I don't mean to insult anyone intelligence, but I'm guessing that most folks here (and in the world at large) don't have a significant background in statistics, sampling methodology etc. In short if you ask enough people the same question you reduce the risk of your study being skewed to a very small amount by people answering falsely. To be 99% sure that your answer is +/- 1% (I'm simplifying the math here) from a group of 300 million Americans, you'd need to survey approximately 16,600 people. If you wanted to be 95% sure that your answer is +/- 1% then you'd need to survey about 9,000 people. If you open it up to say 95% at +/- 3% you need just a hair over a thousand. So it's not difficult to do a statistically valid survey that accounts for self report variations.
-Jenrick
Unfortunately there are some states that take ALL of the guesswork out of it. Illinois is one, with it's FOID card. New Jersey is another.
In Illinois, we have widely disparate gun owner %'s across the state, county by county.
For example, as of 2014 (the last FOIA documentation I have from the State Police);
In Cook county (Chicago), there are 5,240,700 people with 424,784 gun owners... 8.1%.
In my county (Tazewell), there are 136,352 people with 31,931 gun owners, which is 23.4%.
Calhoun county, population 5,059, has 2,360 gun owners; or 46.6% - the highest percentage in the state, and, not surprisingly, the 3rd least populated one.
The overall state percentage is 12,882,135 people, with 1,792,220 gun owners - or 13.9% of the population.
The 10 LEAST populated counties (the most densely populating having a mere 6,860 souls) combined total 55,337 people, with 18,318 gun owners, or 33.1% of the population.
Meanwhile the 10 *most* populated counties, having a total population of 9,419,469 (73% of the population of the state), has 991,241 gun owners; or 10.5% of the population.
I would wager that across the country you would find that urban centers track similar to this; the bulk of the population wrapped up in high density housing, with a very low percentage of gun owners.
Meanwhile, sparsely populated counties would have a much higher percentage of gun owners.
If you do a survey of random numbers distributed geographically over the country, you'd most assuredly get skewed results which are much higher than true results.
Meanwhile, if you weight-balance your calls so that higher population centers get a higher percentage of phone calls (e.g. 1035 people polled in Chicago for the ONE lonely soul you call in Calhoun county), you'd get better statistical results.
However the uneven geographical distribution of gun owners would make even the most well thought out survey useless.
The only numbers we can trust, are those from states which actually track gun owners as part of a systematic registration process, such as Illinois or New Jersey. From those we can make educated assumptions of the distribution of gun ownership on urban vs. rural.
A randomized calling system of numbers will not yield accurate results, by any means. Not unless the sample size is so sufficiently vast that the majority of America is called. This is simply due to the vastly uneven distribution of gun owners in the country. If you had full data of how they were distributed you could take an accurate survey; but without taking a full accounting you couldn't realize that distribution, which is a Catch-22 preventing any survey from being anywhere close to accurate.