Nonresponse error is a systematic bias that occurs when the final sample differs from the planned sample. Nonresponse error occurs when a sufficient number of the preselected prospective respondents in the sample refuse to participate or cannot be reached. Nonresponse is caused by many factors. Some people do not trust the research sponsor or have little commitment about responding, while others resent what is perceived as an invasion of their privacy. The differences between people who do respond and those who do not can be striking.
For example, some research has shown that for mail surveys, respondents tend to be more educated than nonrespondents and have higher scores on related variables such as income. In addition, women are more likely than men to respond. Methods for improving response rates include multiple callbacks (or online contacts), follow-up mailings, incentives, enhancing the credibility of the research sponsor, indicating the length of time required to complete online or other types of questionnaires, and shorter questionnaires.