INVESTIGATION OF LINEAR TYPE EVALUATION FOR COLLEGIATE JUDGING CONTESTS

Various time constraints and cow numbers were used to identify the min imum time per cow and the minimum number of cows needed to permit rank ing of students for their skills in linear type evaluation. Three sepa rate trials were conducted involving 23, 15, and 12 students from thre e universities who evaluated 15 traits on 30, 20, and 27 cows, respect ively, using a 50-point scale. Time constraints included 1.5, 2, 3, an d 4 min per cow as well as unlimited time per cow. The mean of the abs olute difference between official and student scores was used to rank students. Lower means indicated greater student accuracy. Mean differe nces and (standard errors) across trials were 7.28 (0.24), 7.41 (0.11) , 7.18 (0.11), and 7.41 (0.18) for 1.5, 2, 3, and 4 min of evaluation, respectively. The overall effect of time was not significant, and no pairs were significantly different. Rank correlations between student means suggested that 2 min of evaluation were adequate, and no advanta ge was found for evaluations of >3 min per cow. Rank correlations for student evaluation means based on 5 versus 10 cows were 0.85 to 0.90. Evaluation of <4 cows dropped rank correlations, but evaluation of >5 cows did not improve correlations greatly. Results suggested that use of 4 or 5 cows, evaluated at 2 to 3 min per cow, permitted accurate ra nking of students for their ability to evaluate linear type traits.