|
Abstract:
|
The National Crime Victimization Survey (NCVS) is a survey of the non-institutionalized U.S. population aged 12 and older and utilizes a 7-wave rotating panel design. In a survey such as the NCVS which is used to measure crime retrospectively, the panel design is useful in that it provides bounded reference periods for all but the first interview. This can improve survey accuracy by reducing recall bias among respondents, though it also introduces another source of potential bias not present in cross-sectional surveys: respondent fatigue. Since participants remain in the sample for a period of up to three years, and since each interview can be time-consuming, the potential for wave nonresponse, panel attrition, and underreporting is high. Each can have a negative impact on the quality of estimates. This paper addresses how the authors assessed the impacts of these error sources, with emphasis on the methods used to measure fatigue-induced rate deflation. After accounting for potentially confounding sources of bias, the impact of fatigue on estimates of personal and property crime is shown to be substantial, suggesting that respondent fatigue is a source of bias in the NCVS.
|
