Multivariate procedures for evaluating factors afffecting palatability of beef

Multiple regression, factor, and canonical correlation analyses were used to study the relationships among palatability attributes, Qualatitative Descriptive flavor and aroma variables, live animal traits, and carcass variables. Data from 301 steers were analyzed to determine those factors which were most valuable in predicting palatability characteristics. Carcass weight was the most important animal characteristic found to predict juiciness, flavor, and overall desirability scores. Addition of more than one of the carcass characteristics to the regression models did not result in a large increase in the coefficients of determination and the results of the factor analysis showed that all carcass variables and slaughter weight were highly related. The number of days the cattle had received a high-energy diet was found to be the best single predictor of palatability characteristics. The strength of carcass variables in predicting change in palatability characteristics appeared to be related to their change with respect to the number of days the cattle had received a high-energy diet. Carcass characteristics, palatability attributes and quantitative descriptive analysis (QDA) factors collected from the carcasses of 108 steers were evaluated by canonical correlation analyses to evaluate the relationship among these groups of varibles. Evaluation of carcass and QDA variables resulted in one significant (P<.0001) pair of canonical variables which related carcass fatness positively to beef-fat and negatively to milky-oily descriptor variables. Analysis of carcass and palatability variables resulted in one pair of canonical variables describing primarily carcass weight and sensory flavor and overall desirability. Two significant (p<.05) canonical variables resulted from analysis of the QDA factors and palatability attributes. The first QDA canonical variable contrasted beef-fat and milky-oily flavor and aroma factors while the second QDA canonical variable was composed primarily of liver flavor. The first palatability canonical variable was highly related to the flavor and overall desirability palatability attributes while the second palatability canonical variable was related to tenderness and juiciness. Increases in palatability attributes and QDA factor were more affected by carcass fatness variables then they were by the other carcass characteristic.