Validation of a New Analysis Program for Law of Comparative Judgement Data
Abstract
The psychophysical Law of Comparative Judgment (LCJ) has been shown to have substantial value in assessing the signature of camouflaged or signature managed (C/SM) military assets when there is a need for a psychophysical measure of effectiveness. However, statistical tools for analyzing LCJ data are not widely available, have until now only had command- line interfaces, and require extensive interaction with the analyst. This prompted the Army to produce a new computer program that combines the functionality of several older command-line programs and has a more convenient windows interface. Exercising normal caution with new computer code, it is necessary to verify the new program to ensure that it correctly performs the desired tasks. This process requires a number of test cases: examples of each different class of results to be expected. Additionally, some of the changes in the program are expected to affect the precision of values for LCJ scales. Thus, since the values obtained with the new program are expected to be different than those obtained with the old one, it is necessary to verify that the results are equivalent. This paper describes the nature of the changes, the validation process, the nature of the validation data, and the results of the validation effort. The conclusion states that the new program does give results that are very similar to those obtained with the old programs.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2002
- Accession Number
- ADP023107
Entities
People
- James J. Crile
- James R. Mcmanamey
Organizations
- Tank-automotive and Armaments Command