Su /nq UNITED STATES DEPARTMENT OF AGRICULTURE Bureau of Agricultural Economics A SURVEY OF QUALITY OF SELECTED BRANDS OF BUTTER SOLD IN ONE-POUND CARTONS AT RETAIL IN NEW YORK AND CHICAGO Gordon W. Sprague, Senior Agricultural Economist Gertrude G. Foelsch, Junior Agricultural Economist and Edward Small , Marketing Specialist 1 UNIV OF FL LIB. DOCUMENTS DEPT. 1 I <"* ry 0C\ 1 U.S. DEPOSITORY Washington, D. C. February 1939 Digitized by the Internet Archive in 2012 with funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation http://archive.org/details/butterbrandOOunit ▲ SURVEY Or QUALITY 07 SELECTED BRANDS OP BUTTER SOLD IN ONE POUND CARTONS AT RETAIL IN NEW YORK AND CHICAGO By Gordon W. Sprague, Senior Agricultural Economist, Gertrude G. Foelsch, Junior Agricultural Economist, and Edward Small, Marketing Specialist A consumer who buys butter that carries a certificate of quality issued by authority of the United States Department of Agriculture is likely to get a product of higher quality than she would have obtained had she bought butter that did not carry the certificate. This fact is borne out by the results of this survey which shows that although a great deal of butter that is sold without quality certificates is just as good as the butter that bears the certificate, a great deal is sold that could not meet the high quality standards required in order to obtain the cer- tificate. Therefore if a butter of high quality is wanted the safest procedure Is to look for the Government certificate when making purchases. A facsimile of the certificate is shown below. It shows the score, which indicates the quality, and the date the certificate was issued. CERTIFICATE OF QUALITY Issued by Authority of the United States Department of Agriculture. This is to certify that the churning of butter from which the butter in this package was taken was graded by an official United States Butter Grader and that the date of said grading and the number of the grading certificate issued are perforated hereon, and that the quality of it when graded was U. S. 92 Score or Higher DISTRIBUTED BY Note: To read the perforations: Read perforations from the back of certificate. First four numerals indicate number of grading certificate; the next two numerals indicate the calendar month, and the last two numerals, the day of month the butter was graded. At present, these certificates are issued only for butter that has been government graded U. S. 92 or 93 score. The scores are not percentages but are simply designating marks. It would be just as use- ful to use an H AA H for the 93 score and H A M for 92 score. Numerical scores, however, are of long standing in wholesale trading and are well understood and have been used by retailers largely for those reasons. The important point here is that both 92 score and 93 score butter are high-quality products. In order to provide the consumers with butter of this quality unusual precautions must be taken all along the line by the dairyman, the manufacturer and finally by the retail merchant. Failure at any point in the line may result in lowering the grade of the product. A little old cream left in a can, or failure to thoroughly - 2 - cleanse a vat, a pipe line, or a churn, may reduce the grade of a whole churning of "butter. Although butter is surprisingly resistant to deteri- oration if properly cared for it is, of course, a perishable commodity. A merchant who does not provide proper refrigeration or who puts the butter in the same compartment with products such as certain meats, fruits, vegetables, and other products that may impart an objectionable flavor to the butter, may actually reduce the quality of the butter. Since the grade of butter as indicated by the Certificate of Quality is determined before the butter is packaged, there is some possibility of a consumer obtaining butter of a lower grade than the certificate indicates, unless merchants pay careful attention to the conditions under which they keep their butter. PURPOSE Of STUDY Considerable criticism has been made of the Department of Agri- culture because of this fact. It is contended that since deterioration may take place the consumer is deceived. This study was not made to refute this argument v ut rather to discover to what extent butter carrying certif- icates of auality measured up to the grade placed upon the certificates at the time the butter was delivered to consumers at the retail stores. Nothing in this report should be construed to mean that butter which scores below 92 or 03 is unv.'holesome. Individual tastes vary greatly and some consumers prefer lower scoring butter. The tendency of consumers, however, appears to be toward a mild-flavored butter of low salt content. The investigation on which this report is based was made by grading actual purchases of butter made at retail stores. A representative of the Bureau bought a pound of butter as any housewife might, and carried it to the office where it was graded by government graders. The grade thus ascertained was compared with the grade stated on the certificate of quality if the butter carried a certificate. If the butter carried no certificate it was compared with its grade name and the claims ma.de on the carton. SCOPE AND METHOD PF STUDY It was decided that a selected sample, applicable so far as possible to large volumes of butter, would be the most satisfactory sub- stitute for a truly random sample. The large number of variables that may affect the keeping quality of butter between the time of manufacture and the time of purchase by the consumer were assumed to be most fully represented in the large consuming areas of the country. Therefore New York and Chicago were selected as representative markets to be sur- veyed. As these two markets have distinctive characteristics as to location, population, and industrial enterprises, most inter-market dif- ferences arising from these factors would be revealed in the respective samples. Upon the basis of these assumptions the following sampling plan was developed: - 3 - (1) Four areas were selected in each of these cities, each of these areas "being sampled once in each 4-week period, from January through April 1938. The areas were selected as follows: (A) A colored section. (B) A low-income white section. (C) A medium-income white section. (D) A very wealthy section. (2) The weekly sample was made up of 20 pounds of "butter as follows: (A) Eight pounds of "butter "bearing certificates of quality, three of which were "bought at large chain stores handling con- siderable volume, three at large independent stores, and two at small independent stores. (B) Twelve pounds of other "butter, five of which were bought at large chain stores, five at other stores handling consider- able volume, and two at small independent stores. In this group, ten of the twelve pounds of "butter "bought were of well- known, advertised "brands. The remaining two were local brands represented to the consumers as of high quality. No attempt was made to sample all qualities of "butter available to the consumers. (3) The "butter was "bought in solid-pound prints, rolls, and quarter-pound prints. Each weekly purchase was made within a 24-hour period, and was usually graded within 24 hours of the time of purchase. It was not possible, in all instances, to ^Main a complete sample of 20 pounds. This applied particularly to the colored section and low- income white section where "butter, especially the desired "brands, was not so generally available. In some stores in these areas only oleomargarine was offered for sale. The person who "bought the "butter asked for a pound of a certain kind of butter and paid for it. After the package was received at the office all identifying marks were removed and the package was given a number. Thus the grader of the "butter had no way of telling what particular brand it was. The "butter was graded "by three graders who were not allowed to talk during the operation. The grade placed upon it "by each grader was written on a slip of paper "bearing the same number as the package. These slips were then compared "by a fourth person who had not participated in the grading. Thus no collusion was possible and the chance of bias was almost entirely eliminated. Between January 1 and May 1 230 purchases were made at New York and 264 at Chicago, making a total of 404 samples. The data gathered on Form 1 (a copy of which is shown on the next page) "by the persons making the butter purchases and the gradings made by the federal "butter in- spectors provided the "basic material for this investigation. ~ 4 - Form 1 UNITEJ STATES DEPARTMENT OF AGRICULTURE BUREAU OF AGRICULTURAL ECONOMICS fiuali_ty_ of_Retail Butter City Score Certified^ _ Area Date Purchased Score at Purchase, Store: Chain_ _Independent_ _Delicatessen_ Days Since Cert 'n_ Retail Firm Brand Name : Claims for Butter Packed "by Certificate No. Price per pound Surface Flavor or Condition Ant. Sold TOcly-This "brand Total Type ©f Package, Roll Carton Solid Print Quartered Print : Days held in retail store : Held under refrigeration, Yes No Grading Name Final Held with other foods, Yes No Name or describe odors in refrigerator Score Description of quality - 5 - PRESENTATION OF DATA The results of these gradings, with the exception of 6 samples which were "No grade", axe shown in tables 1, 2, and 3.. The scores represent the interior quality of the "butter. Table 1.- Quality score of "butter bought at beth New York and Chicago With certificate of quality 2 to 1 chance of Average Standard "brand falling Standard Prand 1/ ( Certificate Packages score deviation within range below error Code Score Number Score Score Score Score A - N.Y. 92 8 91.88 .83 91.05-92.71 .32 Al - Chgo. 92 16 91.56 .90 90.66-92.46 ,23 B - N.Y. 93 15 92.53 .74 91.79-93.27 .20 Bl - Chgo. 93 13 92.08 .86 91.22-92.94 .25 C - N.Y. 93 7 02.57 .53 92.04-93.10 .22 Cl - Chgo. 93 12 91.83 .83 91.00-92.66 .25 Without certificate of qual i-ty D - N.Y. - 14 90.50 .85 89.65-91.35 .24 Dl - Chgo. - 14 89.79 1.12 88.67-90.91 .31 E - N.Y. - 5 91.00 1.12 89.88-92.12 .56 Ei - Chgo. - 10 89.30 1.42 . 87.88-90.72 .47 F - N.Y. - 15 91.87 .92 90.95-92.79 .24 Fi - Chgo. - 15 91.27 1.33 89.94-92.60 .36 G- - N.Y. - 9 . 89.89 .60 89.29-90.49 .21 Cl - Chgo. - 7 89.71 .95 88.76-90.66 .39 H - N.Y. - 13 90.15 1.14 89.01-91.29 • .33 Hi - Chgo. - 16 90.31 .95 89.3o-01.26 .24 I - N.Y. - 4 91.00 1.41 8 9. 59-02. 41 .82 u - Chgo. — 13 90.00 1.60 88.40-91.60 .48 1/ Like code symbols (A-Ai) indicate like brands. * is - Table 2.- Quality score of butter "bought at New York only With certificate of quality 2 to 1 chance of Score on Average Standard "brand falling Standard Sample Certificate Packages score deviation within range "below error Code Score Number Score Score Score Score a 93 10 92.60 .52 92.08-93.12 .17 b 93 6 92.50 .55 91.95-93.05 .24 « 93 8 92.50 .53 91.97-93.03 .20 4 1/ 93 38 92.71 .45 92.26-93.16 .07 el/ 92 5 91.40 1.14 90.26-92.54 .57 • Without certificate of quality f - 12 92.33 .78 91.55-93.11 .23 g - 7 92.43 .53 91.50-92.96 .22 h - 15 90.80 .68 90.12-91.48 .18 i - 6 92.50 .55 91.95-93.05 .24 J - 12 92.00 .60 91.40-92.60 .18 k 1/ - 17 91.24 1.60 89.64-92.84 .40 1/ Includes several brands Table 3.- Quality score of butter "bought at Chicago only With certificate of quality 2 to 1 chance of Score on Average Standard "brand falling Standard Sample Certificate Packages score deviation within range "below_ _e r _ r £ r _ Code Score Number Score Score Score Score 11/ 93 12 92.17 .83 91.34-93.00 .25 m 92 15 91.93 .70 91.23-92.63 .19 n 92 10 90.30 1.25 89.05-91.55 .42 Without certificate of quality o-16 91.81 .91 90.90-92.72 .24 p - 15 91.40 .99 90.41-92.39 .26 q - 16 91.44 .77 90.67-92.21 ,30 r - 15 88.73 1.38 87.35-90.11 .37 s - 8 91.38 .74 9C. 64-92. 12 .28 t - 7 91.14 .90 90.24-92.04 .37 u - 6 92.83 .41 92.42-93.24 .18 v - 6 91.33 1.51 89.82-92.84 .67 w 1/ - 10 91. CO 1.45 89.55-92.45 .48 x - 8 90.62 .74 89.88-91.36 .28 1/ Includes several "brands ■- 7 - Statistically it must "be recognized that although the scores are expressed in numbers, they are not numbers in fact but classifications. As they are not correlated at any point with real numerical values, the problem of measurement is not serious, and for the purposes of this survey the numerical scores are treated as real numbers. For each "brand of "butter an average score is shown. The first three brands in table 1 were sold with certificates of quality and the remainder are other brands. With one exception, the average score for each of the three brands sold with certificates of quality was higher than that for any of the others, indicating that insofar as the national brands of butter are concerned those sold with certificates of quality offer the con- sumers a greater certainty with respect to high quality. Brand F was the only one approaching the same average quality. JLnother characteristic of the average score is revealed by a com- parison of the results from the two markets. It is generally assumed that New York consumers are more quality conscious than those at Chicago. At first these averages seem to point to the verifying of this assumption for in every instance but one the score at New York was higher than at Chicago. This should not be true in the case of butter sold with certificates, how- ever, since the standards of quality were applied without reference to markets. The next point to be considered is the degree of standardization. It is desirable to know how consistently the score of a brand of butter compares with its average. This is shown by the standard deviation that measures the dispersion of individual scores about their mean. In this respect, brand G at New York was the most consistent of the brands that were not Government-graded. It had a standard deviation of only .6 of a quality score. With this single exception, Government-graded butter was the more highly standardized at New York. The standard error of the mean shows the limits within which the average score might be expected to vary if a different sample of the same brand had been bought. The quality scores of butter bought in New York but not duplicated in Chicago are shown in table 2. These brands are local in character. Sample d, for instance, is composed of several brands of butter sold with certificates of quality as 93 score. This represents very high quality and very highly standardized butter, as evidenced by a high average and a low standard deviation. Brands a, b, and c are about equally consistent. In this group of New York brands, however, it should be pointed out that brands f, g, i, and j also have average scores comparable with those of butter sold with certificates of quality and that, in addition, brands g, i, and j are equally well standardized. The quality scores of butter bought in Chicago but not duplicated in New York are shown in table 3. This record, for butter sold with cer- tificates of quality, shows a more serious problem. Brand n, for instance, is low in quality, poorly standardized, and indicates a real need for much closer supervision. Later investigation showed that the difficulty was due - fc - to the production from one creamery. The "butter in this "brand was scored "by the same inspectors both at the tine of grading* for the print room and after purchase, as was the case with all other "brands, but it did not have satisfactory keeping quality. At Chicago only sample 1 was of Government- graded 93 score. These small lots of "butter sold rith certificates of duality are not s© well standardized as they should "be for low-scoring samples appeared in almost every "brand listed. The best brand of butter bought at Chicago in this survey was sample u, while the butter in samples o, p, and q was equal to the "butter sold with certificates of quality. Another measure of the quality of Government-graded butter as com- pared with other "brands is given in column 5 of tables 1, 2, and 3. The ranges given in these .columns were computed "by measuring one standard deviation above and one below the respective average. With the exception cf "brand n, the housewife who "buys butter carrying a certificate of quality would have a 2-to-l chance of getting "butter that grades "between 9C and 93 score. In fact, the lower limit of the range is below 91 for only two brands. In the case of butter not carrying a certificate of quality, the housewife vrould have a 2-to-l chance of getting "butter that grades "between 87 and 93 score, with the lower limit of the range below 9C score for more than half of the brands. Quality in Government-Graded Butter In New York, 84 pounds of butter that carried 93 score certificates of quality were bought. In Chicago, only 37 pounds of comparable butter were bought. The score at time of purchase as compared with the original score is shown in table 4. About 54 percent of the butter that scored 93 was still of that score at the time of retail purchase. The butter that lost one full quality sc~re amounted to. '38 percent of the total 'and 8 per- cent lost more than one quality score. Table 4.- Loss in quality of 93 score butter between dates of grading and retail purchase If City : 93 score Score at_purchase Weighted average score at time 93 : 92 91 90 : of purchase New York Chicago : Lbs. 84 37 Lbs . 54 11 Lbs. 28 18 Lbs. c : 6 Lbs. ! c : Quality score 92.6 92.0 Total 121 65 46 p, 92.4 Percent 100.0 53.7 38.0 6.6 1.7 - 1/ Since the difference in behavior of the New York and Chicago samples was so wide with respect to quality loss, a Chi Square test was made to learn what probability existed that two such samples could be drawn from a universe described by the sum of the two. In this particular instance, the proba- bility was only one chance in a hundred, indicating that either there was a real difference in the character of the butter between the two markets or there was some bias in the selection or measurement of the sample. - 9 - Table 5 shows the loss in quality "between dates of grading and purchase of "butter that originally graded as 92 score, i. total of 54 pounds of butter carrying 92 score certificates of quality were bought at New York and Chicago in this survey. Of this total only 13 pounds were bought in New York, the remaining 41 pounds being collected in Chicago. At New York, 69 percent of the butter graded 92 score retained this quality at time of purchase, while in Chicago 59 percent retained the original quality. The quality scores of the butter when bought had a range of 90 to 93 score at New York. Less standardization was indicated at Chicago whore the butter thrt originally graded 92 score ranged from 89 to 93 score at time of retail purchase, but a Chi Square test showed that greater differences might be expected to occur merely as sampling fluctua- tions. Table 5.- Loss in quality of 92 score butter between dates of grading and purchase * City 92 score Score at purchase ! : Weighted average score at time 93 : 92 91 • 90 89 of purchase Lbs. 13 41 Lbs. Lbs, 7 : 20 1 2 ! 9 : Lbs. 2 4 Lbs. 4 Quality score New York i Chicago i : 2 : 4 : 91.7 91.4 Total 54 6 27 11 6 4 91.5 Percent 100.0 : 11.1 50.0 20.4 ! 11.1 7.4 - Data presented thus far indicate that, for purposes of comparison, New York and Chicago are not homogeneous and therefore in most cases they must be treated separately. In New York, 97 pounds of butter with certificates of quality and 132 pounds without certificates were bought. The fact that the samples were not selected at random prevents any very rigid interpretation of the comparisons of the score at the time of retail purchase of these two lots of butter. The sample sold with certificates was undoubtedly as representa- tive as is possible to make it, for almost all lots were included. In the attempt to include national brands sold in considerable volume the other sample was undeniably biased, showing higher than average quality. Many distributors, for instance, package low-grade brands that were not included in this sample. Some of the local brands were not characterized by consistently high quality, but these were inadequately represented. With these sug- gestions as to their limitations, tables 6 and 7 are offered, showing the distribution in quality at time of retail store purchase of butter both at New York and at Chicago, with and without certificates of quality. - 10 - Table 6.- Butter; Quality at time of consumer purchase in New York Quality score Class 93 32 . 31 90 89 88 Grade Gov't graded Not Gov't gra .ded Pounds 56 21 Pounds 35 40 Pounds 4 32 Found 9 2 24 Pounds g Pounds 4 Pounds 2 Total 77 75 36 26 9 4 2 Within the limits imposed "by the sampling procedure, table 6 shows that, at New York, Government graded, butter offers the consumer the greater assurance of quality. Table G also shows that a large quantity of fine butter was sold which was not Government-graded. The cOi^parison, at Chicago, between the two classes of butter is shown in table 7. Table 7.- Butter; Quality at time of consumer purchase in Chicago Class Quality score 93 92 31 30 83 88 87 BG '.'" Grade Gov ' t graded Not Gov't graded ! Founds 15 : 17 Pounds 38 48 Pounds 15 42 Pounds Pounds Founds 6 4- 41 18 12 Pounds 2 Pounds Pounds 2 4 Total 32 86 57 47 22 12 2 2 4 These comparisons favor certificates of quality from the point of view of the consumer who desires high quality butter. TIME OF HOLDING AS FACTOR AFFECTING INTERIOR QUALITY OF BUTTER A distribution of the quality scores for butter carrying 93 score certificates compared with the time lapsed between grading and retail pur- chase is found in table 8; a similar distribution for butter carrying 92 certificates is found in table 9. Time lapsed from the churn to the print room could not be measured. From a comparison of tables 8 and 9, it is evident that the pro- portion of butter included in the total sample is not equally divided as between 93 and 32 score at New York and Chicago. Seventy percent of the total samples carrying 93 score certificates of quality were bought in New York, while 75 percent of the total samples carrying 92 score certifi- cates of quality were bought in Chicago. A larger number of brands of butter carry certificates of quality in the New York markets than in Chicago. In this survey, samples of 21 brands with certificates of quality were bought in New York, all but 4 of which were 93 score. At Chicago, 8 brands of butter carrying certificates of quality were bought, 5 of which were 93 score and 3 of which were 52 score. - 11 - Table 8.- Butter, graded 93 score: Score at purchase compared with time lapsed between grading and purchase At New York Number 93 : score ! Score when ! Percent showing of ! day_s held bought a1 93 t 92 ! ! retail : 91 : 90 No loss ! 1 pt. ■ loss : More than l_pt .loss Bays 5 & less 6-10 11 - 15 16-20 21-25 26 & over Lbs . 17 21 20 11 i 6 7 Lb s . : Lb s . 11 : 6 14 : 7 » 15 : 5 4 : 6 6:0- 3 : 3 Lbs. : 1 1 | 1 Lbs . ! 1 : Percent ! 64.7 66.7 : 75.0 : 36.4 | 100.0 42.9 Percent : 35.3 35.3 ! 25.0 : 54.5 : .0 42.9 Percent : .0 .0 .0 • 9.1 : .0 14.2 Total 82 : 53 : 27 2 : 64.6 32.9 2.5 At Chicago 5 & less 1 1 ! .0 100.0 .0 6-10 17 7 10 , o 41.2 58.8 .0 11 - 15 9 2 4 2 1 i 22.2 : 44.5 - 33.3 16-20 3 1 1 ! 1 : 33.3 33.3 33.4 21 - 25 2 ! 2 .0 100.0 .0 26 & over 3 o 1 .0 .0 •100.0 Total 35 10 18 : 5 • 2 28.6 51.4 : 20.0 Table 9.- Butter, graded 92 score: Score at purchase compared with time lapsed between grading and purchase At New York Number 92 score Score when ; bought at retail Percent showing of No loss 1 pt. loss More than days held : ! 93 92 91 90 89 1 _pt . loss Days : Lbs . : •Lbs. Lbs. ' Lbs. Lbs. • Lbs. • Percent Percent : .0 • Percent 5 & less • • : .0 ! .0 6-10 • .0 .0 .0 11-15 ! 4 1 1 ' 1 1 : ! 50.0 25.0 25.0 16-20 : 3 3 : ' 100.0 .0 ! .0 21-25 : 2 ! 1 1 100.0 .0 .0 26 & over j 4 ! ' 2 1 1 1 50.0 25.0 25.0 Total : 13 2 7 2 • 2 ! 69.2 15.4 : 15.4 At Chicago 5 & less : 1 • • 1 : .0 : .0 : 100.0 6-10 : 7 ! 4 2 : 1 57.1 28.6 14.3 11-15 : 12 9 2 1 75.0 16.7 : 8.3 16-20 : 6 1 3 1 . 1 66.7 16.7 16.6 21-25 : 9 1 3 3 2 1 44.5 33.3 22.2 26 & over j_ 4 1 - 1 1 ! 1 - • ! 50. C : 25.0 25.0 Total : 39 ! 3 ; 20 9 : 4 3 ■ 58.9 ■ 23.1 ! 18.0 - 12 - For butter losing only one quality score, whether Government graded 93 score or 92 score, no relation v/as indicated in tables 8 and 9 between the number of days the butter was held and the score at the time purchased. The percentages shown in the tables support the assumption that butter with inherently good keeping quality, when well cared for, does not deteriorate even though held a comparatively long time. On the other hand high quality butter occasionally deteriorates rapidly. A point in fact is shown in table l . where, at Chicago, there were two instances of a 3-point loss in quality within 10 days from the time the butter v/as Government graded. This is an example of butter that was fine flavored when graded but lacked the keeping quality to maintain this flavor even for a short time, or was subjected to unfavorable conditions after grading and before final purchase. The two instances in which there was a 3-point loss in quality in 10 days emphasizes the need for very careful supervision of the product from the preparation of the raw material, through the manufacturing processes, and while the finished product is passing through the channels of distribution, in order to insure a fine product for the consumers. Although the number of instances from which conclusions may be drawn are very few, they seom to indicate that, in butter losing more than one point, time is a factor. This evidence suggests that where real latent de- fects are present, time gives them a chance to develop and once started they develop rapidly usually causing deterioration of more than one quality score. TIME ELAPSED FROM THE PRINT ROOM TO THE CONSUMERS A complete measurement of the time required to move butter from the churn to the print room was not available. For Government -graded butter, however, the time elapsed from the print room to the consumers was measure- able. Table 10.- Rate of turnover for butter from print room to consumers by income areas - Total purchases for various time periods Number pounds bought in each time period by income areas Time Period Low-Income Area : Medium-Income Area : High-Income Area Days : . Pounds Pounds Pounds 5 and less 3 9 6 10 and less : 14 : 25 : 24 15 and less 22 50 34 20 and less : 30 61. " 25 and less 37 : 69 42 Total 45 76 43 Number pounds bought in each time period as percent of total purchases Days P re .nt 5 and less 6.7 10 and less 31.1 15 and less 48.9 20 and less 66.7 25 and less 82.2 Total : 100.0 Percent 11.8 ' 32.9 65.8 80.3 90.8 100.0 F re ,-nt 14.0 55.8 79.1 86.0 97.7 100.0 - 13 - At all time periods the rate of turnover was greatest in the high income area, i.e., of the total pounds of butter bought in the high-income area 79 percent had been held 15 days or less after being Government -graded, as contrasted with 66 percent in the medium-income area, and 49 percent in the low- income area. SURFACE FIAVOR ■ The data previously presented and discussed referred to the interior quality of the butter purchased. Another problem in the keeping quality of butter has to do with its surface. The surface of a print of butter - the more exposed part - is first to absorb any flavors or odors and the first to show evidence of oxidation or defect caused by aerobia. Although the depth of penetration is probably less than one-sixteenth inch in most cases and varies with a very large number of conditions, it is none the less evident that any differenc. between the surface and the interior flavor of print butter may be a serious problem. The existence of a surface flavor different from the flavor in the interior of the retail package was noted throughout this survey, and during the latter part an attempt was made to score the surface separately. In high-scoring butter the existence of surface flavor detracts fron the original quality. In an analysis of the data on surface flavor collected at New York and Chicago, as given in table 11, the proportion of butter having a surface flavor different from the interior at New York differs significantly from that at Chicago. Table 11.- Surface flavor different from interior flavor reported on butter bought in New York and Chicago from Jan. 1 to May 1, 1938 City No change in surface flavor : Changed : surface flavor : : Pounds Total • Pounds Changed surface flavor as percent of total : Pounds : Percent New York 151 : 76 : 227 33.5 Chicago 99 : 161 : 260 61.9 Total 250 237 487 48.7 The difference illustrates again the limitation found in doing re- search work in two widely different areas. One of the conditions of proper sampling is that a constant probability exists. If this condition is not fulfilled, biased results will be obtained from the survey. Because of the differences that exist within the total sample, none of the percentages in table 11 can be taken as representative of the proportion of butter that develops a different surface flavor before reaching the consumers, but they do point out that the problem of surface flavor is real in both markets. - 14 - As the reports on surface flavor at New York and Chicago are not homogeneous, table 12 is confined to data that apply to Chicago where the larger sample was taken. This table shows the existence of surface flavor in retail packages at tine of purchase associated with the quality score given the interior at the sane time. Table 12.- Surface flavor different from interior flavor reported on butter bought in Chicago retail stores from January 1 to l.lay 1, 1938 Interior score at purchase Score : No change in surface flavor Pounds Changed surface flavor Pounds Total Pounds Changed surface as percent of P'ercent s flavor total 93 : : 24 32 75.0 92 : 27 59 86 68.6 91 21 36 57 : 63.2 90 : 20 27 47 57.4 89 13 9 22 : 40.9 88 & Less 10 6 16 : 37.5 Total 99 161 ! 260 : 61.9 The percentages given in the last column of this table indicate that surface flavor different from interior flavor is particularly associated with high-scoring butter. This difference in flavor is therefore important in a certificate of quality program in which consumer satisfaction is the ultimate goal. CONDITION OF HOLDING- IN RETAIL STORES Butter is assumed to absorb foreign odors readily, and therefore it was assumed that butter held apart from other foods in the refrigerator of the retail store would be more free from foreign flavors than butter held with other foods. But many retailers do not have separate compartments in their refrigerators for butter and other dairy products. Table 13 shows that in this survey the proportion of butter held under separate refriger- ation at New York and at Chicago was slightly over one-half and one-fourth respectively. The difference shown between the two markets may be a real one, or it may arise from the fact that, at New York, the preponderance of samples was taken in the medium- income and high-income areas where con- dition of holding may have differed from those in the low- income and Negro areas. - 15 - Table 13.- Surface flavor different fron interior flavor as related to con- dition of holding reported on butter bought from Jan* 1 to May 1, 1936 Refrigeration sepr \rate City New York Chicago Total : No change in surface flavor : Pounds : 80 32 : 112 : Changed ! surface flavor : : Pound s : 29 38 : 67 : Changed surface flavor : as percent of total . Pounds : Percent : 109 : 26.6 70 : 54.3 : 179 : 37.4 New York : Chicago : 62 66 Held with other foods 39 : 101 : 121 : 187 ; 38.6 64.7 Total : 128 : 160 : 288 : 55.6 At both cities sone butter was purchased which was not under refrig- eration. New York reported 17 pounds of which 8 pounds had different surface flavor; Chicago reported 3 pounds of which 2 pounds had different surface flavor. The data tabulated in table 13 disclose that a changed surface flavor develops in butter held in separate compar talents of refrigerators as well as in butter held with other foods. This fact is in agreement with research which indicates that some of the flavors that are usually classified as absorbed, may have other origin. But that foreign flavors may be ..Vjorbed, is shown by the fact that the larger percentage of butter held under re- frigeration with other foods developed surface flavors unlike the interior flavor. The difference between the prevalence of a changed surface flavor under the two conditions of holding is significant. It also shows con- siderable resistance to absorption of flavors which may be assumed to result from the fact that this butter was in cartons. Another factor assumed to be important in the change of surface flavor is the length of time the butter is hold from manufacture to con- sumption. Some measure of the time element is available in table 14 which classifies surface flavor according to the number of days that elapsed between the date the butter was graded for packaging with certificates of quality and the date purchased in the retail store. This table indicates that time of holding is an important factor as related to changed surface flavor. Table 14.- Surface flavor different from interior flavor as related to time from the print room to consumers on butter with certificates of quality, bought from January 1 to May 1, 1938 Days held . No change in : surface flavor : Changed : surface flavor Total Changed surface flavor : as percent of total 5 & less 6-15 16 - 25 26 & over Total Pounds 15 42 : 20 5 82 : Pounds 4 47 : 20 12 83 : Pounds : 19 : 89 I 40 17 165 : : Percent 21.0 : 52.8 : 50.0 70.6 50.3 - 16 - CLAIMS FOR BUTTER The statements or advertising which appear on butter when it is offered for consumer acceptance are natters of importance when consumer deception is under consideration. In this survey no consideration was given the oral presentation made by jobbers to retailers or by retailers to consumers. It is quite likely that considerable, though perhaps innocent misrepresentation of butter quality occurs in this way. Printed statements found on cartons and wrappers only were considered. In the case of butter that is sold with certificates of quality, all statements are carefully scrutinized in advance in order that no deceptive statement nay appear. Even in the case of 93 score, which is the finest available in large volume, superlatives are rarely permitted. Users of these certificates are restricted to advertising that is support- ed by facts. The certificate itself, together with the description of the standards under which the score is placed on the certificate, is considered sufficient evidence of the quality and further elaboration is not nccessarr, Claims for butter that is not sold with certificates of quality are not subjected to supervision except that which the policy of the distribu- tor imposes. As a result, printed statements are often irrelevant as re- ferring to butter quality and in some cases are deceptive. Some tabulated generalization with respect to these statements follows: (A.) Descriptions of quality appearing on butter cartons which should apply only to butter that grades 92 or 93 score: (1) Fine, finer, finest. (2) Best, very best. (3) Fancy, strictly fancy. (4) Highest. (B,) Adjectives used on cartons which have an indefinite connotation when applied to quality in butter: (1) Uniform. (2) Warranted pure. (3) Wholesome. (4) Exceptional. (C.) Conservative statements which do not exaggerate or misrepresent the quality of butter offered: (1) Some packages include only a brand name, (2) Pasteurized creamery butter. (3) Made from pasteurized cream. - 17. - In the REVISED TENTATIVE UNITED STATES STANDARDS FOR QUALITY OF CREAMERY BUTTER ^nly two adjectives are used to describe butter having the highest quality - "fine" and "pleasing". Fine flavor applies only to 93 score quality and pleasing flavor is limited to 92 score. Butter having a fairly pleasing flavor may grade 90 to 91 scorn. i In the samples ^f butter collected that did not carry certificates of quality, the words fine, finer, or finest wore U3ed for 7 brands to describe either the quality of the cream used, or the flavor of the butter contained in the carton. Judging from the description on the package, the customer could expect to receive butter that originally scored 92 or 93. Forty-five individual prints advertised on the carton as "fine" were collected in this survey. The quality range was from 88 to 93 score. Eut only one print graded ~8 and one 93. As 26 percent of the samples graded 92 to 93 score, 36 percent 91 score, and 38 percent 88 to 90 score, it is evident that the consumer could not be sure of receiving the quality that the statement on the carton could load him to expect. The remainder of the descriptive terms listed under (A) also imply butter of 92 to 93 score quality since they are superlatives and therefore should represent only tho highest quality. Adjectives like best, fancy, and highest quality, wera included ia. statements found on 8 brands of butter, consisting of 72 one -pound prints. Tho quality score distribution was as f ollows : Within a range of 89 to 93 score, 18 percent graded 93 score, 41 percent 92, 24 percent 91, and 17 percent 89 to 90 score. Al- though there is considerable chance for consumer deception in this group, the percentage of butter scoring 92 to 93 score is sizable, showing that a number of the brands were not misrepresented by the descriptions on the cartons. The group of adjectives under (B) are too indefinite or vague to be *f much value to a consumer in determining the quality of a carton of butter. The sample collected having descriptions like those given under (B) consisted of 42 prints, that had a quality range of 86 to 93 score, distributed as follows: 18 percent 93, 29 percent 92, 17 percent 91, and 36 percent 86 to 90 score. Butter in this group ranged from the poorest to the very finest. From the three groups under discussion it is clear that there is little relation between the claims for brands of butter that do not carry certificates of quality and the actual quality of the butter. As indicated under (C) there are a number of distributors who make no statements, or only very conservative statements, as to the quality of their brands. This group includes some brands that have a high average quality score. - 18 - Quality Variation by Income Aro_as In planning this study of the quality of butter sold at retail, areas that have distinct income characteristics were selected in each city, for the purpose of learning whether specified brands for sale in all sections of the city maintained uniform quality throughout the market. The sample at Chicago was complete for 10 well-known brands, which did not carry certificates of quality. The average quality score for each brand was calculated and deviations from this average were tabulated by income areas. For some of the brands there was no tendency for the quality to be either above or below average, as related to income areas. Other brands were below average in the low-income and Negro areas and significantly above average in the high- income areas, indicating that the brands did not represent the same quality throughout the city. Whole these differences existed, the claims ri.de for certain brands might be representative of the quality for one income area but not for another. THE SIGNIFICANCE OF QUALITY RESEARCH Butter, if properly produced, is one of the most satisfactory forms of fat for human food because of its fine, pleasing flavor and because it includes vitamins, particularly the fat soluble vitamin A which makes the liver oils of salmon, halibut, and cod so valuable. Eutter is also a good source of vitamin D. Butter, like most articles of food, is perishable, although sur- prisingly resistant to the defects of perishability. A very large quantity of fine research has been completed by chemists, physicists, and bacteriol- ogists which has produced a voluminous body of literature on the subject and a very greatly improved technique of production and marketing. All too often this research directs popular attention to the defect rather than to the means of control. Consumers should recognize the fact that the discussion of the problem of quality preservation by scientists is for the purpose of making butter an even more satisfactory food. Dis- cussion of the problems that are met and controlled in the production and marketing of butter should be reassuring rather than alarming to consumers, for they indicate that the industry is alert to its responsibility in pro- ducing a fine food product. The fact that such a very large proportion of the butter bought and scored in the conduct of this study showed little deterioration at the time of purchase is evidence of the good effects of research already done. Yet merchandisers of butter, either with or without certificates of quality, must face the problem that some inherent defects still may be concealed at the time of grading. Such defects, although rare, may develop later in the process of distribution after it has passed out of the original owners control. Continued research on keeping quality problems should help eliminate the occasional inherent defects. At the - 19 - same time technicians trained for advancement in this research, as well as for using the results of research already completed, are needed. Some butter manufacturers do employ such technicians to whom they give broad authority over production operations. In the process of gathering and scoring the 494 samples of butter reported in this survey, 6 samples were found that were "no grade". This means that according to the standards of the Bureau of Agricultural Economics such butter was below the requirements of U. S. 85 score because of its flavor or other conditions. These samples were found in brands of butter that usually score high. The fact that they were not found in butter sold with certificates of quality is probably an accident because it is in- possible, by scoring alone, to detect a condition that nay later cause butter to show narked deterioration. Although butter that shows narked deterioration in the cartons is rare, the fact remains that the occurence is frequent enough to raise a problen that some fims have attacked by testing the keeping quality of a snail sample from each churning. This method is so devised that it is possible to trace butter back to the producing creamery whore experts are usually able to find the cause. In some instances it might seem impractical to hold butter for the period of 7 to 10 days usually required for testing its keeping quality although in some instances this is actually done. Since lack of keeping quality seems often to be associated with the product of certain creameries, the conduct of keeping quality (incubation) tests might still offer a method of control for dealers interested in high quality by indicating the source of butter which deteriorated badly. Even though the churnings were not held until incubation was completed, the incubation of all samples of high quality butter would identify question- able lots and protect the distributor in so far as later lots from the sane creamery were concerned until the source of the defect was placed under control. STANDARDIZATION AND GRADING Standardisation of quality is a complex problen when applied to any perishable commodity. It is particularly conplex when applied to butter. The grading concept implies that the supply of butter can be arrayed from the poorest to the finest quality available. It further implies that this array nay be subdivided into classes, each having connon character- istics and between which the lines of denarcation are distinct. The grading concept for marketing purposes also has a value im- plication. The criteria for establishing the lines of division between the classes in the array of butter quality should be those dictated by demand. The total supply should be adjusted to demand in such a way as to give the greatest value to the product. UNIVERSITY OF FLORIDA 3 1262 08921 4737 - 20 - The grades now in use are baaed on the result of research in the physic.".! sciences and on practical experience. These are fully described in the OFFICIAL TOUTED STATES STANDARDS FOR QUALITY OF C ; ';' BUTTER, issued by the Bureau of Agricultural Economics, U. S. Department of Agriculture • SUMMARY AND CONCLUSIONS The existence and general application of any kind of grading for quality, by whomsoever applied, presupposes that a service of this kind is practical and can be rendered. The results of this research lend strong support to this hypothesis by showing that, as between butter sold with certificates of quality and all other butter, the brands carrying certifi- cates of quality were of higher average quality and were more uniformly standardized. Consumers who wished to buy butter of the character described as 92 and 93 score, but who were unable themselves to sample and judge this quality, were more likely to find it consistently by buying butter packaged with certificates of quality than by following the advertising statenents printed on butter packages. Some brands of butter, however, were found equal in average quality and standardization to that which was Government- graded. A flavor at the surface different from that at the interior was frequently noted. This condition occurred more often in high-scoring butter than in low-scoring butter because against a background of fine flavor, such a difference is more noticeable. Conditions of refriger- ation in retail stores appear to be a contributing factor in the appear- ance of changed surface flavor. The instances in which the quality had deteriorated more than one point from the original score, although few, were still numerous enough to indicate that keeping quality is a serious problem. For the purpose of identification of butter which lacks keeping quality and for the pre- vention of its use in cartons carrying certificates of quality, a wider use of incubation tests of keeping quality is desirable. *** **