Pew’s recent survey on firearm-related attitudes and experiences of U.S. adults found – based on the percentage saying “yes” to the question about whether they are NRA members – that more than 14 million Americans consider themselves NRA members. The real NRA membership of about 5 million falls well short of this measure, even accounting for any statistical error the survey produced.
What this means in terms of polling, and one thing that Pew and others simply do not make clear to the public when reporting on the survey results, is that Pew did not actually survey NRA members. Any views, beliefs, or opinions they ascribe to “NRA members” is a simple guess on their part. Pew does not know what percentage of NRA members support one law or another, how many guns they own, or anything else for that matter. At best, they can claim to have the responses of Americans who SAY they are NRA members, but they certainly cannot say much beyond that.
Unfortunately, this simple truth hasn’t stopped Pew from attempting to present to the public, along with an eager media, multiple claims or measures about our members’ opinions. Nor has it stopped the similar and blatant attempts by ideologically-driven gun control advocates to claim they know, from polling data alone, how NRA members view particular topics.
July 7, 2017
Remarkable Finding from Pew Survey
[In addition to he conclusion outlined above this has rather far reaching implications and one or more of the following:
- Many people are lying about belonging to the NRA and want to skew the poll results
- Many people align themselves with the NRA but haven’t maintained their membership
- Pew sampling is faulty
- Pew is lying
- NRA is lying about their numbers by understating them by a factor of 2.8
It’s difficult to see how it would be to the NRA’s net advantage to lie in this manner so I am mostly discounting this option. It also seems unlikely that Pew would lie in this manner. Could Pew sampling be this far off? I suppose it’s possible but I doubt it.
That leaves options 1 and 2 as the most likely resolution to the discrepancy. I can only come up with weak cases for either option and end up being dissatisfied with the result. What am I missing?—Joe]