47% of households own gun

[spidey07]

The only thing to swing the poll would be a good number of people would say they didn't own firearms even if they did which would skew it in the other direction of MORE household, not less.

I would consider 41% to 47% or about a 15% increase in a single year pretty damn significant. The article states sales were MUCH higher, by a THIRD over the huge panic buying that happened in november 2008. If you weren't involved in firearms back then, it was a mad house of manufacturers not being able to keep up with demand, ammo shortages, prices went through the roof because of supply/demand. It was nuts.

And sales were 1/3rd higher than that. People are stocking up on firearms and tons of first time buyers realizing they just may need to protect what they have.

 

[PrinceofWands]

I have taken some statistics classes smartass and the accuracy of this poll for this particular question is flaky at best. It was open polled as well. Here is how the poll was done

All your sampling nonsense is only accurate when its a control enviroment. That is not the case here. They randomly contacted adults in each state. The oddness is they contacted only 600 landlines. So you talking about literally 12 people per state in that way. Then they contacted 400 people via cell phone. For a poll like this that is quite odd. There are a lot of people who have cell phone numbers from one state but live in another.

Maybe this methodlogy is effective for elections, but to try and achieve a some what accurate answer as to gun ownership per home, you need a greater sample of people. To try and determine an accurate assessment per state for a question such as this, you clearly need more than 20 people per state.

I think you missed understanding the critical two parts:

For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±4 percentage points.

Samples are weighted by gender, age, race, Hispanic ethnicity, education, region, adults in the household, and phone status (cell phone only/landline only/both, cell phone mostly, and having an unlisted landline number). Demographic weighting targets are based on the March 2010 Current Population Survey figures for the aged 18 and older non-institutionalized population living in U.S. telephone households. All reported margins of sampling error include the computed design effects for weighting and sample design.

In other words, they matched the demographic breakdown of the US to within a small margin in that number of respondents, which is a statistical given once you reach a large enough sample size using their methods. It happens at just over 1000 people, which is why nearly all professional polling uses similar numbers.

Please provide for me ANY peer-reviewed arguments against sampling accuracy. Please show me where any significant polling of any type uses samples larger than this.

 

[classy]

Yes. Do you understand what that is telling you? The part that you bolded doesn't mean what you seem to think it means. No part of this poll whatsoever is using a sample of 20 people. None.

Where did you get the idea that the sampling is done on a state by state basis? The explanation of their methodology clearly states they are using random digit dialing methods. They took a random sample of 1,000 Americans. If the sample is truly randomly selected (and I see no reason to believe otherwise), then the central limit theorem clearly states that sample is perfectly fine within this margin of error.

This is basic statistical math. There is no increase in sample size necessary. I think you do not understand what the poll is measuring.

I think you need to read rather than babbling. It was 1005 adults from all 50 states and DC. They then state it was weighted and part of the weight was region. So they roughly took 20 people per state. That's simple math.

 

[classy]

Please provide for me ANY peer-reviewed arguments against sampling accuracy. Please show me where any significant polling of any type uses samples larger than this.

Huh? I linked an article from the Washington Post, all be it older though. That poll was done over 200,000 people. And it showed a 10% lower ownership than the poll that had only 1000 adults.

While I am sure there is break number that would lead to more inaccuracy. I clearly question how sampling only 1000 people over an entire country could even remotely possibly give me a reliable result for a random open question such as this.

And let me be clear, I am not questioning gallups reputation. But for this particular poll, I am questioning the results because the sample size to me appears to be way too low to represent the entire country.

 

[Jhhnn]

It's not 'supported', it's outright declared in the emails. Black and white, straight english, it says they're going to use it to get control measures passed. Period. No interpretation, straight text. You need to read them. They openly declare it.

Well then, you need to link them to support your contention.

 

[fskimospy]

I think you need to read rather than babbling. It was 1005 adults from all 50 states and DC. They then state it was weighted and part of the weight was region. So they roughly took 20 people per state. That's simple math.

Right, it is simple math. If you understand simple math you will see why 1005 adults is a perfectly representative sample. They really didn't even need that many, they could have easily gotten away with half that.

 

[classy]

Right, it is simple math. If you understand simple math you will see why 1005 adults is a perfectly representative sample. They really didn't even need that many, they could have easily gotten away with half that.

There are over 100,000,000 homes. There is no way 1005 people can be an accurate depiction of the state of that many houses as it pertains to this question.

 

[fskimospy]

Huh? I linked an article from the Washington Post, all be it older though. That poll was done over 200,000 people. And it showed a 10% lower ownership than the poll that had only 1000 adults.

While I am sure there is break number that would lead to more inaccuracy. I clearly question how sampling only 1000 people over an entire country could even remotely possibly give me a reliable result for a random open question such as this.

And let me be clear, I am not questioning gallups reputation. But for this particular poll, I am questioning the results because the sample size to me appears to be way too low to represent the entire country.

In order to better understand sample sizes and how 1000 is in fact far more than they need, read about the central limit theorem: http://en.wikipedia.org/wiki/Central_limit_theorem This is one of the foundational principles of all statistics along with the law of large numbers, specifically how they relate to sample sizes and the margin of error.

They show quite clearly and irrefutably that a random sampling of 1000 adults is sufficient to get a reasonable result even in a population of 300 million. If you would like to say that the sample was not in fact random, that would be a good way to undermine the validity of gallup's study. Saying the number is not large enough simply won't do though.

 

[fskimospy]

There are over 100,000,000 homes. There is no way 1005 people can be an accurate depiction of the state of that many houses as it pertains to this question.

Actually they can, and mathematically have been proven to be within the stated margin of error. Read my other post for details.

 

[monovillage]

There are over 100,000,000 homes. There is no way 1005 people can be an accurate depiction of the state of that many houses as it pertains to this question.

Statistically they can......... as long as it's a fair sample of people polled, what the questions are and how the questions are worded. It's possible to skew a poll, but if the poll givers and the people that paid for the poll don't have a dog in the fight, it's going to be fairly accurate. My opinion given earlier is that due to the question the false responses would change the outcome.

 

[McWatt]

I would consider 41% to 47% or about a 15% increase in a single year pretty damn significant.

http://en.wikipedia.org/wiki/Statistical_hypothesis_testing

 

[TheSlamma]

Is it the impending class war/riots

Is this thought you have caused by paint chips, lead in the solder, power lines or mercury?

Oh and let me guess.. you think you will be fighting as one of the 1%'rs? insert the ROTFLMFAO emoticon

 

[Nebor]

Is this thought you have caused by paint chips, lead in the solder, power lines or mercury?

Oh and let me guess.. you think you will be fighting as one of the 1%'rs? insert the ROTFLMFAO emoticon

You don't have to fight as one in order to fight for them. :thumbsup: