Background

It is apparent that multiple factors impact flavor in beef. Researchers have been extensively involved in first defining what flavors are present in beef whole muscle cuts by developing the beef whole muscle flavor lexicon (Adhikari et al., 2011). Then the team examined the effect of cooking method, degree of doneness, and cut on beef flavor attributes and consumer liking. They have determined perceptions of millennials, non‐millennials, heavy beef‐eaters, and light beef‐eaters through consumer studies including Central Location Tests and In‐Home Use Tests to assess consumer liking. While each of these data sets (4 total) have addressed the objectives and hypotheses presented, combining the data sets to model beef flavor has not been completed. The objective was to develop a beef flavor model to understand the major positive and negative flavor for beef of whole muscle cuts and ground beef using trained descriptive flavor and texture attributes, consumer liking attributes, volatile aromatic compounds, and chemical data.

The objectives of this study were to 1) use existing data to develop a model to understand the relationship between consumer overall liking and trained beef descriptive flavor attributes, volatile aromatic compounds, meat chemical attributes, and precursor compounds in whole muscle beef and 2) understand if factors that drive flavor acceptability in whole muscle beef products are similar for ground beef using existing data.

Methodology

Five previous studies were utilized to understand relationships and ability to predict overall consumer liking. The first study (Luckemeyer, 2015) used beef cuts from 10 beef carcasses differing in quality grade, cooking method, and internal cook temperature endpoints. Cuts included: Choice top loins, High pH top loins (pH ≥ 6.0), Select top sirloin butts, Choice tenderloins, Select bottom rounds, and Choice bottom rounds. Top loin steaks were cooked to 58°C or 80°C utilizing a George Forman grill (Choice top loin, High pH top loin, Select top sirloin and Choice tenderloin steaks) set at 191°C, a flat food service grill at 232°C (Choice top loin, High pH top loin, Select top sirloin and Choice tenderloin steaks), or crock‐pot (Select bottom round and Choice bottom round roasts).  

The second study (Laird 2015) used USDA Choice beef top strip loin (cooked to 58 or 80°C on a commercial electric flat grill) and Select bottom round flat roasts (cooked to 58 or 80°C in a crock pot). Millennial and non‐millennial consumers that were either light (eat beef 2 to 4 times per month) or heavy (eat beef 3 or more times per week) beef eaters were used in this study.

The third study (Glascock, 2014) used the same treatments as in study 1 except tenderloins were not included and consumers were defined as light beef eaters in study 1 and heavy beef eaters in study 3.

The fourth study (Berto, 2015) used beef top loin steaks cut to two thicknesses and grilled on a flat top grill at one of two temperatures.

In a fifth study, ground beef was prepared using two fat levels (10 and 20%), two processing methods (traditionally grind and bowl chopping), and four meat sources (commodity trim, round, chuck and sirloin beef cuts). The subsequent ground or chopped beef was formed into patties, cooked on a flat electric grill to a medium degree of doneness and evaluated using consumer and expert trained descriptive sensory panelists.

In each study, trained descriptive flavor and texture attributes were determined using the expert panel that helped develop the Beef Lexicon with over 25 years of experience. Consumers were selected in Portland OR, Kansas (either Manhattan or Kansas City), Houston TX, Griffin GA, or Pennsylvania (Philadelphia or College Park). Approximately 80 consumers per city were randomly recruited based on selection criteria for each study. Consumers evaluated beef in a Central Location test where samples were prepared as defined for descriptive attribute evaluation. Samples from the consumer evaluation were used to determine volatile aromatic compounds associated with beef flavor from steaks prepared at one consumer location per study. Raw chemical attributes were determined on beef prior to cooking where myoglobin, nonheme iron, and percentage of lipid and moisture were determined. Fatty acid content was also determined on the raw samples. These data were used to first understand factors impacting beef flavor and then prediction equations using multivariate analyses were calculated.

Findings

Nine Partial Least Squares prediction equations were presented. Of these equations, the five equations to predict overall consumer liking for top loin steaks, beef cuts and ground beef are presented in the Summary Table. These equations had moderate strength of prediction and can be used to understand factors that impact consumer liking without conducting extensive consumer testing across the United States.

Implications

Beef flavor is a complicated sensory attribute that is impacted by multiple factors such as beef cut, cooking method, degree of doneness, and USDA beef Quality grade. Research has documented that beef flavor is one of three driving factors of consumer overall liking and actually contributes slightly more to consumer overall liking than tenderness and juiciness. With the completion of 5 extensive consumer studies, prediction of consumer overall liking was evaluated. The purpose of the study was to develop equations where beef can be evaluated by an expert, by a trained descriptive attribute panel and by differences in treatments and whether it can be determined if the differences are enough to impact consumer liking. Additionally, the potential to measure volatile aromatic compounds during cooking in real‐time or on raw beef would provide methods of detecting beef that had negative flavor attributes. Nine final prediction equations can be used that use either expert descriptive flavor and texture sensory attributes, volatile aromatic compounds, or raw chemical values from beef.