Persons who conduct in-person interviews are much more effective if they share important characteristics of the respondents they are interviewing. For example, black interviewers get better cooperation in black neighborhoods than white interviewers, and vice versa. Failure to match interviewers and respondents can produce high refusal rates and responses that are less than candid.
Regardless of how the data are gathered, in person or by telephone, the polling organization must decide how to deal with people who refuse to be interviewed, people who refuse to continue the interview and terminate it before finishing, and people who cannot be contacted. In the case of a failure to make contact, most polling organizations make repeated efforts to reach the person targeted for the interview to avoid bias that might result, for example, from substituting a homemaker who is at home for woman who has a job. Keeping that in mind, a large volume of polling is done in the evening, when all members of the household are most likely to be home. When additional individuals are selected to be respondents to replace others who could not be used, it is important that the same selection procedure be used for the substitutes as for the original list.
With the rapid growth of polling organizations, Americans increasingly find themselves inundated with surveys. An understandable response to one too many phone calls is to refuse to participate. Since there is no way to know exactly who is refusing to respond, it is difficult to know how to correct for any potential bias in the sample interviewed. The increasing refusal rate presents a challenge for all polling organizations.