gun accidents that usually stem from increases in gun sales. I know it's a bold assumption to make, but if you can get ahold of the paper (sorry I only have a hard copy), you can see that he makes a strong argument that an increase in guns leads to an increase in crime. Unlike a lot of pay for results studies, and since he's an economist, he includes several potential limiting factors and insight into why this may not be 100 percent correct, such as when measuring increases in gun ownership, it is very hard to capture the exact timing of purchase as to did the crime occur and then the purchase for defensive reasons, or did the purchase precede the crime. The paper is definitely worth a read.
The problem here with electing to participate in the monitoring process is that you lose your random sample and ability to control for data manipulation. The basis of the Nielsen system, and statistics in general, is that randomness is essential in approximating the behavior of a population, and important (with this issue) to limit networks ability to manipulate data. While you raise the prospect of "everyone" electing to participate, it is highly unlikely to happen, especially among the skittish/. crowd who would be required to give up those rights to privacy they hold so dear to keep their shows on TV. If the industry based its selection of shows off of people who "choose" to participate, then perhaps we would have an endless supply of sci-fi from the geek crowd (probably not with aforementioned geek privacy issues), or even worse/better off depending on your view, the networks would be able to jack their ratings by "electing" to view their own shows, thus creating fake ratings for advertisers while erasing any ability to target demographics. This would lead to telecom style price wars on advertising (due to demographics being tossed out and price being the only factor), which we know leads to, first and foremost, the deaths of certain networks like sci-fi channel who cannot compete with the big guys (ABC, etc), which will lead to nonstop episodes of the Bachelor, Survivor, and When Animals Attack (i.e. all the junk shows with no substance) for the instant gratification of the idiotic viewers that already plague the market. So be careful what you wish for, because in this case, in attempting to get more worthwhile shows on the air you are far more likely to end up killing the few ones that you already have. The absolute worse case scenario is that somehow networks cannot jack numbers and we come to find out that more geeks have Nielson boxes than the total population numbers indicate causing all sci-fi shows to be cancelled.
Or maybe I'm just dumb (However, I do know that I'm drunk)
Slashdot Mirror
A large fundamental problem with attempting to capture the effect of an increase in gun ownership to crime rates is the data available on guns. I read through the posts quickly, since it's finals week so I may have missed similar responses, but I saw a few mentions of Lott's More Guns, Less Crime, which I recall generating lots of fanfare when it came out a few years ago. I am not entirely sure how Lott got his measurements on actual gun ownership, since it is not readily available or accurate, and whether or not he attempted to separate handguns (involved in more crimes) from other types of firearms. An economist, Mark Duggan, has a working paper called "More Guns, More Crime" where he uses a proxy variable of subscriptions to Guns N Ammo magazine to represent increases in gun ownership, and since Guns N Ammo mainly showcases handguns, it seems like a good proxy. He spends some time showing results from FBI reports and General Health Statistics reports to show that increases in the magazine subscription have a direct relationship in increases in gun accidents that usually stem from increases in gun sales. I know it's a bold assumption to make, but if you can get ahold of the paper (sorry I only have a hard copy), you can see that he makes a strong argument that an increase in guns leads to an increase in crime. Unlike a lot of pay for results studies, and since he's an economist, he includes several potential limiting factors and insight into why this may not be 100 percent correct, such as when measuring increases in gun ownership, it is very hard to capture the exact timing of purchase as to did the crime occur and then the purchase for defensive reasons, or did the purchase precede the crime. The paper is definitely worth a read.
The problem here with electing to participate in the monitoring process is that you lose your random sample and ability to control for data manipulation. The basis of the Nielsen system, and statistics in general, is that randomness is essential in approximating the behavior of a population, and important (with this issue) to limit networks ability to manipulate data. While you raise the prospect of "everyone" electing to participate, it is highly unlikely to happen, especially among the skittish /. crowd who would be required to give up those rights to privacy they hold so dear to keep their shows on TV. If the industry based its selection of shows off of people who "choose" to participate, then perhaps we would have an endless supply of sci-fi from the geek crowd (probably not with aforementioned geek privacy issues), or even worse/better off depending on your view, the networks would be able to jack their ratings by "electing" to view their own shows, thus creating fake ratings for advertisers while erasing any ability to target demographics. This would lead to telecom style price wars on advertising (due to demographics being tossed out and price being the only factor), which we know leads to, first and foremost, the deaths of certain networks like sci-fi channel who cannot compete with the big guys (ABC, etc), which will lead to nonstop episodes of the Bachelor, Survivor, and When Animals Attack (i.e. all the junk shows with no substance) for the instant gratification of the idiotic viewers that already plague the market. So be careful what you wish for, because in this case, in attempting to get more worthwhile shows on the air you are far more likely to end up killing the few ones that you already have. The absolute worse case scenario is that somehow networks cannot jack numbers and we come to find out that more geeks have Nielson boxes than the total population numbers indicate causing all sci-fi shows to be cancelled.
Or maybe I'm just dumb (However, I do know that I'm drunk)