Differences among Appraisers in the New York Type Appraisal Program.

The purpose of the New York type appraisal program is to a t tempt to rate type traits objectively. Approximate ly 34 traits are included, of which 25 are rated by the appra i se r and 9 by the herd mml~ger. This pape r reports differenees among 18 appraisers, including S professional judges, 7 New York Art i ficial Breeders ' Cooperative fieldmen, and 3 other sire selection personnel all of whom rated 38 cows in a single herd. There were statistically significant differences (P ~ .05) among appraisers for all traits. Average scores of the professionals were different from those of the fieldmen for all except six traits. Differences among the professionals were also large, as were the differences among the fieldmen. The results suggest that the fieldmen as a group appraised with as much consistency as the group of professional judges. The Extension division of the Animal Husbandry Depar tment of the New York State College of Agricul ture at Cornell Univers i ty in 1953 began a program of objective ra t ing of type traits, in an at tempt to uncover characteristics early in life which would influence herd life. This repor t considers the differences among appraisers, professional judges, fieldmen of the New York Artificial Breeders ' Cooperative, their sire analysts, and the chairman of their Holstein sire selection committee when faced with the same decision s i tua t ion-the same eows in one herd. EXPERIMENTAL PROCEDURE All appraisers were to rate all cows in a Holstein herd for all traits. The list of traits and the way they were broken apar t for analysis are shown in Table 1. Where the rat ings for a t rai t were in a l inear pat tern, the scores for a t ra i t were analyzed as a single variable. I n other eases, when the possible scores were not linear, each possible score was treated as a different binomial variahle with scores zero or one. The eight professional judges were accredited in New York. There were seven N Y A B C fieldmen, two N Y A B C sire analysts, and the dairyman chairman of the Holstein sire selection committee. F i f t y cows from the Griswold herd of Cortland, New York, were ineluded in the experiment. Only 38 of these cows were rated by all 18 appraisers. The analysis of variance was for a factorial a r rangement -18 appraisers and 38 cows--wi th one observation on each variable per subclass. l~eeelved for publication June 23, 1964. 61 The sum of squares for 17 degrees of f reedom associated with appraisers was part i t ioned into single degree of freedom contrasts. The analyses are summarized in Table 2. The columns of the analysis of variance are described below. Degrees of Source of variation freedom Among appraisers ] 7 Professionals versus fieldmen 1 Among professionals 7 A m o n g f i e ldmen 6 Professionals versus other (analysts and dairyman) 1 Amm~g others 2 Among cows 37 Error 629 All significance tests were at the (P ~ .05) level. RESULTS AND DISCUSSION The general result, as can be seen f rom Table 2, is that there were statistically significant differences among appraisers for all traits except for a few of the (0 or 1) variables within a trait. The scores of the professionals were significantly different f rom those of the fieldmen for 19 of the 25 traits. The six traits not scored significantly different by the two groups were shoulder (8), hind legs sideview (16), hind l egs r ea r view (17), pasterns (18), teat pos i t ion rea r (53-55), and teat pos i t ionfo re (56-58). There were several significant differences within the professional and fieldmen groups for four of these t ra i ts ; shoulder scores, hind legs--s ide view, hind l egs r ea r view, and pasterns. F o r only two traits where there were sig62 L. D. VAN VLECK AND R. ALBRECTSEN