Skip Header

Deborah H. Griffin, Sandra L. Clark
Component ID: #ti379941825

Survey managers acknowledge that paper questionnaires must be designed to accommodate the requirements of data capture systems. It is important, however, to recognize that such design elements can have an effect on respondent behavior. This paper reminds us that even fairly minor changes in the format of a questionnaire can jeopardize successful completion by respondents. In 2005, the paper questionnaire used in the American Community Survey was redesigned to transition from a key-from-paper to an imaging and key-from-image (KFI) data capture methodology. On the surface the proposed changes appeared fairly minor--adding additional white space and removing horizontal lines between questions. The KFI system experts proposed most of the changes and little attention was given to possible respondent implications. The Census Bureau conducted a pretest of this revised questionnaire and analysis of the results identified an increase in nonresponse for selected questions. A closer review of the changes identified the potential for respondent navigation errors. A questionnaire design team was established to redesign this questionnaire to meet technical KFI requirements and address respondent visual miscues apparent in the new questionnaire. This paper reports on the test of the initial and redesigned forms. It provides a valuable lesson in the need to pay attention to both the technical requirements and the visual dynamics of respondent-administered survey instruments.

Back to Header