Telephone surveys are a valuable method of obtaining data. Sometimes they are the only means by which data on a particular topic can be gathered. But there are times when they can give you very bad data.
Suppose you did a phone survey asking if there were any fugitives from justice living at your house. My guess is the data obtained would differ considerably from the actual value.
Suppose you did a phone survey asking if you had expensive jewelry. Probably to a lesser extent but still my guess is the answers given will not be accurate.
Furthermore my hypothesis is the differences from reality to the two questions above would be be fairly constant over time. If you asked about fugitives living at your home my guess is that you might get 1 out of 100 to admit a fugitive was living there when there was one and that would not change much from 25 years ago to today.
Now suppose you did a survey on how many homosexuals live in your home. If you did the survey in the 1950s my guess is the report of a “none” when in fact there was one or more would be very high. But if the survey were done today the false reporting would be much lower.
Surveys on gun ownership suffer from similar problems. There is a threat of oppressive regulation and confiscation in some jurisdictions and there is the threat of theft by non government actors as well. Gun owners have a vested interest in not giving accurate data to some random person calling up on the phone.
Hence, I am skeptical of the latest survey on gun ownership. The huge increase in gun sales during late 2008 and 2009 was do to fear of heavy restriction. This climate of fear would have been motivation to falsely report to survey takers.
Note that this is not a VPC survey but just a VPC interpretation of a survey. Had it been a VPC survey additional motivation for error introduction could have been present as well. The VPC conclusions are almost for certainly flawed. If they were correct then Brady Campaign and VPC “membership” would be up and the NRA membership would be down. But this is not the case. I don’t think we can determine what the true gun ownership numbers are but I’m nearly certain it is much higher than that reported in the survey.