Wash Post falls for bogus numbers about gun ownership

[Jack Burton]

Really... who is going to tell the truth to a total stranger who just happens to call on the phone and ask if you own a gun. Especially in states such as Illinois and California.

Gun Ownership by State (washingtonpost.com)

 

[Arickosmo]

Well, during this political cycle, I've taken to answering the phone ALOT. I'm having fun with them so if they had called me, I would have tipped the scales. The NRA or any gun favorable entity isn't going to call and ask so I would assume they hate guns and let em have it.

 

[Smokepole]

Not inclined to believe those numbers. Like Texas has 36% owning and 64% not while Wisconsin is 44% and 56%. Not buying it.

 

[Lancem]

Well that was back in 2001, and they polled like what 1% or less of a states population?? How is that going to get you any kind of accurate number?

 

[JOLLYKART]

They have convinced them selfs that the same cross section of America they have been using sense the 40's still represents what America thinks. Shame. I think everyone would agree, if you want information, ask for it on a voters ballet, otherwise it just becomes worthless information not suitable to wipe your rear with.

 

[Mr Evilwrench]

Statisticians have some magical formulae they've developed that in most cases seem to represent reality with pretty small samples, but their logic is sometimes faulty, and they fail to account for motivations of the surveyed. This is an excellent example. Asking questions of people, many of whom have been given good reason for skepticism, and expecting honest answers, is a recipe for failure. If I thought there was any prospect they really wanted to learn, rather than support an ideology, I would be more likely to be charitable, and consider them scientists, but I'm not buying it.

 

[CarmelHP]

Response bias, as mentioned above, in this type of poll is a problem. The sample sizes cited are not a problem as long as all in the target population were randomly sampled and all of the population was equally likely to be sampled. Law of Large Numbers calculations point to statistical power topping out at about 3000 sample size with very little gain in accuracy above that regardless of population size. Response bias, though, is highly motivated is this type of poll. Unless I know you, my response is a standard "No" in response to gun ownership questions. I don't want anyone's crooked teenager or nephew of step-brother hearing any loose talk and come looking for an easy score when they think no one is home. I would think the same kind of worry applies to an anonymous caller.