Beef flavor has been defined as an important component of beef demand. Beef flavor is not a “single” attribute, but has multiple attributes associated with it. The beef industry took the first big step in addressing beef flavor by funding the development of the beef flavor lexicon that identified major and minor beef flavor components. Now that the beef lexicon has been developed, researchers are working to understand which compounds are responsible for each flavor attribute in the lexicon and what flavor attributes drive consumer liking of beef.
In the past two years, research studies have used light and moderate-to-heavy beef eaters to evaluate beef flavor. Differences in beef flavor were presented to consumers, induced by using different cuts cooked using different cooking methods and cooking to medium rare and well done degrees of doneness. This model effectively created flavor differences in beef. One-on-one exit interviews with ~20% of the consumer participants were conducted to understand factors affecting consumer liking. From this research investigators have shown that consumers, whether light or moderate-to-heavy beef consumers, predominantly like beef cooked to a medium rare degree of doneness on a high temperature grill.
The previous data sets were not designed to address how demographics affect beef flavor liking. Research from the marketing group has shown that millennials, individuals ages 18 to 34 years, do not consume beef at the same proportion as non-millennials. Non-millennials are people older than 34 years and include the consumer classifications of Generation X and Baby Boomers. The question is, “Why do millennials not eat as much beef; why do they eat more chicken; and what flavor attributes drive their liking of meat and non-meat proteins?” The millennial generation has strong purchasing power. They are the beef industry’s next powerful consumer group, but to date, they do not cook, they eat more chicken than other protein sources and they are very connected to digital media. It is imperative that the beef industry understands millennial's perceptions on beef and beef flavor. Previous research included millennials, but they were either light or moderate-to-heavy beef eaters. This research sought to address millennials as a consumer group and examine what factors drive consumer perception for beef and how to increase their willingness to eat more beef.
The objectives of this study were to select four consumer groups, millennials and non-millennials that are either light (eat beef 2 to 4 times per month) or heavy (eat beef 3 or more times per week) beef eaters in four cities (Portland, OR; Olathe, KS; University Park, PA; Atlanta, GA) and determine their perceptions of beef liking based on being presented with beef, chicken and pork that has been cooked differently to create differences in flavor.
USDA Choice beef Top Loin Steaks and Select Bottom Round Flat Roasts; chicken Breasts and Thigh Meat; and boneless pork Loin Chops and Inside Ham Roasts were purchased commercially. Within each cut, cuts were cooked to one of two internal cooked endpoint temperatures, 137°F for beef and 145°F for chicken and pork (medium rare) or 176°F (well done). Top Loin Steaks, chicken Breasts and pork Chops were cooked on a commercial electric flat grill set at 350°F. Beef Bottom Round Roasts, chicken Thigh Meat and Inside Ham Roasts were cooked in a crockpot using the high setting. These cooking methods have been shown to induce differences in Maillard reaction products and heat-induced lipid oxidation. Internal temperatures were monitored by iron-constantan thermocouples inserted into the geometric center of the cut.
The intent was to create a set of steaks, chops and/or roasts that differed in key flavor attributes. These steaks, chops and roasts were evaluated by an expert trained meat descriptive attribute panel that helped develop and validate the beef lexicon, developed the pork lexicon and has used the chicken lexicon extensively. Flavor, juiciness and tenderness attributes were measured using a 16-point scale within each lexicon (0=none and 15=extremely intense).
Volatile compounds were captured from the same steaks evaluated by the panelists. Volatiles were evaluated using the Aroma Trax gas chromatograph/mass spectrophotometer system with dual sniff ports for characterization of aromatics. This technology provided the opportunity to separate individual volatile compounds, identify their chemical structure and characterize the aroma/flavor associated with the compound at parts per trillion. In addition, raw meat pH, fatty acid composition (neutral and polar lipids), myoglobin content, and non-heme iron content was determined for each sample.
After trained descriptive attribute sensory evaluation and the Aroma Trax chemical evaluation, the same samples were used for consumer evaluation. Consumers (n=120 per city) were randomly selected in four cities so that geographical areas represented the Midwest, the East coast, the South and the West coast. In each city, up to 6 consumer sessions with approximately 20 consumers per session were conducted. Within each city, consumers will be selected to be either millennials (ages 18 to 34; n=60) or non-millennials (n=60; ages greater than 34) and within age categories to be either light (n=30 per age group; eat beef 2 to 4 times per month) or heavy beef eaters (n=30 per age group; eat beef 3 or more times per week). Consumer demographics for age, sex, income, ethnicity, household size, employment level, meat cooking methods, degree of doneness, flavor profile preferences, meat consumption levels of beef, pork, chicken, fish, eggs and non-meat proteins at home and away from home and meat shopping habits were determined. The ballot included appearance, overall, flavor, juiciness and tenderness like/dislike using 9-point hedonic scales. After completion of each consumer session, four consumers were asked to participate in one-on-one exit interviews to determine attitudes toward beef and beef flavor.
Cut, cooking method and final cooked internal temperature endpoint affected meat flavor descriptive attributes. As expected, beef cuts had higher beef identity; pork cuts had higher pork identity; and chicken cuts had higher chicken identity. For beef cuts, cooking to higher internal cooked temperature endpoints increased beef flavor identity, but internal cooked endpoint temperature did not affect identity of pork and chicken cuts. Treatments differed in juiciness, tenderness and the major flavor attributes. Therefore, these treatments help to illustrate how millennial and non-millennial light and heavy beef eaters responded to meat flavor, juiciness and tenderness.
Samples as described previously, were used for consumer evaluations in Griffin, GA, Olathe, KS, Portland, OR and University Park, PA. Consumer demographics showed that there were slightly more females in the study and consumers were somewhat evenly distributed for age groups, household income and household size. Consumers were heavy consumers of chicken, beef, pork, fish and eggs and tended to eat these protein sources at home and away from home. For beef, chicken and pork, consumers used outside grilling for cooking followed by pan-frying and oven baking.
Consumers were segmented into four consumer groups (millennial light beef eaters, millennial heavy beef eaters, non-millennial light beef eaters and non-millennial heavy beef eaters). Millennial and non-millennial light beef eaters rated the beef, pork and chicken samples lower for overall liking and tended to rate samples lower for overall flavor liking and beef/pork/chicken flavor liking. Consumer group did not affect how consumers rated grill flavor, juiciness and tenderness liking. To further understand relationships between consumer attributes to overall liking, stepwise regression was conducted to predict overall liking using the other consumer attributes. Overall flavor liking was the first variable to enter into the regression equation and accounted for 78% of the variability in overall liking indicating that flavor was the strongest driver for overall like. Tenderness liking was the second variable to enter the equation and accounted for 4% additional variation in overall liking. These results indicate that overall flavor was the biggest driver of consumer liking and that tenderness, while not accounting for a great amount of variation, was still contributing to variation in overall consumer liking. Other attributes of flavor did not contribute appreciably to overall consumer liking.
Simple correlations were calculated between consumer and trained panel descriptive attributes. Relationships between consumer and trained sensory panel attributes were evaluated using partial least squares regression. As expected, flavor liking attributes were clustered together and the brown/roasted attribute was most closely related to overall liking. Tenderness and juiciness liking, while somewhat closely related to overall liking were closely associated with each other. Brown/roasted was the trained sensory attribute most closely related to overall liking and flavor liking attributes. Chicken identity and sensory tenderness attributes were closely associated, as the chicken treatments were more tender. Fat-like, burnt and juiciness were closely clustered with trained descriptive sensory attributes for juiciness. Warner-Bratzler shear force was negatively related to juiciness and tenderness attributes as would be expected as higher Warner-Bratzler shear force value have been related to tougher and drier meat. Rancid, cardboard, liver-like, warmed over, nutty, sour aromatics, bloody/serumy and metallic were negatively related to overall liking indicating that as levels of these attributes increased, overall liking decreased. Astringent, sour milk and spoiled putrid flavor and sour basic tastes were negatively related to overall liking. Pork identity, overall sweet and beef identity flavor attributes and sweet, salty and umami basic tastes were closely associated and clustered most closely with meat flavor liking. These results show that consumer liking and trained descriptive attributes are related across beef, pork and chicken products cooked to differ in flavor, tenderness and juiciness.
Raw chemical attributes were measured across treatments to determine if raw chemical attributes in the meat were predictors or indicators of consumer liking. These results showed moderate relationships between raw chemical data and consumer liking. Additional analysis examined if raw chemical data was related to consumer and trained descriptive attribute sensory evaluation using partial least squares regression. Raw chemical attributes were not closely clustered with consumer sensory attributes, but relationships were found between raw chemical composition and trained panel descriptive attributes and may explain some of the variation in flavor attributes in beef, pork and chicken.
Volatile aromatic chemical compounds (n=186) were identified in the cooked beef samples. To understand if volatile aromatic chemicals were associated with consumer liking, and trained descriptive flavor attribute data, stepwise regression equations were calculated for overall liking, beef flavor identity, pork flavor identity, chicken flavor identity, brown/roasted, bloody/serumy, fat-like, metallic, liver-like and umami. For flavor attributes, equations accounted for about 42 to 64% of the variation. For beef, pork and chicken identity, different volatile aromatic compounds were used to predict the dependent variable; however, chemical compounds associated with lipid degradation and Maillard reaction products were used in the equations. Of the 186 volatile aromatic compounds reported in the study, 117 of these compounds were used in the regression tables. These 117 volatile aromatic compounds were used in partial least squares regression bi-plots to further understand how volatile aromatic compounds were related to consumer and trained descriptive attributes. A number of volatile aromatic compounds were associated with beef, chicken and pork identity. 2,3-Butanedione and dimethyldisulfide were related to the metallic flavor attribute. Cardboardy and fat-like flavor aromatic were closely associated with lipid oxidation products. Volatile aromatic compounds were closely related to sour and bloody/serumy flavor aromatics. These results indicate that volatile aromatic compounds are related to trained descriptive sensory flavor attributes and could be used to measure flavor in meat products.
These data show that non-millennial light, non-millennial heavy, millennial light, and millennial heavy consumers responded to the 12 treatments that differed in flavor, juiciness and tenderness similarly. All consumer groups responding similarly to all palatability traits indicated that factors, other than palatability, drive millennial light beef eaters to not purchase beef. These factors may include social issues, time limitations or food preparation knowledge.
Figure 1. Partial least squares regression biplot of consumer liking sensory attributes (9-point hedonic scales; in blue) and trained meat descriptive attributes (0=none and 15=extremely intense; in red).