Beef knuckles (n = 60) were chosen from a population of 328 knuckles to test a selection procedure and determine the potential causes of liver-like off-flavor. In phase I, 2 independent panelists were allowed to smell the aromas or smell and taste samples of cooked beef to determine the presence or absence of the liver-like off-flavor and off-flavor intensity. The panelists tested knuckles from 5 feedlots, but only identified 29 as having an off-flavor. A kappa statistic was generated to assess the level of agreement of the 2 panelists that indicated the panelists moderately or substantially agreed when judgments were based on smell exclusively or smelling and tasting, respectively. Although the agreement was acceptable (kappa = 0.57 and 0.76), there was not enough variation in the liver-like off-flavor and off-flavor intensity for the 2 panelists to detect differences compared with an independent sensory panel. Phase II identified factors that led to the development of the liver-like off-flavor in beef. The M. rectus femoris from knuckles identified from phase I were used. Sensory analysis, proximate composition, heme iron, mineral content, and fatty acid analyses were conducted. Stepwise regression was used to identify factors contributing to the liver-like off-flavor. Specifically, Na, 16:1, cis 18:1(n-7), 20:2(n-6), and 20:3(n-6) fatty acids explained (P = 0.021) 46% of the variation of the liver-like off-flavor. Although previously reported as playing a role in the development of the liver-like off-flavor, iron, heme iron, and pH had no effect in this study.