The objective of this study was to predict overall consumer liking (OLike) or Warner-Bratzler shear force (WBSF) using either consumer tenderness (OTend) and juiciness (OJuice), trained descriptive tenderness (Tend) and juiciness (Juice) attributes, or WBSF.

Data where consumer sensory, trained meat descriptive tenderness and juiciness attributes, and Warner-Bratzler shear force were used. Study 1 used top loin steaks cooked to 58°C or 80°C utilizing a George Forman grill (191°C) or a flat food-service grill (232°C). A second study used Choice beef top loin steaks cooked to 58 or 80°C on a flat electric grill (176°C). Companion steaks were cooked for WBSF (kg). A third study used 60 top loin steaks cooked on an open hearth electric grill (176°C) and served to a trained meat descriptive attribute panel. For study 1, consumers (n = 80 per city) were recruited in Portland, OR; Olathe, KS; State College, PA and study 2 recruited 120 consumers per city from Olathe, KS; State College, PA; Portland, OR; and Griffin, GA.

Consumer ratings (n = 3228), Warner-Bratzler shear force values, and 400 trained descriptive attribute values (n = 400) were used to develop 13 equations (Table 1) . Equation [1] used Tend ratings to account for 24.6% of the variation in WBSF. Inclusion of Juice and their interaction did not appreciably increase the amount of variation accounted for by Eq. [3] (R² = 0.25) compared to Eq. [1]. Equations to predict OLike using Tend, Juice or WBSF had very low R². OTend and OJuice were better predictors of OLike (R² = 0.500, 0.492, 0.554 for Equations 10, 11, and 12, respectively) than when Juice, Tender or WBSF were used.

WBSF is more highly related to trained descriptive tenderness ratings then to consumer tenderness liking values. Overall consumer liking is difficult to predict using trained descriptive attribute and WBSF values and is most highly related to consumer sensory liking ratings for tenderness and juiciness. Juiciness ratings, either trained or consumer, did not appreciabilty improve predictability of regression equations to predict either WBSF or consumer overall liking.