EFFECTS OF NUMBER OF JUDGES UPON SCALE VALUES IN THE ANALYSIS OF SMALL GROUP TASKS,
Abstract
The data from a previous study of a method for scaling small group tasks were partially reanalyzed to determine the effects of number of judges upon scale and Q values. Five of the 10 dimensions used in the earlier study (Cooperation requirements, Decision verifiability, Difficulty, Goal clarity, and Solution multiplicity) were selected for reanalysis. These dimensions represented a range of interjudge agreement, as indicated by Q values. The first twenty and the last five judges recruited for the original study were used as new groups for the computation of scale and Q values. Correlations between scale values and between Q values were computed for varying number of judges. The results indicated that reliable scale values could be obtained from as few as five judges for the most reliable dimensions and as few as 20 judges for the less reliable dimensions. Stable Q values require a larger number of judges, probably 50 or more. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1963
- Accession Number
- AD0422985
Entities
People
- J. Michael Blum
- Marvin E. Shaw
- Morgan Worthy
Organizations
- University of Florida