United States Department of Transportation
i
A Methodology for Analyzing General Categorical Data with Misclassification Errors with an Application in Studying Seat Belt Effectiveness
-
1977-06-01
[PDF-2.95 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
DOI:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final report
-
Corporate Publisher:
-
Abstract:Author's abstract: In this report, a methodology for analyzing general categorical data with misclassification errors is developed and applied to the study of seat belt effectiveness. The methodology assumes the availability of an original large sample based on a fallible classifier, and requires obtaining a small supplementary sample that is cross-classified by both the fallible and a "true" (usually more expensive) classifying device. For the study of belt effectiveness, the original sample is drawn from North Carolina accident reports for the first eight months of 1975. The "true" classification of a sub-sample of these occupants is assumed to be obtained through hospital reports for the injured occupants and telephone interviews for the non-injured. Chapter I presents some background material, while Chapter II outlines the mathematical derivation of the proposed methodology. (Details of the statistical
methodology are given in Appendix A.) In Chapter III, the data to which the methodology is applied in examining safety belt effectiveness is described. Results of the investigation are given in Chapter IV. An examination of the nature and magnitude of misclassification errors in this police-reported belt and injury data is described in Chapter V. A brief discussion-of this work along with suggestions for further research are included in Chapter VI.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
