Public Attitudes Toward Gun Control - Evaluating Public Opinion Polls

Public opinion polls, like all sources of information, must be used with care. Pollsters select sample populations because it is impossible to interview every American on a given question. The selection is usually performed randomly, a process made easier today by computers. Most major pollsters interview between one thousand and two thousand people to establish a valid sample. Other pollsters may interview far fewer than five hundred, and the sample could be too small to fairly represent the opinion of the American people. Generally, the larger the sample, the greater the chance for an adequate representation and a valid result.

The polling errors that concern most people are those caused by bias in the presentation of questions. This can be a factor in how a question is answered. Is the question vague? Is it too long? Is it threatening? Were the questions asked over the telephone or by mail? If asked in an in- person interview, was the interviewer too forceful or threatening? Did respondents provide answers they thought would please interviewers? Were respondents disqualified because of membership in gun control or gun rights organizations? What was the purpose of the poll? Who hired the polling organization, and what is its stand on the issue?

While most pollsters are not purposely or even consciously biased, they are subject to the same influences as everyone else. In the case of gun use surveys, it is entirely possible that respondents would be unwilling to candidly discuss their use of weapons. Finally, polling does not always determine how important a person considers an issue. The issue may be of absolutely no concern to the respondent, but when asked, the respondent then thinks about the topic and provides an answer. Five minutes after the question has been asked, the issue may completely disappear from his or her mind.

The polling organization might not include the number of times there was no response. If these "no replies" come predominantly from one group, it might influence the answers so that they do not truly represent national opinion on a given issue. Pollsters are aware of these weaknesses and, for that reason, usually indicate how reliable they consider their polls. Many polls indicate a plus or minus (+/−) two to four percent accuracy rate.