Citation

Arthur W, Tubre T, Day EA, Sheehan MK, Sanchez-Ku ML, Paul D, Paulus L, Archuleta K. Hum. Factors 2001; 43(1): 1-11.

Affiliation

M University, College Station 77843-4235, USA. wea@psyc.tamu.edu

Copyright

(Copyright © 2001, Human Factors and Ergonomics Society, Publisher SAGE Publishing)

DOI

unavailable

PMID

11474755

Abstract

In the crash involvement literature, it is generally assumed that archival and other "objective" criterion data are superior to self-reports of crash involvement. Using 394 participants (mean age = 36.23 years), the present study assessed the convergence of archival and self-report measures of motor vehicle crash involvement and moving violations. We also sought to determine whether predictor/criterion relationships would vary as a function of criterion type (i.e., archival vs. self-report), and if a combination of both criteria would result in better prediction than would either by itself. The degree of agreement between the two criterion sources was low, with participants self-reporting more crashes and tickets than were found in their state records. Different predictor/criterion relationships were also found for the two criterion types; stronger effects were obtained for self-report data. Combining the two criteria did not result in relationships stronger than those obtained for self-reports alone. Our findings suggest that self-report data are not inherently inferior to archival data and, furthermore, that the two sources of data cannot be used interchangeably. Actual or potential applications include choosing the appropriate criterion to use, which, as the finding of this study reveals, may depend on the purpose of the investigation.