U.S. Department of Agriculture
Food Safety and Inspection Service
Washington, DC 20250

David Soderberg, Emerging Issues Branch
Michael K. Hoffman, Emerging Issues Branch




A study addressed the question, "How often does ground beef, purchased at retail and cooked at home, turn brown before reaching 160 F?" The study used only pure ground beef and ground beef patties.

Two hundred and forty pairs of ground beef packages, each package weighing 1.25 pounds or more, were scheduled to be purchased at retail in nine locations across the country. To conserve the chemistry of the meat color, samples were bought near the laboratories or shipped to them by overnight mail. None of the samples were frozen before shipment to the laboratory. USDA personnel were instructed to purchase the ground beef samples as though they were buying them for their own use.

At the laboratory, one of the two packages from each sample pair was frozen for use one week later. The other package was made into several 4 ounce patties of about 5/8 inch thickness using a hand patty mold with a diameter of 3.5 inches. These patties were then cooked "fresh" using an electric griddle set at 325 F. (This is Treatment A in Table 1, below.) Patties were cooked to one of four temperatures: 135 F, 150 F, 160 F and 175 F. Patties were cooked for 2 minutes on each side, then flipped at 1 minute intervals until the endpoint cooking temperature was reached. The cooked patties were cut in half and evaluated for visual meat color, pinkness of the juices and visual cooked texture of the cut surface. Other supporting data, such as instrumental color, percent cook yield, cook time, pH, and percent fat, were also collected. Instrumental color was recorded using a reflectance colorimeter. The redness (a*) parameter was of most interest.

Most of the frozen packages were held for one week and then were thawed by one of four methods commonly reported by consumers (see Table 1 below). For three of the thawing treatments (B, C, and D), the samples had been made into patties before freezing. Samples that were to be thawed overnight (treatments E and F) had been frozen in the bulk retail package. The statistical importance of each of the sample handling techniques may be estimated from the USDA 1995 Continuing Survey of Food Intakes by Individuals. The different treatments that were tested in the laboratories and their prevalence of use by consumers are shown in Table 1.

(A)Cooked Fresh (estimate)29.0%
(B)Frozen; thawed in a microwave.7.5%
(C)Frozen; thawed 2 hours at room temperature8.0%
(D)Frozen; thawed 4 hours in a refrigerator15.0%
(E & F)Frozen; thawed overnight in a refrigerator5.5%

Thawing for treatments E and F is the same, so these treatments cannot be separated in Table 1. Treatment E samples were stored frozen for one week. Treatment F consisted of twenty extra samples that were stored frozen for an entire month to determine whether storage time might also have been a relevant factor in premature browning.

The color of the cooked ground beef patties was visually assessed using a single, standardized viewing procedure that compared color to the Kansas Agricultural Experiment Station Ground Beef Patty Cooked Color Guide. This guide has pictures of ground beef patties from red to brown, with scores from 1 (most red) to 5 (most brown). The numeric scoring used in this study against these pictures covered the range of 0.5 to 5.0, in 0.5 increments. Verbal descriptors of color were not used in these evaluations, with one important exception - scores of 4.0 and above could not contain any pink. This meant that 3.5, the next decremental score below 4.0, was selected when observers perceived just a trace of pink remaining in a cooked ground beef patty.

The color of the juice was visually assessed for the presence of pink color. Analysts were also asked to describe the appearance of the texture of cut surface - whether it appeared cooked or not, and to record the texture using adjectives such as "wet," "glossy," "soft," or "mushy."


Visual color scores were reported that covered the entire reportable range of 0.5 to 5.0 at all four temperatures. As the temperature increased from 135 F to 175 F the scores tended to shift from red towards brown. The score of 3.5 was prominently reported at all temperatures except 135 F. Since a score of 3.5 may be interpreted as the borderline score between pink and brown, it is clear that many observations fell at this border where a slight remnant of pink was difficult to discriminate from brown. The fact that ground beef color scores representing all colors from red through brown were observed for all cooking temperatures, including those temperatures less than 160 F, and that many were borderline observations, clearly shows the food safety and public health problems inherent in a "cook-until-brown-in-the-middle" message. That is, "premature browning" is a real and serious problem that must be explicitly addressed.

This conclusion was again confirmed when median visual color scores for the patties were subjected to a linear regression against the endpoint cooking temperatures. The two correlated at 0.54 (r2= 0.29), so that temperature accounted for about 30% of the overall variability of the visual color scores. A regression of the average instrumental color a* values for the patties versus cooked color yielded a similar result.

Analysis of Variance (ANOVA) of visual scores was performed for the variables, COOKTEMP, FRESH, TAKENBY, and LAB compared to the median visual score for each patty. These variables are defined as follows:

COOKTEMPendpoint cooking temperature,
FRESHfreezing/thawing treatment (B,C,D,E, or F), or cooked fresh (A)
TAKENBYsample was purchased by the laboratory or by a Compliance Office
and shipped to the laboratory,
LABthe laboratory where the samples were cooked and analyzed.

The variables COOKTEMP, FRESH, and LAB all had a significant correlation with visual color. The variable TAKENBY was not significant. The most important meaning of this result is, of course, that the cooking temperature (COOKTEMP) does definitely affect the result of the visual color evaluation.

The treatments encompassed by the FRESH variable formed three clusters. Treatments E and F (both treatments were frozen as a whole package and thawed overnight in a refrigerator) form one such cluster. Treatments B, C, and D (all frozen as patties and thawed in a relatively short time) form a second cluster. Treatment A formed the third cluster.

When ground beef was frozen in whole retail packages and then thawed overnight in a refrigerator, premature browning was greatly increased. At 150 F, as many as two thirds of ground beef patties prematurely turned brown. By contrast, when ground beef that had been frozen as patties had been thawed only a few hours, or in a microwave, and cooked to 150 F, fewer than 15% turned brown.

Treatment A (cooked fresh) showed less premature browning than the E and F cluster, but more than the B, C and D cluster.

The fact that these three different correlations were found means that the way consumers handle ground beef prior to cooking is an important factor in the amount of premature browning encountered.

ANOVA's were also calculated at each temperature. The variable TAKENBY was again shown to be unimportant. The variables FRESH and LAB were once again shown to be important.

The effect of FRESH on visual scores is clear and consistent. If the proportion of patties that were scored brown (greater than 3.5 or 4.0) at 135 F and 150 F is calculated, this effect becomes even more obvious. This effect, summarized below (Table 2), is a critical outcome of the study.

Temperature 135 F 135 F 150 F 150 F
Visual Color Score>3.5>4.0 >3.5>4.0

Treatment A5%2%28%12%
Treatment B9%2%14%7%
Treatment C7%2%13%7%
Treatment D8%6%13%6%
Treatment E18%7%64%48%
Treatment F22%7%63%33%

An analysis of the cooking data also showed that cooking to 160 F cannot ensure that a ground beef patty will turn brown. Nearly half of all patties retained some pink color when cooked to 160 F and nearly one fifth still retained some pink color at 175 F.

The instrumental color scores, a* (redness), were also subjected to an ANOVA analysis at each temperature. Differences in a* values for the variable FRESH show a strong similarity to the differences in visual score at each temperature. This suggests that the observed color effects are due to physical/chemical differences, and are not just an observer phenomenon.

The correlation of the raw color of the uncooked ground beef with the color of the cooked ground beef was also determined. This is important because the myoglobin chemistry is believed to have a strong effect upon the temperature at which the ground beef changes color. Since myoglobin chemistry also affects the color of the raw ground beef, it was natural to ask whether initial differences in color could be used to predict the extent of premature browning seen upon cooking.

Instrumental color measurements were obtained from both the surface (external) and the interior (internal) for each package of ground beef. The external raw instrumental color and the internal raw instrumental color were poorly correlated, with correlation coefficients ranging from ~0.16 - 0.24. Because this correlation is small, it would not be possible to predict the internal color of ground beef from the external color of a package of ground beef. No useful correlation between external raw color and the color of the cooked ground beef patties was found. Internal raw color does have a small correlation with the cooked color at the lower cooking temperatures, but this becomes insignificant as cooking temperatures increase above 160 F. In summary, a consumer cannot judge whether ground beef will turn brown before reaching 160 F by looking at the raw product.

The effects of cook time, percent cook yield, pH, and total fat on the visual color of the cooked ground beef patties were also tested. An ANOVA showed that cook yield, pH and fat were significant factors. Cook time, which is the variability in time needed to cook to any specific temperature, was not a significant factor.

The juices are more heat resistant than is the visual color of the ground beef patty itself. At 160 F, almost 75% of the cooked patties still had pink juices. The different freeze/thaw treatments affect juice color similarly to the way they affected the meat color. That is, three groupings were found: treatments B, C and D; treatment E and F; and treatment A.

The analysis of the texture data paralleled the findings for the juices.


The ultimate objective of this study was to try to determine, "How often does ground beef, purchased at retail and cooked at home, turn brown before reaching 160 F?" The answer , based on weighting of different consumer freezing and thawing practices, taken from the USDA Continuing Survey of Food Intake by Individuals, is as follows:

Final Cooking
Percentage with Visual Color
Scores of 3.5 or Higher
Percentage with Visual Color
Scores of 4.0 or Higher
135 F7.5%3.2%
150 F20.6%10.2%
160 F41.3%23.3%

This study has confirmed that the premature browning phenomenon is real.

At 135 F, up to 7.5% of ground beef patties were considered to be brown or nearly brown (Table 3). At 150 F, more than 25% of all fresh-cooked patties (Table 2, Treatment A) were scored as brown or nearly brown.

Freezing and thawing adds further complexity to color evaluation. As many as two thirds of ground beef patties made from previously frozen meat were judged to turn brown prematurely (Table 2, Treatments E and F).

Even cooking ground beef patties to 160 F or higher cannot ensure they will turn brown. Nearly half of all patties retained some pink color when cooked to 160 F, and almost 20% still retained some pink color at 175 F.

Brown color is not a good indicator that a hamburger has been cooked to 160 F.

*These results must be interpreted with much caution. Many uncertainties arise both in generating the experimental values and in combining the different sources of information that go into these calculations.

Click here for further information about this publication.

Click here to learn more about the Office of Public Health and Science.

URL: http://www.fsis.usda.gov/ophs/shortres.htm
Last Updated On 06/05/1998.