This study was designed to determine the extent to which the results of an employment survey of graduates of a teacher preparation program would have been affected by changes in response rate. At the University of Tennessee, Knoxville, a followup of teacher education program graduates is conducted annually. A total of 284 graduates of the 1992 teacher education program were identified as the target population for the 1993 survey. A total of 184 individuals responded to the mail survey (64.8%), with 33.8% responding to the first wave, 17.2% responding to the second, 6.3% to the third wave, and 7.4% responding late. Telephone calls elicited information from 40 additional persons (14.1%). Supplemental information about employment from other sources resulted in the eventual determination of the occupations of 265 of the 284 individuals (93.3%). There was no evidence that data collected after about 50% of the sample had responded resulted in any meaningful differences in survey results. These results suggest that concentrating on potential nonresponse bias may not be as important as attending to other aspects of survey methodology, such as sample size and questionnaire design. (Contains 2 tables and 28 references.) (SLD)