Zapka J G, Stoddard A M, Lubin H
Department of Health Policy and Management, University of Massachusetts, Amherst 01003.
Med Care. 1988 Jan;26(1):27-33. doi: 10.1097/00005650-198801000-00003.
Dental records and client survey reports were compared as sources of data for dental research. Further, two forms of the client survey were completed, a mail survey and telephone interview. A random sample of 271 clients of a dental group practice based within a university health service was selected. The Kappa values indicated a reasonable level of agreement of client survey and dental record with regard to the types of services received but no with regard to actual number of visits. Responses to the mail and the telephone versions of the client survey were compared for service satisfaction, hygiene behavior, and attitude and belief variables. Two potential sources of bias: nonrespondent/respondent bias and other response biases, such as acquiescent responses, were detected. The study illustrates the importance of weighing which source of response bias is more crucial to the general validity of a study based on the research questions to be addressed.
