Accession Number:

AD0416050

Title:

PAIRED-COMPARISON EXPERIMENTS INVOLVING FIVE RESPONSE CLASSES,

Descriptive Note:

Corporate Author:

RESEARCH TRIANGLE INST DURHAM N C

Personal Author(s):

Report Date:

1963-03-29

Pagination or Media Count:

21.0

Abstract:

A paired comparison is made by presenting two treatments to a judge who is asked to choose the one that he considers the better on the basis of some common characteristic. For example, the treatments may be two blends of coffee and the comparison consists of selecting the blend with the better flavor. A paired-comparison experi ment involves the pairwise comparisons of two or more treatments by one or more judges. It is proposed to extend the method of Glenn and David to cover paired-comparison experiments involving five response classes. This will be accomplished by making the assumptions that a tie will occur whenever the difference between the judges re sponses to two stimuli under comparison lies be low a first threshold, a mild preference will be declared whenever the difference lies between the first threshold and a second threshold, and a strong preference will be declared when the difference exceeds the second threshold. There will thus be five admissible response classes, viz. the two mild-preference classes, the two strong-preference classes, and the no-preferences or ties class. Author

Descriptors:

Subject Categories:

Distribution Statement:

APPROVED FOR PUBLIC RELEASE