A STUDY OF COinMER PREFERENCE FOR TURKEYS WITH VARIOUS GRADE DESIGNATEONS AND PRICES T5953: {or ”to Dsqm of pH. D. MICHIGAN STATE UNEVEB’SITY Wilfiiam Clearer: Mills. Jr. 1963 This is to certify that the thesis entitled A Study of Consumer Preference for Turkeys with Various Grade Designations and Prices presented by William Clearon Mills, Jr. has been accepted towards fulfillment of the requirements for Ph. D. degree in Poultry Science jor professo J ‘/ ’ jéme/ fl . Mn Edi/(“J Date-l Z‘Z ht é. {63 ‘ LIBRAR Y Michigan State ‘ University I l ABSTRACT A STUDY OF CONSUMER PREFERENCES FOR TURKEYS WITH VARIOUS GRADE DESIGNATIONS AND PRICES by William Clearon Mills, Jr. This study was undertaken in order to arrive at a better understand- ing of price influence and the effect of "Prime" and "Choice" as grade designations on purchases of turkey hens. Two main questions were investigated: (1) How much more would con- sumers pay for turkeys labeled as "Prime", "Choice", and "Good" compared with Grade A and B; (2) At what price differential would a change in pref- erence occur due to price. Judgment sampling techniques were used to select stores in Charlotte, North Carolina; Grand Rapids, Lansing and Detroit, Michigan. Turkey hens (10-14 pounds) labeled "Prime", "Choice" were sold dur- ing August-October and before Thanksgiving (1962) and before Easter (1963) in competition with turkeys labeled Grade A. Various other combinations were also sold using two different labeling methods. Price differentials of 2, 4, and 6 cents per pound above the prevailing Grade A retail price (control group) were used. Consumer preference studies were made to evaluate the effect of "Prime", "Choice" and "Good" as grade designations on raw turkey parts, and cooked rolls with and without prices. The base price was the prevail- ing retail price and 2 and 4 cents price differentials were added to the test labels. The results of the sales tests were analyzed by Chi Square and the "t" test. Confidence intervals for all proportions were obtained. The William Clearon Mills, Jr. theoretical sales ratio of Grade A to "Prime" or "Choice" was 2:1. The sales ratios (price differentials, "Prime" and "Choice" com- bined) were: Grand Rapids 55:45; Lansing 75:25; Detroit 62:38. The combined sales ratio was 63:37 which is shown by Chi Square not to be significantly different from the theoretical sales ratio. Confidence intervals for the combined sales of "Prime" and "Choice" are 36 to 40 per cent. The variability between the several store sales within cities was much less than the variability of sales between cities. The panel results indicated that "Prime", "Choice", and "Good" are preferred as grade designations on raw turkey parts and cooked products. Based upon the results obtained in this study, it is concluded that: l. One-third of the turkey hens can be sold for 4 cents per pound more than the current price of Grade A turkeys when adequately differ- entiated and labeled "Prime". 2. Consumer satisfaction was apparently increased as measured by the willingness of some consumers to pay premium prices for turkeys in this study. 3. Some of the criteria by which consumers judged turkey quality in this experiment are: appearance, meatiness, wholesomeness, minor defects. 4. Consumers reacted favorably to grade labels "Prime", "Choice", and "Good" as labels for whole turkeys as well as for raw turkey parts and cooked turkey products. 5. One might expect aggregate sales of turkey hens in the United States to be similar to the sales ratios obtained in this experiment. 6. Consumer grades for turkey hens should be "Prime" and "Choice". 7. All other turkeys not meeting the specifications for the grades William Clearon Mills, Jr. of "Prime" and "Choice" should be classified as "Good" and used for canning and other processed foods. 8. The proposed grading system apparently more nearly reflects consumers' opinions of quality than the present grading system. 9. Total industry revenue from a given size of crop would likely be increased under the proposed grading system. This study has pointed the way to possible increased profits for producers, processors, retailers, and greater satisfaction for the con- sumer. The turkey industry can profit by grade revision, produce dif- ferentiation, and proper pricing. A STUDY OF CONSUMER PREFERENCE FOR TURKEYS WITH VARIOUS GRADE DESIGNATIONS AND PRICES BY William Clearon Mills, Jr, A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Poultry Science 1963 f‘ a‘ixxs 7/8/94 ACKNOWLEDGEMENTS To my wife and North Carolina State College, I am deeply grateful for their encouragement and assistance in making it possible for me to pursue graduate study. To the National Turkey Federation goes much of the credit for sponsoring this study. Without their sponsorship, this study would have been delayed. I owe many thanks to the Poultry Science Department, to the Agri- cultural Economics Department, and to the Food Science Department, Mich- igan State University, for their support of this study. Especially to Professor J. A. Davidson, my Committee Chairman, Dr. Henry Larzelere, Director of my thesis, Dr. H. C. Zindel, Dr. Lawrence Dawson, and Dr. T. H. Coleman, I wish to express my gratitude for their counsel and guidance during the study and the writing of this thesis. Without the cooperation of the retail stores, this study would not have been possible. Mr. Bud Schmidt, Lansing, and Mr. Van Valkenburg of Grand Rapids, and Mr. Harris of Charlotte, North Carolina, were most helpful in cooperating with the author in conducting sales in their sup- ermarkets. The other cooperator wishes to remain anonymous, but was no less helpful. To the United States Department of Agriculture, Poultry Division, Mr. R. W. Kempers (Chief, Grading Branch) and to the Michigan Department of Agriculture (Division of Foods and Standards, Mr. Lyle Littlefield, Director) and to the North Carolina Department of Agriculture (Mr. L. Y. Ballentine, Commissioner), the author expresses appreciation for their cooperation in allowing experimental grade labels to be used. ii To Mrs. M. J. Bostick, Department of Home Economics, Wayne State University, the author expresses thanks for her assistance in conducting the panel portion of this study. Appreciation is expressed to Dr. W. D. Baten, Professor Emeritus of the Statistics Department,for assistance with the statistical ana- lysis. iii TABLE OF CONTENTS ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES INTRODUCTION REVIEW OF LITERATURE Development of Grading Early History of Standards Poultry Legislation and Grade Deve10pment 1933 Grade Standards for Turkeys 1960 Grade Standards for Turkeys Grade Labeling Determination of Standards PROCEDURE Sales Experiments Preference Panel Studies RESULTS Sales Tests Preference Panel Studies DISCUSSION Sales Tests Grading in General Present System Proposed System Prime Grade Specifications Choice Grade Specifications RECOMMENDATIONS FOR FURTHER RESEARCH SUMMARY AND CONCLUSIONS BIBLIOGRAPHY APPENDIX iii Page ii iii iv vi vii 86 9O 92 Table 10 ll 12 13 LIST OF TABLES Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled Prime (4 Cents per Pound Price Differential) . . . . . . . . . . . . . . . . Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled Prime (6 Cents per Pound Price Differential) . . . . . . . . . . . . . . . Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled Choice (4 Cents per Pound Price Differential) . . . . . . . . . . . . . . . Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled Choice (6 Cents per Pound Price Differential) . . . . . . . . . . . . . . . . . . . Sales Summary, by Cities, of Grade A Turkeys with Combined Sales of Prime and Choice Turkeys (4 Cents per Pound Price Differential) . . . . . . . . Sales Summary, by Cities, of Grade A Turkeys with Combined Sales of Prime and Choice Turkeys (6 Cents per Pound Price Differential) . . . . . . . . . . . . . . Summary of Turkey Sales (Price Differentials Combined). Turkey Sales Comparing Choice and Prime as Grade Labels . Turkey Sales Comparing Good and Choice as Grade Labels Turkey Sales Comparing Good with no Grade Label - Grade B Turkeys . . . . . . . . . . . . . . . . . . . . . . . Influence of Grade Label on Preference for Cooked Turkey Rolls by the Detroit Preference Panel. Afternoon, April, 1963 O O G O O G O O O O O O O O O O O O O O O O O O 0 Influence of Grade Label and Price on Preference for Cooked Turkey Rolls by the Detroit Preference Panel. Evening, April, 1963 . . . . . . . . . . . . . . . . . Influence of Grade Label on Preference for Cooked, Sliced and Packaged Turkey Roll by the Detroit Preference Panel. Afternoon, April, 1963. . . . . . . . . . . . . . . . . . iv Page 40 42 43 44 45 46 47 47 48 50 52 52 53 Table 14 15 16 17 18 19 20 21 22 23 24 25 Influence of Grade Label and Price on Preference for Cooked, Sliced and Packaged Turkey Roll by the Detroit Preference Panel. Evening, April, 1963. . . . Influence of Grade Label on Preference for Raw Turkey Breast by the Detroit Preference Panel. Afternoon, April, 1963. Influence of Grade Label and Price on Preference for Raw Turkey Breast by the Detroit Preference Panel. Eyening, April, 1963. Influence of Grade Label on Preference for Raw Turkey Thighs by the Detroit Preference Panel. Afternoon, April, 1963. Influence of Grade Label and Price on Preference for Raw Turkey Thighs by the Detroit Preference Panel. Evening, April, 1963. . . . . . . . . . Influence of Grade Label on Preference for Raw Turkey Legs by the Detroit Preference Panel. Afternoon, April, 1963. Influence of Grade Label and Price on Preference for Raw Turkey Legs by the Detroit Preference Panel. Evening, April, 1963. Calculated Gross Returns to Producers for Two Grading Systems, Turkey Hens . . . . . Calculated Returns to Retailers in this Study vs. All Turkeys Sold as Grade A . . . . . . . . . Detroit Turkey Sales by Stores, Grade Designation and Price, April, 1963. . . . . . . . . Grand Rapids Turkey Sales by Stores, Grade Designation and Price, November, 1962. . . . . . . . . . Lansing Turkey Sales by Stores, Grade Designation and Price, November, 1962. . . . . . . . . Page 53 55 55 56 57 58 58 80 81 93 94 95 Figure 10 11 12 13 14 LIST OF FIGURES Illustration Showing Turkeys as Labeled for Sales Test (Charlotte, Lansing and Grand Rapids). Illustration Showing Turkeys as Labeled for Sales Test (Detroit). . . Illustration of Labels Used in Charlotte, Grand Rapids and Lansing. Illustration of Store Label Attached to Grand Labels. Illustration of Turkeys Labeled in Detroit Sales and Actual Labels Used. Turkey Rolls as Labeled for Panel Test. Cooked, Sliced Turkey Roll as Labeled for Panel Test. Raw Turkey Breasts as Labeled for Panel Test. Raw Turkey Thighs as Labeled for Panel Test. Raw Turkey Legs as Labeled for Panel Test. Percentages of Total Prime Turkey Sales. Percentages of Total Choice Turkey Sales. Illustration of Demand Curve Movement. Illustration of Demand Curve Movement. vi Page 22 23 28 28 31 36 36 37 38 38 63 63 64 64 Appendix LIST OF APPENDICES Detroit Turkey Sales by Stores, Grade Designation and Price (April, 1963). . . . . . . . . . . . . Grand Rapids Turkey Sales by Stores, Grade Designations and Price (November, 1962). . . . . . . . . Lansing Turkey Sales by Stores, Grade Designations and Price (November, 1962). . . . . . . . . Computation of Coefficient of Concordance and Chi Square for Panel Data. Statistical Analysis of Sales Data. Chi Square for Table 1 . Confidence Interval for Table 1 Confidence Interval for Table 2 Chi Square for Table 2 . . Chi Square for Table 3 . . Chi Square for Table 4 . . . . . . . . . . . . Confidence Intervals for Tables 3, 4 . . . . . . . Chi Square for Table 7 . . . . . Confidence Interval for Table 7 . . Differences Between Stores - Grand Rapids . Differences Between Stores - Lansing . . . Differences in Sales Ratios of Grade A Turkeys (between Cities) . . . . . . . . . . . . Confidence Intervals for Real Difference in Grade A Sales, Lansing and Detroit . . . . . . . vii Page 93 94 95 96 97 97 98 98 99 100 101 102 103 103 104 105 106 106 INTRODUCTION The question of mandatory consumer grades for turkeys is one that has spawned widespread interest and verbosity. Not only are there strong economic arguments for such grades, but also strong arguments against such grades. If mandatory grade labeling comes to pass, What system of grade labeling will be used? Is the present letter grade system best, or is there another system that is more adequately suited to labeling for consumer acceptance? Before an intelligent decision can be made regarding mandatory grade labeling, it is necessary to decide what grade labels to use and to determine the most satisfactory grading system to use. What grade label has the highest psychological significance of quality to the person buying turkeys? It appears to some that the present system is prolix and no longer adequately expresses current consumer preferences and advances in technology in either production or marketing. The tmprovement of turkeys is related to the problems of grading and product design. The diversity in quality of many agricultural products has often been considered a positivebenefit to consumers and/or producers. Diverse qualities are thought to match the supposedly diverse preference maps of consumers, so that community satisfaction exceeds that which would prevail with a uniform quality. When the proper grading system is instituted, it will accurately reflect what consumers want and are willing to pay for, and it will return to producers the highest marginal value for their turkeys. 2 The effectiveness of price appeals for turkeys classified under our present system as "inferior" and "superior" has diminished due to today's technological knowledge, production and distribution methods in turkey production which have reduced the cost between grades to an insignificant level. Poultry meat grades were designed primarily to facilitate whole- sale trade, and recently increasing interest has been directed toward having grades more accurately reflect consumer preferences. This has led to a debate within the turkey industry as to the desirability of ‘mandatory grade labeling as well as the advisability of changing the present grade labeling system. Any grade labeling system faces the problem.of defining and measuring quality. In order for grades to be effective in differentiating quality for consumers they must be based upon those qualities that consumers look for in purchasing turkeys. This brings up the question of quality and how it can be defined. For the purpose of this experiment, quality is defined as that combi- nation of attributes that have significance in determining the degree of acceptability of a product by a consumer. Using this definition the question then arises, who should determine quality and how it should be labeled? Based on the above definition of quality, the author suggests that the consumer is the one who should determine quality, the degree of acceptability and how such quality should be labeled. Many human variables enter into the behavioral and preferences picture. Examples of these variables are age, education, sex, occupa- tion, nationality, environment and income. Exogenous factors such as seasonality of product, packaging and advertising also must be con- sidered in the study of consumer behavior. 3 "In so far as economic and marketing research is concerned, the important question is not to uncover the inner motivations for par- ticular actions but rather to identify motivations on a level where remedial steps can be carried out", Ferber and Verdoorn, 1962 (9). Mass production and marketing dictate that collective preference pat- terns be evaluated; therefore, the author is not particularly inter- ested in individual grade preferences pggdgg, but in the total market demand which is made up of a.mass of individual preferences. Purposes of this study were: to evaluate turkey grade labels and their quality meaning to the consumer; to determine the effect of price differentials on preference; how much of a price differential consumers would pay; and, at what price differential a decline in preference would occur due to price. Earlier work of Mills,'g£‘§1. 1960 (17, 18); Makens, gg‘gl. (14) had shown that "Prime", "Choice" and "Good" are preferred by consumers to letter grades A, B and C as grade labels for turkeys. This study has enlarged upon their work via the route of re- tail turkey sales in various market areas. With this in mind, this experiment was designed to test the following hypotheses: (l) The consumer will purchase turkeys based primarily upon appearance and labeled with the grade label which has the highest perceived quality meaning to her; (2) A portion of the consumers will pay a higher price for turkeys labeled "Prime", "Choice" and "Good" than they will for turkeys labeled Grade A or Grade B. REVIEW OF LITERATURE Development of Grading Brunk and Darrah, 1955 (6)1/say the following about the develop- 'ment of grades for agricultural products: "Many of the state and federal grades (there are now over several hundred in effect) were developed from grades originally established by the trade. The Chicago Board of Trade established the grading of grain in this country about one hundred years ago. Until grading was set up, buyers or buyers' agents inspected the grain in person. When this became cumbersome and impractical, some sellers began to sub- mit written statements guaranteeing their product. When the Chicago Board of Trade employed grain inspectors, other ’ markets soon followed suit. In 1871, the State of Illinois established the first governmental grading and inspection service, but as other states followed suit, the wide differ- ences in their standards did little to alleviate the diffi- culties inherent in trade-created grades. Even so, it was not until 1916 that the federal government assumed the responsibility of grading and inspecting grain. "Even before Congress passed the Grain Standards Act, there had been debate in the House of Representatives con- cerning the establishment of an inspection service for fruits and vegetables but not until August, 1917, did Congress appropriate funds for this purpose. Before that, some fruits and vegetables were sold under brand names but the great bulk was offered on the market unclassified to local buyers who, in turn, shipped to terminal markets. Under this system it was common practice for receivers to do much repacking and sorting at the terminal market in order to be able to meet trade requirements. "It was not until 1923 that the government undertook the grading of meat, but there was not as great a demand for _ meat grades as for grades of other products because a large proportion of the meat products were processed in relatively few packing plants with well-established brand names. It might also be argued that the qualities of meat, although highly variable, are not as subject to day-to-day ravages of weather as the qualities of fruits and vegetables. Feeding practices and processing methods make a certain degree of quality control of meat possible. Even to this day a large proportion of meat is sold under packer brand rather than U. 8. grade unless it is sold under price control laws, which often require federal grading. In 1941, slightly over 1] pages 277-279. 5 1% billion pounds of meat and meat products were federally graded. From 1943 through 1945, when price control legis- 1ation made such grading mandatory, about 12 billion pounds were graded annually. By 1950 the volume of graded meat and meat products had decreased to about 2.7 billion pounds. "For many years, the principal grades for eggs were established by the trade, which resulted in a wide variety of grade terms and grade interpretations. A definite effort was made in 1922 to develOp and to encourage the use of a federal grading system for eggs. Although federal grades have never been in wide use, except for government pur- chasing during and immediately after World War II, they have been used quite generally by state governments and the trade as the basis for state and private grading systems. In more recent years, consumer grades for eggs have been adopted as a result of the enactment of laws in many states requiring the grading of eggs sold at retail. Federal standards and grades for live and dressed poultry were first proposed in 1928. Few packers have adopted U. S. grades; they use, instead, special grade names which have been advertised over a long period of time. "It is difficult to determine the origin of grading and standardization, but the practice undoubtedly extends back many centuries. There was little need for grading or standardization as long as the products which people con- sumed were produced nearby or at home. Products purchased in the local market could be selected or rejected on sight, so that, in effect, the grading was done by consumers. But specialization in agriculture made it increasingly difficult for producers to sell directly to consumers, with the result that there was an ever-increasing demand for effective grading systems. The step from private market brands to federal grading was, for many products, a real milestone in the history of marketing." Early History of Standards Kohls, 1955 (11)g/has this to say about the early history of stand- ards. "Our grading standards have evolved hand in hand with the commercialization of agriculture. Some have been developed by trade groups and then have been formalized by federal and state agencies. Others have come about through the results of research directed toward the establishment of standards. Practically all have been changed with the 'passing years as weak points became evident. .2/ pages 140-141. 6 "In most instances, pressure developed for reform of the grading system from within the trade itself. Trade groups and organizations attempted to systematize nomenclature and grades. Generally, however, real permanent progress was not made until the federal government stepped in to coordinate the efforts to improve the grading system. In 1907, Congress appropriated funds to study federal standardization. The passage of the Cotton Futures Act in 1914 and the Grain Standards Act of 1916 initiated a series of laws which have gradually broadened the area of federal responsibility in promulgating uniform standards." Baker, 1954 (4)2/wrote that, "The official United States standards for market classes and grades of carcass beef were set in 1926. Grade standards have been extended since then to include the feeder and slaughter classes of most species of market livestock and meat derived therefrom. All the grade standards have re- mained practically unchanged since their origin, except for two changes. "The beef standards, which previously had provided for grading each class of beef (steer, heifer, cow, stag and bull) on a separate standard, were modified in 1939 to pro- vide for the grading of all steer, heifer and cow beef on a single standard. "The beef standards were again modified in 1950, when the Prime and Choice grades were combined as a single grade Prime, the Good grade became Choice and the beef from the younger animals in the upper part of the Commercial grade was designated Good. Corresponding changes were made at the same time in the standards for slaughter cattle. In 1951 similar changes were made in both carcass and slaughter standards for veal and calf and for lamb and mutton. Stand- ards were set in 1952 for slaughter barrows and gilts and their carcasses." Q] pages 160-161. 7 The USDA classification and grading schedule for slaughter live- stock and their carcasses are as follows: Class or Kind Subclass Grades Beef Calf Veal Lamb (Steer, heifer, cow . . . Prime, Choice, Good, Commercial, Utility, . . . . . . . . Cutter, Canner. (Bull and stag . . . . . Choice, Good, Commercial, Utility, Cutter, Canner. . . . . . . . . . . . . . . . . . . . . . Prime, Choice, Good, Commercial, Utility, Cull. O O O O O O O O O O O O O O O I O O O O 0 same as calf. . O O O O O O O O O O I O O C O C O O O 0 Prime , ChOice , GOOd , Utility, Cull. Yearling mutton . . . . . . . . . . . . . . . Same as lamb. Mutton . . . . . . . . . . . . . . . . . . . . Choice, Good, Utility, Cull. Pork . . . . . . . . Barrows and gilts . . . Choice No. 1, Choice No. 2, Choice No. 3, Medium, and Cull. 1Cows are not eligible for the Prime grade. Poultry Legislation and Grade Development According to the United States Department of Agriculture, 1956 (2y&{ "There are many Federal, State and municipal laws and regulations affecting grading and marketing. Only those that apply most directly or indirectly to grading poultry will be described or listed here. "The Federal Food, Drug and Cosmetic Act as amended, and regulations for its enforcement by the Food and Drug Administration of the U. S. Department of Health, Education and Welfare are extremely important. This act was designed to prohibit the movement in interstate commerce of adulterated and misbranded food, drugs, devices and cosmetics, and for other purposes. ‘3/ pages 30-31. 8 "The Packers and Stockyards Act of 1921 was enacted to regulate interstate and foreign commerce in livestock, live- stock products, dairy products, poultry, poultry products, and eggs, and for other purposes. It was amended in 1926, 1938, 1939, and 1942. Whereas the Food and Drug Act applies mostly to processed poultry in its effect on the public health, the Packers and Stockyards Act of 1921 is designed primarily to protect producers and consumers against various unfair, deceptive, and fraudulent practices and devices in interstate and foreign commerce. The regulations under the Packers and Stockyards Act are enforced by the U. S. Depart- ment of Agriculture. _- "There is an increasing trend toward the development of state, county, and municipal laws governing the processing and marketing of poultry. The primary purpose of these laws is to protect the consuming public. They may serve also to encourage consumption of poultry providing the provisions of the laws are sound and they are adequately enforced. There are many differences in state, county, and municipal laws at present. Since poultry like other farm products move in interstate commerce, it would be desirable if the provisions of the laws and regulations were uniform. "From an analysis of the laws which affect poultry grading and marketing the various state, county, and municipal laws may be grouped into the following categories: 1. Food and Drug laws are patterned after the Federal Food, Drug and Cosmetic Act. In addition to the general provisions of the food and drug laws, some states and local governments have certain sections which apply specifically to market poultry. 2. ‘Licensing laws require poultry dealers to obtain a license before engaging in buying and selling poultry. Some require a different type of license for poultry-processing plants or slaughterhouses. 3. Standardization laws generally authorize the establishment and maintenance of standards and grades for farm poultry. A few of the states have established standards and grades for some or all classes of poultry. In most instances, standards and grades are voluntary and are usually patterned after Federal standards and grades. In several states, grading and grade labeling of poultry offered in retail channels are compulsory. 9 4. Laws pertaining to proof of ownership of poultry in transit. In some states, it is illegal to transport poultry unless it is accompanied by a bill of sale, or a memorandum signed by the vendor. Producers transporting their own poultry are usually exempt. Legally authorized common carriers are exempt in most cases. This is primarily to guard against thieving. 5. Laws pertaining to limitations on hours of business that make it unlawful to sell or offer for sale during certain periods of the day or on certain days. 6. Inspection laws. A few of the laws require in- spection before poultry may be diSplayed, delivered, or offered for sale in certain forms. Other laws provide for the establishment of voluntary inspection service. A number of counties and cities require post-mortem examination during processing. 7. Laws affecting the sale of imported poultry pertain to the licensing of buyers, sellers, and processors of poultry outside of the state, county or city jurisdiction, regarding sanitation." / Benjamin and Pierce, 1937 (5)2 describe early turkey grading in the following manner. "Ordinarily turkeys are grouped in three grades, according to quality. As quoted in The Producers' Price-Current, the New York market recognizes fancy, choice and medium grades and most trading is done under these nomenclatures (young hens, young toms, old hens and old toms). There are no authori- tative specifications describing these grades as such, but they are approximately similar to and somewhat higher than the grades tentatively proposed by the United States Government under the terms "prime", "choice" and "commercial" as described in Table XXVII. 2] pages 283-284. 10 Table XXVII Tentative Specifications for U. 3. Standards and Grades of Dressed Turkeys Quality Specifications for Individual Birds -- Young Hens U. 8. Special "AA" U. S. Prime "A" U. S. Choice "B" Young, fine-grained, soft-meated female bird with broad full-fleshed breast, and with entire carcass fully covered with fat. Must be well bled, well dressed, and practically free from pin feathers, and have empty crop. No flesh bruises allowed; and only very slight skin abrasions, bruises or discolorations permitted, none of which shall be on the breast. Slightly dented breast bones (not to exceed % in. in depth) permitted but no crooked breast or other deformities allowed. A broken or disjointed wing above the wing tip, or a broken or disjointed leg, not permitted. Must be dry-picked or semi-scalded and must be dry packed. Young soft-meated female bird, with well-fleshed breast, and with entire carcass well covered with fat. Must be well bled, well dressed, with breast practically free of pin feathers and only a few scattered pin feathers over remainder of carcass. Crop must be empty. Only very slight flesh or skin bruises, abraisions, or dis- colorations permitted, with breast practically free of such defects. Slightly dented breast bones (not to exceed % in.) permitted, but no crooked breasts that would interfere with the slicing of the meat, or other deformities allowed. Broken wings above the wing tips or broken legs not permitted. A disjointed leg or wing permitted if only slightly bruised. Birds with crops properly removed and sewn up may be included in this grade. Must be dry-picked or semi-scalded and must be dry packed. Young female bird, with fairly well-fleshed breast, and with carcass fairly well covered with fat. Must be fairly well bled and dressed, and may show scattered pin feathers over the entire carcass. Crops may con- tain not more than 4 ounces of feed. Slight flesh or skin bruises permitted, but not more than three such defects on any bird. Slight skin abrasions or dis- colorations permitted. Abrasions or tears more than 3 inches in diameter not allowed, unless properly sewn up. Dented or slightly crooked breast bones or other slight deformities permitted. One broken wing or one broken leg permitted if bone does not protrude through the flesh and if not showing excessive bruise or blood clot. 11 U. S. Young female bird which may be poorly fleshed and Commercial with carcass poorly covered with fat. May show "C" evidence of poor bleeding and have numerous pin feathers over the entire carcass. Skin abrasions and discolorations permitted. Hunch back or other deformities allowed if birds are fairly well fleshed. Birds badly bruised so as to make any appreciable part of the carcass inedible not per- mitted. Birds showing emaciation or external evidence of disease or other condition which would render them unwholesome or unfit for human food not permitted." Chicken grades likewise were classified in the 1930's as Benjamin and Pierce, 1937 (STQ/list the following as U. S. Grades for dressed broilers, fryers, roasters and fowl: ”U. 8. Special or U. S. Grade AA; U. S. Prime or U. S. Grade A; U. S. Choice or U. S. Grade B; U. S. Com- mercial or U. S. Grade C." In 1953 the Poultry Division, Agricultural Marketing Service, USDA, changed the turkey grades and grade standards. Marsden and Martin, 1955 (15yz/describe in detail the 1953 grade standards and state that, "Briefly, U. 8. Grades are now three, formerly four, given in descending order of quality: U. 8. "AA" formerly Special (now discontinued) U. 8. Grade A formerly Prime U. 8. Grade B formerly Choice U. S. Grade C formerly Commercial" ‘g/ pages 166-167. 1/ page 378 . 12 The present specifications for standards of quality for individual carcasses of ready to cook turkeys are provided by the United States Department of Agriculture, Agricultural Marketing Service, 1960 (3). Summary of Specifications for Standards of Quality for Individual Carcasses of Ready-To-Cook Turkeys and Geese (Minimum Requirements and Maximum Defects Permitted) Factor A Quality B Quality C Quality CONFORMATION: Normal Practically normal Abnormal Breastbone Slight curve or Dented, curved, Seriously dent slightly crooked crooked Back Normal (except a Moderately Seriously slight curve) crooked crooked Legs and Wings Normal Moderately Misshapen Melissa. FLESHING: Well fleshed, Fairly well flesh- Poorly moderately long ed on breast and fleshed and rounded legs breast FAT COVERING: Well covered con- Sufficient fat on Lacking in sidering class breast and legs fat cover- to prevent a dis- ing over tinct appearance all parts of flesh through of carcass skin PINFEATHERS: Breast Breast and Legs Elsewhere and Legs Elsewhere Nonprotruding Pract. Pract. Few Few pins and hair free free scat- scat- Scattering tered tered grgtruding pins Free Eggs Frag Free Free CUTS, TEARS and MISSING smzl Free " g: 6" No my; 3 vgscmourgons : 2 2" g" 3;" it" 1.. W 13 Factor A Quality, BgQuality C Quality Disjointed house 1 l or 2, if no No limit broken bones Broken bones None 1 Nonprotruding No limit Missing parts Wing tips and Wing tips, 2nd Wing tips tail wing joint, and wings and tail tail FREEZER BURN: Few small (%" Moderate-dried Numerous diameter) areas not in pockmarks pockmarks excess of %" and large in diameter dried areas 1Total aggregate area of flesh exposed by all cuts, tears, and missing skin. 2Flesh bruises and discolorations such as "blue back" not permitted on breast and legs of A Quality birds. Not more than one-half of total aggregate area of discoloration may be due to flesh bruises or "blue back" (when permitted), and skin bruises in any combination. 3No limit on size and number of areas of discoloration and flesh bruises if such areas do not render any part of the carcass unfit for food. Grade Labeling The trend in Federal grading has been toward more grading. Each year a higher percentage of the turkey crop is Federally graded than was graded the year before. Effective January 1, 1959 all poultry processing plants concerned with interstate commerce were required to operate under compulsory Federal inspection regulations, and all poultry (turkeys) processed therein and packaged for consumers was required to be labeled by age (young, yearling, mature or old). This law has inspired much debate about the future of compulsory grading of turkeys. Much of this debate evolves around consumer grades and what grading system should be used. Since the topic is still controversial, it is fitting that the subject of grade labeling be reviewed. What is 14 grade labeling? Indeed, what is grading? What standards will be used? For whom will the standards be applied? For whose benefit? Shall the present system of A-B-C grading be used or is there another system more suitable for product differentiation for the consumer? These are but a few of the myriad questions which could be asked in analyzing the question of grade labeling for turkeys. Waugh, 1955 (21)§/stated that, "Grading has been promoted by producers and traders, and largely because they stood to gain by it; but grades must rest solidly on consumer's_preferences or on basic utility to consumers if they are to be effective. (under- lining added) Consumers will not pay more for one grade than another if it makes no difference to them which grade they buy. Furthermore, the fundamental economic justification of grades likewise is that they afford a means for consumers to register their preferences more accurately and more effec- tively, so that, if the grading system is carried all the way back to the producers, consumers are better able to encourage the production of the grades they prefer and to discourage production of the less desirable grades. ",..It's the consumers who determine the effectiveness of the grades set up. The grades established have been effective in prOportion as they have reflected real differ- ences in consumer's preferences." In pointing out problems relating to consumer grades, Brunk and Darrah, 1955 (6)2/stated, "All too frequently, consumer grades are based on the opinions of technical workers concerning what consumers should want rather than on actual measurement of quality factors that are meaningful to consumers. A pen that will write under water is a great technical achievement, but there are probably few people in the market seeking a pen with that particular quality. _8_/ page 3040 12/ page 288. 15 "Many of the opponents of consumer grades maintain that such grading is often misleading. To be helpful, it must be based on the ultimate use to be made of the product, a use that cannot be known ahead of time. Grade C tomatoes, for example, may be as satisfactory for use in some cooking as Grade A tomatoes, but the housewife buying them does not know this because she is not acquainted with the standards used to determine grade. All she knows is that Grade C tomatoes are not as good as Grade A. Precisely what "good" means in this context she does not know. Consequently, in some cases the lower grade designation results in unjustified discrimi- nation." (underlining added; In discussing the effectiveness of letter grades in communicating / the image of quality to the consumer, Kohls, 1955 (ll)l9 wrote that, "The "A-B-C" labels have the benefit of simplicity. But they will be useful only if the attributes of each grade are known by the purchaser. The Opponents of this pr0posal point out that "C" quality products may be the best product for some uses. But the average consumer will not know this and will discriminate against the lower grades as products not useful for consumption. For example, both "A" and ”B" eggs are adequate for table use. The ”C" grade eggs may actually be better for some cooking uses, such as when whipped egg whites are wanted. All are nutritious and clean products. But it is maintained that the terms "B" and "C" discriminate against their use because of the consumer's association of these terms with unusable products." Determination of Standards 11/ Kohls, 1955 (ll)—' writing about the objective of "ideal standards" states that, lr—I \ IH \ "The principal objective of an ideal standard should be to aid the consumer in telling the producer what he considers desirable in a product for the particular use to be made of it. Ideal standards set up a chain of information between the consumer and the producer. "Grading of agricultural products is a method of differ- entiation of the products to meet the various desires of the consumers. We have learned that consumers are characterized by a large range of incomes and preferences. Some desire and can pay for the highest quality. Others must be satis- fied with a somewhat lower quality at a lower price. A grading system attempts to differentiate the product in such a way that the various consumers are tapped for all that they page 154. page 142-143. 16 will pay. Or from the consumer's point of view, a grading system attempts to aid the consumer in obtaining the particu- lar product he desires. The best grading system will be such that it will move the greatest total amount into con- sumption and secure the greatest total price for that amount. The purpose of grading is not to assure the marketing of only top quality products. Those who conceive a grading system as a vehicle for the elimination of lower quality products are ignoring the wide range of consumer preferences and uses which exist. ”The major problem is developing grading standards then becomes one of determining differences in products which are economically significant and then developing methods for measuring these differences in the products. Only differ- ences which users are willing to pay for are pertinent to the establishment of grades. Whims and desires not backed by the willingness to pay are not justifications for grade differentiation. Neither are the opinions of a few "experts" of what should be desired workable foundations for grades." "The claims made for A-B-C grading and labeling are certainly specific in their language and positive. They may be summed up as follqws, as far as possible in the pro- ponents' own words (1)—— : l. A-B-C grading is a scientifically reliable and objective method measuring quality. Grade A repre- sents the finest obtainable. 2. It is based on consumer preferences. 3. A grade letter on a label is a simple, truthful and concise statement of quality and provides a common language for buyer and seller since grades can be carried through from producer to consumer. 4. A-B-C labeling is a practical and honest way of selling; its application can be made uniform by proper training of graders. 5. A-B-C labeling reduces uncertainty in buying; grade letters are impossible to misunderstand; buying by grade is simple and accurate and requires no technical knowledge. It throws an x-ray of know- ledge into the darkness surrounding consumer buying. 12/ pages 30-31. l7 6. It enables the consumer to compare prices and values. 7. It gives the consumer protection and a guarantee of minimum quality." Quality grading, a form of standardization, is done to reduce un- 13/ certainty and waste in buying and selling. Fairoletti, 1939 (8)- states that, ards "Devising a set of grades for a given foodstuff is an orderly and generally accepted method of accounting for sig- nificant differences in the characteristics that make up quality. Like the dollar and the ounce, a set of grades is a unit of account in exchange transactions, but it is applicable only to the commodity defined. "These defined standards of quality are not made up of exact, unchanging, or even easily measurable factors. For example, changes in methods of production or distribution, or the introduction of new varieties of the product, or even changes in the techniques of measuring the quality factors themselves, may require changes in the standards. Conse- quently, in different food trades or industries, there are found different degrees of development and acceptability of quality standards." / Waugh, 1953 (21)"1‘it in discussing the establishment of grade stand- stated, "One of the first problems encountered in the estab- lishment of standard grades is that of locating boundaries between grades. These take the form or provisions in the 'specifications' for the several grades. When such specifi- cations are changed, there usually follows a change in the proportions which graders will place in the different grades affected. ....Presumably grading is done to maximize returns to sellers. It does this by dividing_given products into 'grades' on the basis of attributes which buyers of different classes consider significant. ....Just where the boundaries between grades should be placed willgthen depend upon the degree to which the various users will pay premiums for certain qualities rather than substitute adjacent qualities within the ranges available." (underlining added), page 307. page 307. 18 Many who criticize the various grading systems may not be too familiar with the grade standards of the system. Brunk and Darrah, 1955 1 (6T'2/state that, "One of the greatest criticisms of grading systems is that frequently the standards developed have little or no relation to either price or use. Although the preferences of buyers should be reflected in grades, there is evidence that these are sometimes ignored. Perhaps the most serious short- ,coming of present-day grading is that the selegtign pf gualities for consideratigg and the determinatipn of theig respective_importance are not based on scientifically ‘gpasured tests of buyer preference." (underlining added). Some guides toward a better understanding of grade labeling (l)l§/ are, "The scoring systems used for many products should be re- examined to determine to what extent each score takes account of qualities important to consumers and whether weights have a reasonable relationship to consumer preferences. If important qualities have been omitted from the scoring, rea- sons for such omission should be examined. If objective tests for an important quality are lacking, renewed effort should be made to develop a satisfactory test. It may be that attention should be given to scoring products for certain nutritional values." I In discussing the potential effect of research on grade standards, Fairoletti, 1939 (8)ll/stated that, "Research in consumer preferences may discover that for some foodstuffs, at least, the grade standards used in a trade or industry are not so closely related to these prefer- ences as they might be for Optimum consumer satisfaction. If such discoveries are made, a reappraisal of the grade stand- ards will be in order. Until then, however, it may be assumed that there is a real and significant relationship between a set of grade standards used in the trade and consumer prefer- ences. Consequently, if it is technically ossible to measure quality differences with a degree 0 objectivity acceptable to the trade, in all probability this objectivity can be transmitted to consumer grade labels." ‘12/ page 282. lg/ pages 45-46. ‘11/ pages 367-368. l9 Permeating much of the literature on consumer grades is the explicit statement that consumer grades should be based upon what the consumer looks for in a product. To what extent she will pay a higher price for a grade if such a grade is preferred, has not been fully explored. Neither has consumer preference for grade labels been fully explored. Mills, 25.31., 1960 (16) conducted work in an attempt to discover if the consumer preferred the A-B-C system of grade labeling or if another system embracing the words "Prime", "Choice", and "Good" would be more preferable. They used the Detroit Preference Panel (12) to secure consumer Opinions, attitudes and preferences for turkeys labeled with various grade labels and prices. A variety of qualities was used. Preferences during a series of repeated tests were in favor of turkeys labeled with the word grades. With no price differentials, preference for the word grade labels was overwhelming. When prices were increased 2 cents per pound, preferences actually increased in most cases. Their study showed that price was a nexus of turkey quality and cannot be over- looked Or ignored in evaluating preferences for turkeys. Preference for turkeys appears to be directly influenced by the psychological sig- nificance of the grade label and price. Throughout the study, the panel members were apparently little concerned with minor defects. Discolorations due to moderate size flesh bruises and lack of flesh were the two main defects which consumers considered when selecting a turkey. The rejection of turkeys with these defects was consistent even when the preferred grade labels were used. Many consumers demon- strated that they did not like and would not buy turkeys labeled Grade B, but, they did like and would buy these same turkeys when they 20 were labeled "Good" or "Choice”. Expressed preference in a panel is one thing, but expressed prefer- ence at the retail sales counter is still another. Mills,lg£.§l., 1961 (17) conducted retail sales tests in Lansing, Michigan supermarkets in order to determine the validity of the panel results. They reported that the sales tests confirmed the consumers preference (panel results) for the grade labels ”Prime" and "Choice" to Grade A. Further, about 69% of total sales were turkeys labeled with word grade labels, and 25% of total sales were at a 2 cents per pound higher price. They con- cluded that the panel results were valid. Makens,.g§.§1., 1960 (13) reported that consumers said that they would be willing to pay an additional amount for prOperly graded and labeled birds. Personal observation was a major determinant in the selection of turkeys and consumers preferred grade labels "Prime", "Choice", and "Good" to letter grade labels A-B-C. Swanson and Hess, 1961 (19) reported that "in addition to size, the most important features considered in the individual turkey when the consumer makes a purchase were 'plumpness' and skin color". In the same study they stated, "of the 21 percent who have purchased a parts- missing turkey and the same percentage who have purchased skin tears, all but 13 and 10 percent, respectively, would buy parts missing and skin tear turkeys again. Chief reasons were low price and the feeligg that those turkeys were just as tasteful and of the same Quality as other turkeys," (underlining addedx No other research work was found pertaining to word versus letter grades on turkeys. PROCEDURE Sales Experiments In order to enlarge upon earlier work and in an attempt to arrive at a better understanding of price influence on purchases and the effect of "Prime" and "Choice" as grade designations on purchases of turkey hens over a wider geographical and a more representative cross section of the American consumer market, a market research project was undertaken. Two primary investigations made were: (1) How much of a price dif- ference per pound the consumer would pay for turkeys labeled "Prime" and/or "Choice", (2) Consumer preference for turkey grade labels "Prime", "Choice", and "Good” contrasted with Grade A and brand name (Grade B). To eliminate time and spatial factor influences, various combina- tions of grade designations and prices were tested simultaneously within a single market area and/or between market areas. USDA Grade A turkeys were used as controls, i.e., results were compared to the movement of Grade A turkeys at the prevailing retail market price. In most stores Grade A turkeys were sold in competition with tur- keys labeled either "Prime" or "Choice", Figure 1. In some Detroit stores, turkeys labeled "Choice" were used as controls and "Prime" as the experimental classification. In other Detroit stores, turkeys labeled "Good" were the controls and turkeys labeled "Choice" were the experimental, Figure 2. In still other Detroit stores, turkeys with a brand name (no grade label) were the controls, and turkeys labeled "Good" were the experimental classification. 21 22 Figure 1. Illustration Showing Turkeys as labeled for Sales Test. (Charlotte, Lansing and Grand Rapids). srosfiooufisvs!‘ CHOICE Tbnxavs Figure 2. Illustration Showing Turkeys as Labeled for Sales Test (Detroit). 24 Point of sale posters were used to call attention to the selection of turkeys available under different grade labels and at different prices, Figure 2. NO attempt was made to deceive the customer; on the contrary, the meat market managers were cautioned against this. Any advertising that was done by the store management calling at- tention to the turkey sale was the advertising of Grade A turkeys at the regular market price. NO mention of the test was made in any advertise- ment, or to the customer, unless the customer inquired as to the differ- ence between the turkeys. In this case the meat market manager was in- structed to tell the customer that the store was cooperating with the hniversity on a market research project in which they were testing a proposed new grading system for turkeys. He was cautioned to be perfectly honest in his answer to the customer and to preserve the customer's good will toward the store. Insofar as possible, equal numbers of turkeys labeled with the dif- ferent grade labels being compared were maintained in the frozen food dis- play area. This made it possible for the customer to have a variety of turkeys from which to make his or her selection. In addition, a variety of weights, within the 10-14 pound range being tested, was maintained insofar as possible. Only selections actually made by the customer were recorded. That is, no telephone orders were counted in the results (many such orders were noted in Lansing), and no instances in which the customer was aided in his or her selection by any employee of the store were counted in the results. The meat market managers were asked to report to the investi- gator any comments favorable or unfavorable regarding the study so that consumer reactions and attitudes could be noted. 25 The main reason for selecting the August-October sales period was the desire to secure information about off-season sales in response to the stimuli of prices, and the grade labels "Prime" and "Choice". It was believed, however, that little sales volume would be Obtained dur- ing this period, and this shbsequently proved to be the case. November is a heavy turkey sales month and was naturally included. The Easter season is a time when many turkeys are purchased; therefore, this period was also selected in which to conduct the studies. Some researchers in consumer marketing studies begin with a zero price differential and increase prices until they can observe a resist- ance to increasing prices. In this experiment, all price differentials to be investigated were used simultaneously. In this manner both resist- ance to price increases and the effect of decreasing prices could be Ob- served. Even-numbered price differentials were used in order to follow the usual merchandising practice of off cent pricing (29¢, 43¢, $1.69, etc.). Price differentials of 2e, 4e, and 6c per pound were used. In all cases the differential was 22223 the retail market price prevailing at the time the tests were conducted in a particular market area. Where Grade A turkeys were used as the controls, they were priced at the pre- vailing retail market price and turkeys labeled "Prime" or "Choice" were priced at specified differentials. Price differentials and treatments in each store were changed each Mbnday morning during the test. For the most part the base price was con- stant during the test; however, in some instances prices changed due to competitive forces, and the price also changed after Thanksgiving. This caused no problems since the main price factor under investigation was 26 the price differential. Price levels, while they would probably alter sales figures, were of secondary importance in this study. For this experiment, grade labels and price differentials were es- sentially the main factors under investigation. This is not meant to imply that other factors are of no importance but only that no attempt was made to control these factors. Since turkeys are sold under diverse conditions it was believed that a more representative and accurate ap- praisal Of grade label and price influence could be Obtained only if these two variables were controlled. An attempt was made, however, to secure a cross section of the shoppers within each market area. This was done in the selection of the stores to be included in the test. Purposive or judgment sampling techniques were used in selecting the stores because it was believed that a better representation of the city could be Obtained in this man- ner. In each market area, the management of the supermarket chain was asked to select the stores in his organization that would be the most representative of the income, ethnic groups, and educational levels within the city. In conducting market research, one of the main factors involved is finding management that is willing to cooperate with the investi- gator in conducting marketing tests. This was one Of the factors in- fluencing the selection of the cities in which to conduct the tests. In addition, the cities selected were convenient to the investigator, thus reducing the cost of the project, and the cities were judged to be representative of the geographical area in which they were located. 27 Charlotte, North Carolina, was judged to be representative of the Southeast; Grand Rapids, Michigan, of the Northern‘Mid-western area. Lansing had been used by Mills, 55 31., 1961 (17), and was used in or- der to obtain a comparison of sales between 1960 and 1962. Detroit, Michigan, was judged representative Of the large metropolitan areas of the United States with a diversity of educational and income levels and ethnic groups. In all test areas, the turkeys labeled "Prime" and "Choice" were selected for appearance (conformation, flesh, finish, serious defects); that is, these were the best turkeys in the USDA Grade A shipment. The author selected the birds used in the Charlotte, North Carolina, study at the processing plant prior to their being frozen. In the Lansing and Grand Rapids, Michigan, studies, the selections were made within the retail stores by the meat market manager after he was instructed by the author in how the selection should be made. However, due to the small number of turkeys available for the study, quality differences were minor. There was more uniformity within the "Choice" and "Prime" classifications where the birds were selected in the processing plant. In the Detroit tests, however, supplies of turkeys were not large enough to make ade- quate selections; therefore, essentially no separation was possible and the "Prime", "Choice", and Grade A turkeys were of comparable quality. In the Charlotte, North Carolina; Lansing and Grand Rapids, Mich- igan, tests, the USDA Grade A label was not covered, but was left plainly visible on the package. Grade identification of "Prime" and "Choice" turkeys was accomplished by means of a shipping tag wired on to the posterior end of the package around the metal clip used to seal the pack- 28 GRADE A PRIME CHOICE Figure 3. Illustration of Labels Used in Charlotte, Grand Rapids and Lansing. Figure A. Illustration of Store Label Attached to Grade Labels. 29 age. The grade label was stamped on this tag, Figure 3. On the other side of this tag the store label showing the weight, price per pound, and total price was attached. Thus, the customer was forced to look at the label in order to know the weight, price per pound, and total price before making his or her selection, Figure 4. In the Charlotte, North Carolina, test six stores were available from a local chain Of supermarkets. The design for the experiment was set up as a Latin Square with six stores involved in a six-weeks test. In this manner all three price treatments were used in each store, and "Choice" and "Prime" were used with each price differential. Two six-week periods were set as a test during the period August 5, 1962, through Oc- tober 15, 1962. Three stores were used in the test in Grand Rapids, Michigan. These stores were a part of a small local supermarket chain. In this market area price differentials Of 4 cents and 6 cents per pound were used. Grade A turkeys were sold in competition with either "Prime" or "Choice" in each store. The treatments and price differentials were changed weekly. The sales were conducted for a four-week period from November 5, 1962, through December 1, 1962. The market test in Lansing, Michigan, was conducted in cooperation 'with a local chain of supermarkets, in four stores, for a four-week per- iod from November 5, 1962, through December 1, 1962. As in the other sales tests, prices and treatments in each store were changed weekly. Grade A turkeys were sold in competition with either "Prime" or "Choice" turkeys. The customer had a choice Of a Grade A turkey at the regular ‘retail price or a "Prime" or a "Choice" turkey at the specified price differential above Grade A prices. 30 In Detroit, fourteen stores from a national chain of supermarkets were made available to the investigator. The study was thus more val- uable since both large and small supermarkets were used and the reac- tions to the study were available from the respective managements. Again, Grade A turkeys were used as the control birds with turkeys labeled "Choice" and "Prime” priced at 4 cents and 6 cents per pound above the prevailing retail market price. In some Detroit stores, turkeys labeled "Choice" were used as controls and "Prime" as the experimental birds. In others, Grade B turkeys labeled as "Good" were the contrdls at the prevailing retail price for Grade A turkeys, and Grade A turkeys labeled as "Choice" were the experimental birds. In still other stores, turkeys with a Brand Name (no grade label) were the controls, again at prevailing Grade A prices, and turkeys labeled "Good" were the experimental birds (in this instance, both were USDA Grade B turkeys packed under the processor's brand reserved for B grade turkeys). In the Detroit test, the USDA Grade A label was covered on the experimental turkeys with a white, rectangular pressure sensitive adhesive label. On this label the word "Prime”, "Choice", or "Good" was printed in blue ink as near to the shade used in the USDA grade label as possible. See Figure 5 for sample of labels. No additional label was attached to the USDA Grade A turkeys to signify the grade. The regular supermarket label was used to inform the customer as to the weight, price per pound, and total price of the turkey. This label was attached directly to the carcass near the grade label. 31 Ptime Clio/ca Good Pigtu'e 5. Illustration of Turkeys Labeled in Detroit Sales and actual labels used. 32 .In summary, the sales experiments were designed as follows: Number Time City of in Price Grade Label Stores Weeks Differentials Variables Charlotte 6 12 2c, 4c, 6¢ Grade A, Prime, Choice Grand Rapids 3 4 4c, 6¢ Grade A, Prime, Choice Lansing 4 4 4¢, 6¢ Grade A, Prime, Choice Detroit 14 2 4¢, 6¢ Grade A, Prime, Choice Good, brand name (no grade label) Prices and labels were changed weekly, i.e., store no. 1 would sell turkeys labeled Grade A.@ 35¢, and "Choice" @ 39¢ for one week. The next week the same store wOuld offer Grade A turkeys @ 35¢ and "Prime" @ 41¢. Each week changes were made in like manner until the end of the test period. A The sales data were analyzed by the Chi Square method and the "t" test. The Chi Square method was used to test the hypothesis that the sales ratio of Grade A to "Prime" and/or "Choice" would be 2:1. The 2:1 sales ratio was used to determine the theoretical sales for the pur- pose Of analyzing the sales data. Further analysis by Chi Square using a calculated sales distribution frequency was made, Dixon and Massey, 1957 (7). The "t" test was used to determine confidence intervals for the sales proportions. This method was used also to determine sales differ- ences between cities and between stores within cities as well as to de- termine which of these differences were greatest. The "t" test was used also to determine confidence intervals for the difference in sales pro- 33 portions between cities and between stores within cities, Dixon and Mas- sey, 1957 (7). Preference Panel Studies NO investigation had been found relative to the influence of word grade designations on turkey parts and cooked turkey; therefore, it was decided to test the influence of "Prime”, "Choice", "Good", and Grade A as grade labels on turkey parts, rolls, and cooked sliced rolls. The Detroit Preference Panel was reported by Mills, 35 31., 1960 (16), to be a valid means of measuring preference for turkeys. There- fore, it was decided to utilize this Panel for this portion Of the study. In this experiment there were 146 panel members (male and female) who came to Wayne State University and evaluated the turkey parts as to their acceptability or ranked them in order of preference. About one- half of the panel members made their rankings in the afternoon, and the remaining members made their rankings in the evening (12). Typewriter symbols %, &, #, *, () were used to identify the turkeys to eliminate any possible meaning which might have been associated with letters or numbers if they had been used in identifying the products be- ing ranked. These symbols together with the word or letter grade were attached to the package by means Of pressure-sensitive adhesive labels. The consumer ranked the products in order of preference from most to least acceptable by merely putting the number 1 by the symbol on the panel report card that corresponded to his or her first preference, and the number 2 by the symbol that corresponded to his or her second choice, etc. A tie for first or any other preference could occur. Four turkeys were selected as a source of the parts which were eval- uated by the Detroit Consumer Preference Panel. Three Of these turkeys 34 were of Grade A quality and one of Grade B quality. The bird selected for the "Prime" parts was of excellent conformation, fleshing, and fin- ish with an absence of defects. The two birds selected as a source of parts for the "Choice" and Grade A parts were of good conformation, well fleshed, but were lacking in the degree of finish evident on the "Prime" turkey. No defects were present. The bird selected for the "Good" parts was of Grade B quality and somewhat lacking in flesh, finish, and conformation, but with no serious defects. It was slightly red due probably to improper bleeding. These turkeys were divided into parts as follows: whole breasts, thighs and legs; packaged in Cryovac bags; sealed with a metal clip and heat treated. Then, they were placed in the liquid freezing facilities Of Michigan State University until the desired color was attained, after which they were held in the air blast freezer at -20° F. until ready to use. The rolls used in this study were purchased from the manufacturer, roasted in the original package material to an internal temperature of. 160°F. and transferred hot to a plain Cryovac bag and sealed with a metal clip. The heat from the rolls was sufficient to shrink the Cryovac to make a neat package. The rolls were held at a temperature of 35°F. until evaluated by the panel. For this series Of tests, no prices were placed on the products in the afternoon; however, for the evening panel, prices were included as a part of the labeling. The prevailing retail market price was used as the base price and price differentials were used. The parts and products labeled "Good" were priced at the prevailing retail price; Grade A and 35 "Choice" were 2 cents per pound higher in price, and "Prime" was 4 cents per pound higher in price. In Series I, cooked turkey rolls, prices for the evening panel were: Grade A and "Choice" , $1.55 per pound; and "Prime", $1.59 per pound, Figure 6. In Series II, cooked sliced turkey roll, the "Prime", "Choice", and Grade A packages consisted Of four uniform slices that were placed on clear white paper board, wrapped in cellophane and heat sealed. For the "Good" package, slices that had become separated during slicing were used. All packages had a net weight of 6 ounces and were made from the same roll. For the evening panel, prices were: "Good", 75¢ per package; Grade A and "Choice", 77¢ per package; and "Prime", 79¢ per package, Figure 7. Series III consisted of raw turkey breasts. In the afternoon no prices were used. For the evening panel the turkey breasts were priced as follows: "Prime", 79¢ per pound; Grade A and "Choice", 77¢ per pound; and "Good", 75¢ per pound, Figure 8. Series IV consisted Of raw turkey thighs. In the afternoon no prices were used. Prices attached in the evening were: "Prime", 79¢ per pound; Grade A and "Choice", 77¢ per pound; and "Good", 75¢ per pound, Figure 9. Series V was made up of raw turkey legs (drumsticks). In the af- ternoon, preference was tested with no price differentials. In the ev- ening, the following prices were added: ~"Prime", 69¢ per pound; Grade A and "Choice", 67¢ per pound; and "Good", 65¢ per pound, Figure 10. 36 Figure 6. Turkey Rolls as Labeled for Panel Test. Figure 7. Cooked, Sliced Turkey Roll as Labeled for Panel Test. 37 Figure 8. Raw Turkey Breasts as Labeled for Panel Test. 38 Figure 9. Raw Turkey Thighs as Labeled for Panel Test. Figure 10. Raw Turkey Legs as Labeled for Panel Test. 39 For analyzing the Detroit Preference Panel data, the coefficient of concordance was used Kendall, 1955 (10). This method of analysis measures the correlation (O to 1 scale) or communality of preference of a group of respondents. The larger the coefficient, the probability that the rankings could have happened by chance is less. Significance by this method does not necessarily mean that there is, or is not, a significant difference between any two classes in a series. The co- efficients were tested for significance by the Chi Square method. (See appendix page 98 for calculation example.) RESULTS Sales Tests In order to evaluate consumer preference for the words "Prime", "Choice", and "Good" as grade labels for turkeys, and to investigate how much of a price difference per pound the consumer would pay for turkeys labeled "Prime", "Choice", and "Good", a series of sales tests were conducted in Charlotte, North Carolina; Detroit, Grand Rapids, and Lansing, Michigan. See Tables 23, 24, and 25, appendix pages 93-95 for sales prices in each store, by cities. Table 1. Summary, by Cities, comparing Sales of Grade A Turkeys with Turkeys Labeled "Prime" £42 per Pound Price Differentialz. Per cent of Number Sold Total by Total Sales City Grade A Prime City Grade A Prime Detroit 23 22 45 51 49 Grand Rapids 107 80 187 57 43 Lansing 91 34 125 73 27 Totals 221 136 357 62 38 Table 1 shows the summary of the sales in three Michigan cities where Grade A turkeys were sold in competition with turkeys labeled "Prime" and priced 4c per pound above the Grade A retail price. The data show a sales ratio of Grade A to "Prime" of 62:38. When these data were analyzed by the Chi Square method using the theoretical sales ratio of Grade A to "Prime" of 2:1, the difference is significant at the 1 per cent level. When analyzed by the same method, using a calculated sales frequency, the results are also highly significant (appendix page 97 ). 4O 41 Using the "t" test, confidence intervals were obtained to measure the range of Grade A sales, expressed in percentage of total sales, which could be expected from repeated experiments. The 95 per cent confidence intervals are 57 to 67 per cent, and 55 to 74 per cent are the 99 per cent confidence intervals (see appendix page 98 for calculations). Based upon these data the expected sales of turkeys labeled "Prime" will be from 26 to 45 per cent of the total 99 per cent of the time, and from 33 to 43 per cent of the total 95 per cent of the time. These data show that if this experiment were repeated many times, one can be confident that the sales of turkeys labeled "Prime" and priced at 4 cents per pound higher than Grade A turkeys would consti- tute one-third of the total sales 95 per cent of the time. Further, one can be confident that the sales of turkeys similarly labeled will constitute 26 per cent of total sales 99 per cent of the time. From the data obtained in these tests, it is obvious that about one-third of the hen turkeys may be sold at a higher price when the grade designations are changed to those used in this study and at the prive level prevailing during this study. Table 2 shows the sales summary comparing Grade A turkeys with turkeys labeled "Prime" and priced at 6c per pound higher than the Grade A retail price. 42 Table 2. Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled "Prime" 56% per Pound Price Differential). Per Cent 0 Number Sold Total by Total Sales City Grade A Prime City, Grade A Prime Detroit 34 16 50 68 32 Grand Rapids 192 112 304 63 37 Lansing 72 _ 23 95 76 24 Totals 298 151 449 66 34 These data show a sales ratio of Grade A to "Prime" of 66:34. Analysis of these data by the Chi Square method revealed a significant difference from the theoretical sales ratio (2 Grade A:1 "Prime") at the 5 per cent level but not at the l per cent level. This also was true when the calculated sales ratio was used as the theoretical expec- tation. (See appendix page 99 for calculations.) Confidence intervals were obtained for Grade A sales by means of the "t" test. The 95 per cent confidence intervals were 62 to 70 per cent, and the 99 per cent confidence intervals are 60 to 72 per cent (see appendix page 98 for calculations). A price resistance was noted in the above analysis. When the price differential was increased from 4 to 6c per pound, there was a decline in percentages of sales in the "Prime" grade designation. This also was shown by the statistical analysis as measured by the failure of the sales to be significantly different from the theoretical ratio of 2:1 at the 1 per cent level, whereas at 4c per pound the sales were differ- ent. This difference was due to an excess of "Prime" turkey sales over what was expected. 43 Although a price differential of 6c per pound was sufficiently broad to cause an incipient decline in preference for the "Prime" labeled turkeys, these data show that from 25 to 35 per cent of the hen tur- keys were sold for 6¢ per pound more than the price of the Grade A labeled turkeys. Table 3 shows the summary of sales when Grade A turkeys were sold in competition with turkeys labeled "Choice" and priced at a 4c per pound price differential. Table 3. Summary, by Cities, Comparing Sales of Grade A Turkeys with Turkeys Labeled "Choice" (49 per pound Price Differential). Percentage of Number Sold Total Sold Total Sales City Grade A Choice by City Grade A Choice Detroit 37 29 66 56 44 Grand Rapids 134 140 274 49 51 Lansing 114 37 151 75 25 Totals 285 206 491 58 42 These data show a sales ratio of Grade A to "Choice" of 58:42. When these data were analyzed by the Chi Square method, a highly sig- nificant difference was found from both the theoretical ratio of 2:1 as well as the calculated ratio. This difference was due to the ex- cess sales of "Choice" turkeys over the expected sales (see appendix page 100 for calculations). Confidence intervals were also obtained from these data. The 95 per cent confidence intervals for the range of Grade A sales are 54 to 62 per cent, and the 99 per cent confidence intervals are 52 to 64 per cent (see appendix page 102 for calculations). 44 These data show that when Grade A turkeys were sold in competition with "Choice" turkeys (selected for quality) priced at 4c per pound above Grade A turkeys, 35 per cent of the hen turkeys now being sold as Grade A were sold at a higher price when labeled "Choice". Further, one can be confident, based upon these data, that this will occur 99 per cent of the time as shown by the 99 per cent confidence limits. Table 4 shows the summary of the sales when Grade A turkeys were sold in competition with turkeys labeled "Choice" and priced at a 6¢ per pound price differential. Table 4. Summary, by Cities, Comparing Sales of Grade A Turkeys with , Turkeys Labeled "Choice" 56¢ per Pound Price Differential). Percentage of _ Number Sold Total Sold Total Sales City Grade A Choice by City Grade A Choice Detroit 35 12 47 75 25 Grand Rapids 64 75 139 46 54 Lansing 136 41 177 77 23 Totals 235 128 363 65 35 4 These data show a total sales ratio of Grade A to "Choice" of 65:35. When these data were analyzed by the Chi Square method, the differences were found to be highly significant from the theoretical ratio of 2:1. This significance is due largely to city differences between Grand Rapids and Lansing. Sales of Grade A turkeys (percentage) in Grand Rapids were consistently lower than sales in Lansing. (See appendix page 101 for calculations.) Confidence intervals of these data for Grade A turkey sales com- pared with "Choice" sales are almost identical with the confidence in- 45 tervals comparing Grade A with "Prime" at the same price differential. The 95 per cent confidence intervals are 60 to 70 per cent and the 99 per cent confidence intervals are 59 to 71 per cent. This means that, based upon these test data, one can expect the percentage of Grade A sales to fall within these limits with the appropriate confidence un- der the conditions of this experiment. Again, these data show that with a price increase from 4c per pound to 6c per pound above Grade A prices, the percentage sales of turkeys labeled "Choice" declined, which demonstrates the effect of wide price differentials. However, even at 6c per pound above Grade A prices, these data show significant differences in the sales when compared to the expected ratio of 2:1. When "Prime" was compared with Grade A, there was a highly significant difference between cities which indi- cates that the sales ratio of 2:1 should not be expected in all cities, but it may be expected from aggregate sales from several cities. Table 5 shows the summary of Grade A turkey sales compared with turkeys labeled "Prime" and "Choice" at a 4¢ per pound price differ- ential. Table 5. Sales Summary, by Cities of Grade A Turkeys with Combined Sales of "Prime" and "Choice" Turkeys (4c Per Pound Price Differential). Percentage of Number Sold Total Sales Prime & Total Sold Prime & City Grade A Choice by City Grade A Choice Detroit 60 51 111 54 46 Grand Rapids 241 220 461 52 48 Lansing 205 71 276 74 26 Totals 506 342 848 60 40 46 These data show highly significant differences from the theoreti- cal sales ratio of Grade A to "Prime" and "Choice" of 2:1. Based on these results one-third of the turkeys now being marketed as Grade A may be sold for at least 4¢ per pound higher price if they are labeled "Prime" or "Choice". Table 6 shows the summary of Grade A turkey sales compared with turkeys labeled "Prime" and "Choice" at 6c per pound price differential. Table 6. Sales Summary, by Cities, of Grade A Turkeys with Combined Sales of Prime and Choice Turkeys (6c per Pound Price Dif- ferential). fl Percentage of Number Sold Total Sales Prime & Total Sold Prime & City Grade A Choice by City Grade A Choice Detroit 69 28 97 61 39 Grand Rapids 256 187 443 58 ‘42 Lansing 208 64 272 76 24 Totals 553 279 812 66 34 These data show highly significant differences from the theoreti- cal sales ratio of Grade A to "Prime" and "Choice" of 2:1 by cities, but the total sales ratio is almost identical to the theoretical ratio. Based upon these data, approximately one-third of the turkeys now be- ing sold as Grade A.may be labeled "Prime" or "Choice" and sold for 6c per pound higher price. The results (total sales) are not significantly different from the theoretical ratio. Table 7 shows the combined sales data. 47 Table 7. Summar of Turke Sales Price Differentials Combined . Percentage of Number Sold Total Sales Prime & Total Sold Prime & City Grade A Choice by City Grade A Choice Detroit 129 79 208 62 38 Grand Rapids 497 407 904 55 45 Lansing 413 135 548 75 25 Totals 1039 621 1660 63 37 The "t" test was used to compute 95 per cent confidence intervals from these data. Based upon these data, it is expected that from 35 to 39 per cent of total sales would be "Prime" and "Choice" in repeated tests. The 99 per cent confidence intervals are 33 to 40 per cent. Earlier work with consumer panels showed that there was no differ- ence between the preference for turkeys labeled with the words "Prime" and "Choice" at the 5 per cent level of significance. To further test this observation in actual sales tests, two Detroit stores were selected. Turkeys labeled "Choice" were the controls and turkeys labeled "Prime" carried the higher price. The sales results are shown in Table 8. Table 8. Turkey Sales Comparing "Choice" and "Prime" as Grade Labels.* Grade Price Number Grade Price Number Label Per Pound Sold Label Per Pound Sold Choice .39 11 Prime .43 7 Choice .39 35 Prime .45 2 *All turkeys Grade A quality. These data show that when the price differential was 4c per pound there was no difference between sales, from a statistical point of view (5 per cent level). However, from the point of view of 48 this experiment, they show that with product differentiation by grade labels "Prime" and "Choice", returns can be increased. Further, the data support earlier work showing that there is little or no difference in preference between the two terms as grade labels for turkeys. At a 6c price differential, these data show that there was a de- cided shift in preference in favor of turkeys labeled "Choice". No doubt much of this shift is a direct influence of the 6¢ price differ- ence. These data show that after the 4¢ price differential is passed, an incipient decline in consumer preference occurs due to price. To test the movement of Grade B turkeys labeled with the word "Good" when sold in competition with Grade A turkeys labeled with the word "Choice", two stores were selected. The results are shown in Table 9. Table 9. Turkey Sales Comparing "Good" and "Choice" as Grade Labels.* _ m Price Price Grade per Number Grade Per Number Label Pound _ Sold L Label Pound Sold Good - .38 38 Choice .42 '26 Good .39 2 Choice .45 12 * "Good" turkeys were B quality; "Choice" turkeys were A quality. These data in Table 9 show that at a 4¢ per pound price differen- tial, turkeys labeled "Good" outsold the turkeys labeled "Choice". When the price differential was 6c per pound, turkeys labeled "Choice" outsold the turkeys labeled "Good". These data show that if Grade B turkeys are labeled as "Good“ they will sell in volume at prevailing Grade A retail prices, and many turkeys presently referred to as Grade A turkeys might sell for 4c per pound more when they are labeled "Choice" 49 It should be pointed out that the Grade B turkeys in this test had better conformation, fleshing, and finish than did many of the Grade A turkeys that were labeled "Choice". The exact cause of down-grading could not be determined except in a few instances. Some had missing skin and tears on the legs and backs. Others were reddish in appear- ance. The consumers demonstrated in this test that "red" turkeys are highly undesirable and unacceptable. Sales of turkeys labeled "Good" practically ceased when a box of red turkeys was displayed. When these were finally sold, sales of turkeys labeled "Good" returned to a high level. This unacceptable condition probably accounts for some of the "Choice" sales. These data were subjected to statistical analysis at a theoretical ratio of 1:1 and were found to be statistically different at the 5 per cent level. This difference probably is due to the small number of "Good" sold at 39¢ in comparison to the larger number of "Choice" sold at 45¢. One store had some turkey fryer-roasters and toms in stock, all of which were USDA Grade B. It was decided to test the preference relationship between Brand Name (no grade label) and the grade label "Good". The results are shown in Table 10. 50 Table 10. Turkey Sales Comparing "Good" with No Grade Label - Grade Grade Per Number Grade per Number Label Pound Sold Label Pound Sold Brand Name# .53 6 Good# .57 6 Brand Name* .39 4 Good* .43 6 #Beltsville Small White Turkeys 6-8 pounds. *Tom Turkeys 16-18 pounds. These very limited data indicate that consumer preference fOr word grade labels as consumer grade designations is probably just as appli- cable to fryer-roaster and toms as to hens. The most important thing in this table is the price level of the fryer-roasters and toms. These data would seem to indicate that the general price level is of lesser importance in determining whether consumers will pay a price differ- ential for turkeys labeled with preferred grade labels than formerly thought. One will note that the toms were priced the same as Grade A hens in other stores. No statistical analysis was made of the results in this table. The volume of sales in Charlotte, North Carolina, was insufficient for statistical analysis. When sales at all three price differentials (2c, 4e, 6e) were combined, 62.5 per cent of the total sales were tur- keys labeled with either "Prime" or "Choice" as compared to turkeys labeled Grade A during the first six-weeks period. During the second six-weeks period sales of turkeys labeled "Prime" and "Choice" were 51 per cent of the total. 51 No conclusions can be drawn from these data. At best it is inferred that "Prime" and "Choice" are preferred by many consumers to Grade A as grade labels. Word grade labels appear to have more meaning as quality communicators than does Grade A. It is also inferred from these data that one might expect consumers in the South to respond to the grade labels "Prime" and "Choice" as consumers in other sections. While the percentage of sales of "Prime" and "Choice" were higher than in other areas tested in this study, it is doubtful if any parti- cular significance should be attached to this since the number of sales was small. If sales volume had been large, it is doubtful if Grade A turkeys would have sold in sufficient numbers to cause any change in the total findings of this study. Preference Panel Studies In order to obtain information on consumer preference for raw tur- key parts and cooked turkey rolls (whole and sliced) labeled with the grade label "Prime", "Choice", or "Good", a series of tests were con- ducted with the Detroit Preference Panel. For the afternoon section of the Panel, no price was used on the rolls. Grade A, "Prime", and "Choice" were used as the grade designations with each roll being as identical as possible except for the grade designation. For the evening section, prices were added to the grade label. A price of $1.59 was placed on the roll labeled "Prime", and $1.55 was placed on the Grade A and "Choice" rolls. The base price of $1.55 was selected, based upon the wholesale price of raw rolls at 75¢ per pound plus 80¢ per pound to compensate for weight loss and the service of cooking. No other change was made between the afternoon and evening. 52 Results of Series I are shown in Tables 11 and 12. Table 11. Influence of Grade Label on Preference for Cooked Turkey Rolls by the Detroit PreferencefPanel. Afternoon April! 1963. Z lst Rank Symbol Grade Label Place 1 2 3 Persons Z Prime 65.0 39 12 6 * Grade A 20.0 12 21 24 () Choice 15.0 9 23 25 Table 12. Influence of Grade Label and Price on Preference for Cooked Turkey Rolls by the Detroit Preference Panel. Evening, April, 1963. Price 1% Grade per lst Rank 4§ymbol Label Pound Place 1 2 3 Persons % Prime $1.59 45.3 39 18 27 * Grade A 1.55 24.4 21 32 31 () Choice 1.55 30.2 26 37 20 The data in Table 11 show that 65 per cent of the respondents pre- ferred the roll with the "Prime" grade designation. This percentage declined, probably due to the influence of price, to 45 per cent in the evening section (Table 12). While the roll labeled "Choice" in the af- ternoon was the least preferred, when price was added it was preferred by 6 per cent more respondents than the roll labeled Grade A. In Series II, cooked, sliced turkey rolls were labeled with a variety of grade labels to observe indicated preference. In the afternoon no prices were used on the packages, while at the evening panel prices were added. No other changes were made. Results 53 are shown in Tables 13 and 14. Table 13. Influence of Grade Label on Preference for Cooked, Sliced and Packaged Turkey Roll by the Detroit Preference Panel. Afternoon, April, 1963. Z Grade lst Rank Symbol Label Place 1 2 3 4 Persons * Prime 32.8 20 14 13 10 # Grade A 16.4 10 13 23 11 Z Choice 36.1 22 21 10 4 () Good 14.7 9 13 8 27 The data in Table 13 shows little or no difference in preference between the packages labeled "Prime" and "Choice", and little or no difference in preference between the packages labeled Grade A and "Good" as per cent of first place. Table 14. Influence of Grade Label and Price on Preference for Cooked, Sliced and Packaged Turkey Roll by the Detroit Preference Wn=__ A-r111963~ s, Price Z ’ Rank Grade per lst Symbol Label Pkg. Place 1 2 3 4 Persons * Prime $.73 22.4 19 8 28 30 # Grade A .71 12.9 11 30 35 9 Z Choice .71 35.2 30 38 12 5 () Good .69 29.4 25 13 10 37 When.prices were added, as the data show in Table 14, indicated preference shifted somewhat. The package labeled "Choice" was then preferred by most respondents, with the package labeled "Good" in second place, and the package labeled Grade A was clearly preferred 54 by the least number of respondents. This emphasizes the fact that the quality of cooked, sliced turkey meat is difficult to distinguish. Products from deformed carcasses and carcasses lacking in flesh would appear to be acceptable in cooked form. It also appears that standards of quality for raw turkey meat are relatively useless as a measure of the quality of cooked turkey meat where little or no nutritional or tenderness differences exist, unless grades have a merchandising ad- vantage and do not needlessly downgrade the product. These data show that price was an influence in the purchase of cooked, sliced turkey meat, and that consumers had a tendency to be somewhat conservative when purchasing this product. However, the 48 per cent who indicated a willingness to buy the same meat at 2¢ per package higher price, and the 22 per cent who indicated a willing- ness to pay 4¢ per package more is important from the standpoint of profit maximization to the manufacturer. Seventy-one per cent indi- cated a willingness to pay more for the meat in the package of their preference. These data were subjected to statistical analysis (coef- ficient of concordance) which showed the coefficient's significance at the l per cent level. From time to time it may be profitable to offer cut-up turkey for sale. Several series were included to evaluate the relative preference for turkey parts labeled with word grades as compared to Grade A label. Table 15 shows the preference scores for raw turkey breasts (Ser- ies III). In this series, the turkeys were selected as to quality. The breast labeled "Prime" had better conformation and better finish than the breast labeled "Choice" and Grade A, both of which were of 55 comparable A quality. The breast labeled "Good" was from a Grade B turkey which was lacking in flesh and finish, and slightly red on the sides.' Novprocessing defects such asrpinafeathers, cuts or bruises were present on any of the breasts. As in previous tests, no price was attached to the packages in the afternoon, but was added for the evening panel. Results of this series are shown in Tables 15 and 16. Table 15. Influence of Grade Label on Preference for Raw Turkey Breast Persons () Grade A 11.7 7 15 25 10 Z Good 3.3 2 6 16 33 * Choice 33.3 20 22 12 3 & Prime 51.7 31 14 4 8 Table 16. Influence of Grade Label and Price on Preference for Raw Turkey Breast by the Detroit Preference Panel. Evening, April, 1963. J — Price Z Grade per lst Rank Symbol Label Pound Place 1 2 3 4 Persons () Grade A $.77 10.6 9 22 33 21 Z Good .75 15.3 13 15 14 43 * Choice .77 34.1 29 35 14 7 & Prime .79 40.0 34 15 24 12 These data in Table 15 show a clear preference for the breast lab- eled "Prime", and the breast labeled "Choice" preferred to the one labeled Grade A. These data in Table 16, with prices added, showed a 56 reduced preference for the "Prime" package, but it was still ranked first by most respondents. The breast labeled "Choice" gained in pre- ference score when prices were added. 7 Probably the most notable change occurred with respect to the breast labeled "Good". The gain in preference score of the breast labeled "Good" was almost exactly equal to the decline in preference score of the breast labeled "Prime". These data also show that price on turkey breasts was an influence on preference and that a high per- centage of the consumers indicated that they would pay more for turkey breasts when they were labeled with the word grade labels used in this study. Analysis of coefficients of concordance showed signifi- cance at the l per cent level. Series IV was included to evaluate the relative influence of word grade labels on preference for raw turkey thighs. The thighs were from the same respective birds as the breasts. In the afternoon no prices were attached. In the evening prices were included in the labeling. Results are shown in Tables 17 and 18. Table 17. Influence of Grade Label on Preference for Raw Turkey Thighs by the Detroit Preference Panel. Afternoon, April, 1963. 1.. Grade lst Rank Symbol _ Label Place 1 7 2 3 4 Persons # Prime 43.1 25 14 15 3 () Good 1.7 l l 5 50 Z Grade A 34.5 20 l8 l7 2 * Choice 20.7 12 24 20 l 57 Table 18. Influence of Grade Label and Price on Preference for Raw Turkey Thighs by the Detroit Preference Panel. Evening, April, 1963. Price 7Z7 Grade per lst Rank Symbol Label Pound Place 1 2 3 4 Persons # Prime $.79 32.2 28 16 26 14 () Good .75 11.5 10 8 ll 55 Z Grade A .77 35.6 31 25 24 4 * Choice .77 20.7 18 35 23 8 These data in Table 17 again show a preference for the thighs lab- eled "Prime". In this test, thighs labeled Grade A outranked thighs labeled "Choice" in number of first place rankings. These data in Table 18 show that when prices were added the thighs labeled Grade A received more first place rankings than did the thighs labeled "Prime". The thighs labeled "Choice" received the same percentage of first place rankings in the afternoon and evening. As in the case of the breasts, the preference for thighs labeled "Good" showed the greatest shift in relative preference. These data also show that a high percentage of the consumers indicated a willing- ness to pay more for turkey thighs when they were labeled with a word grade label. This further supports price influence in purchasing tur- key parts. Coefficients of concordance showed significance at the l per cent level. Tables 19 and 20 show the results of Series V when raw turkey legs were labeled with a variety of grade labels. In the afternoon panel, no prices were included in the labeling; however, as in previous series, 58 prices were added in the evening. The price addition was the only change. Table 19. Influence of Grade Label on Preference for Raw Turkey Legs by the Detroit Preference Panel. Afternoon, April, 1963. Z Grade lst Rank Symbol Label Place 1 2 3 4 Persons & Choice 25.0 14 27 10 4 * Good 5.4 3 7 11 34 # Prime 60.7 34 13 6 2 Z Grade A 8.9 5 12 27 ll Table 20. Influence of Grade Label and Price on Preference for Raw Turkey Legs by the Detroit Preference Panel. Evening, April, 1963. Price Z Grade per lst Rank Symbol Label Pound Place 1 2 3 4 Persons & Choice $.67 32.9 28 32 15 10 * Good .65 17.6 15 10 15 45 # Prime .69 31.8 27 25 19 14 Z Grade A .67 17.6 15 21 37 12 These data in Table 19 show a clear indication of preference for the legs labeled "Prime" and "Choice" in that order. When prices were added as shown in Table 20, these data show a marked change in prefer- ence rankings. There was no significant difference between the number ranking the legs labeled "Prime" and "Choice" as first place, and no difference in the per cent ranking legs labeled Grade A and "Good" as their first choice. These data show that with turkey legs a high per cent of the consumers indicated that they would pay more for turkey 59 legs when labeled with grade designations "Prime" and "Choice" rather than Grade A. They also show that with a small price differential the turkey legs from a Grade B turkey labeled as "Good" may sell about equally with turkey legs from Grade A turkeys labeled as Grade A. Coefficients of concordance showed significance at the l per cent level. DISCUSSION Sales Tests Mills, 35 $1., (1961) used Lansing stores in a preliminary sales study and reported that 25 percent of the turkeys of comparable quality could be labeled "Prime" or "Choice" and sold for at least 2 cents per pound higher price. This thesis supports that work, however, in this study the price differentials were increased to 4 and 6 cents per pound. The same turkey sales ratio as reported by Mills, 23 31., (1961) was observed. This similar finding in Lansing appears to be important because it indicates that the same results obtained in the other cities are likely to reoccur in repeated tests. It further increases the probability that the results of this study was valid. At the l per cent level of probability there is no difference in the ratio of turkey sales in Lansing in 1960 and in 1962. Combined turkey sales observed in this study differed from the theoretical sales ratio of 2 Grade A:l ”Prime" or "Choice" at the l per cent level of probability. This difference was due to the excess sales of turkeys labeled "Prime" and "Choice" over that which was expected. The differences in the sales ratio in the various cities used in this study were much greater than the differences between stores within cities as shown by the "t" value in the analysis, although some stores ‘within a city might have a noticably different sales ratio. When Grand Rapids and Detroit were tested against Lansing, the differences had "t" values of 8 and 7 respectively. This shows that the differences between these cities are highly significant and could not have happened by chance based upon the data obtained in the experiment. At the 99 60 61 per cent level of confidence, the "t" value would have been 2.58. When the difference between Detroit and Grand Rapids was analyzed, the "t" value was 1.87. Since a "t" value of 1.98 is the 95 per cent confidence level, there appears to be no significant difference in the sales ratio of Grade A turkeys to "Prime" or "Choice" in these cities. Store differences in Grand Rapids were analyzed and "t" values of 2.39, 2.53 and 0 were obtained. Comparing these values with the "t" values obtained for the analysis between Grand Rapids and Lansing, it is observed that the differences between stores within Grand Rapids are much smaller than the differences between the two cities. See appendix pages 104-108. Store differences in Lansing showed "t" values of 2.52, 0.67 and 2.998. Two of these comparisons are significantly different at the 5 per cent level (a ”t" value of 1.98), and the other is not. While there are some differences between stores within Lansing, this differ- ence is not as great as between Lansing and Grand Rapids. Store number 4 in Lansing was not tested for differences because of insuf- ficient sales to make the comparison meaningful. Although the sales per store in Detroit were too small for adequate statistical analysis, the "t" test for differences was used for indications as to the nature of the differences. These differences followed the same trend as those between stores in Lansing and Grand Rapids. The meaning of these differences is that one should not expect each Inarket area to behave in the same manner or to show the same sales ratio. However, one should expect, based upon these data, that the stores with- in each market area would experience quite similar sales ratios. The 62 aggregate sales from several cities, should follow a similar pattern as was discovered in this study. Two competitive price levels were involved in this experiment. In Grand Rapids, the prevailing retail Grade A price was 35 cents per pound. In Lansing, the prevailing retail price during the same period was 39 cents and in Detroit the base price level (Easter) was also 39 cents. Price differentials above these levels were added. The Detroit market appeared to be more sensitive to a price differ- ential of 6 cents than was either Grand Rapids or Lansing. In Detroit, the decline in purchases of turkeys labeled "Prime" or "Choice" at the 6 cents price differential averaged 18 per cent as a percentage of total sales. This could have been due to the actual price increase of the turkeys or it could have been due to the city-wide promotion of hams during the Easter season. Another factor that may have had some influence in this decline was the fact that one major food chain sold Grade A turkeys at 37 cents per pound although this would appear doubt- ful except in instances where the consumer shops strictly by adver- tised prices. The ratio of sales in Detroit at the 6 cents price differential was practically the same as in Lansing. In Grand Rapids and Lansing, there was little net change in the sales ratio when the price differential was increased from 4 cents to 6 cents per pound. The fact that there was some decline in purchases of the "Prime" and "Choice” labeled turkeys indicates that maximum sales at differentiated prices may occur at about a 4 cents per pound higher price than the next lower grade. Any increase above this there will perhaps be a decline in purchases. 63 Figures 11 and 12 illustrate the percentage sales of "Prime" and "Choice" labeled turkeys at the price differentials used in this study. For example, in Figure 11, when turkeys labeled Grade A and "Prime", with the "Prime" priced 4 cents per pound above the Grade A, the number of "Prime" sold was 39 per cent of the total. When the price differ- ential was 6 cents per pound, the number of "Prime” sold was 34 per cent. Figure 12 reflects the relationship when ”Choice" and Grade A turkeys were offered with the two price differentials. Lines d connect the two points that link the per cent of sales at two different prices during the study and lines d' and d” represent the range within which one might expect the per cent of sales to occur in repeated sales (95 per cent confidence interval@. By dropping a per- pendicular line from the confidence intervals to the base line, one can observe that at each price differential there is a possibility that at either price differential some overlapping of sales may occur. This overlap is represented by the letter 0 in each figure. 6¢ dl d d" 6‘: d: .d Ad" | I | ' 8's l | | ' C1 G i g g I I | I t3 9" I I I l f” “ | I S a: | . a) a) l I l I U «U .3 c l l I u m u 3.0 I l I ' 4c ' a. ' 3 n c J 4 -...~ .' {<— o a! I ‘ Tc»: I 30 40 45 30 40 45 Figure 11. Percentages of Figure 12. Percentages of Total "Prime” Turkey Sales. Total "Choice" Turkey Sales. 64 Figures 13 and 14 illustrate the movement of the price-quantity sold relationships (demand curve) as occurred in this study when the turkeys were differentiated by grade label and price. Line d represents a hypothetical demand curve based on two points of the price-quantity sold relationship if all turkeys had been sold as Grade A. This would represent the common trade practice today. Line d' represents the demand curve as a result of blending the prices at the two price levels in these sales tests. The blend price was obtained in the following manner: Grade A sales X Price X weight + Word Grade sales X Price X weight Total sales X weight 'o a . 8 .40 ~ ‘ ‘ d! m I 3 \ m .35 d g m - U s m U .30 m .3 H . . . . m 200 300 400 500 Quantity Sold Figure 13. Illustration of Demand Curve Movement (4 Cents Price Differential). vo g .45 o a. u m \ u ~_ - a -¢ 3 .35 d ~ 1 g m U 'L‘ m .30 200 300 400 500 Quantity Sold Figure 14. Illustration of Demand Curve Movement (6 Cents Price Differential). 65 By product differentiation and pricing, the demand curve was shifted to the right on the basis of this study. This is based on the assumption that all sales occurred in the same market. The assumption that Grand Rapids and Lansing are the same market appears to be a valid assumption since the sales ratios in Grand Rapids changed very little between the two price differentials. The Lansing sales ratio was practically identical at both price differentials. Grading in General In our competitive system, every turkey producer has the right to establish and market his own private brand and promote it in the market place to build consumer confidence in his brand. Conversely, his distributors or retailers may develOp their private brands from his turkeys. This freedom of producing, brand labeling, and marketing assures both producer's and consumer's freedom of choice. The confidence which the producer of turkeys places in the con- sumer makes it possible for him to plan his production in a more business like manner. It enables him to invest in capital improvements, to support research designed to improve quality and to reduce production and marketing costs, to enter into contractual arrangements with suppliers and processors so that they will be supplied with an adequate amount of turkeys, and to engage in advertising to keep consumers in- formed about turkeys. These activities which are designed to provide the consumer with high quality turkeys at reasonable prices help main- tain and promote greater confidence. With a system built upon mutual confidence, albeit highly competi- tive, safeguards against possible abuse of their freedom.must be built into the system. These safeguards have taken the form of laws and 66 regulations providing protection against fraud, adulterated food, unwholesome products, misrepresentation, and monopoly. Enforcement of these safeguards is a function of the State and Federal Governments. The Congress of the United States, in providing basic legislation, has been broad in its scope, constructive, and cognizant of the needs of an expanding industrial society and agriculture and the rising level of living. Under our free enterprise system, American agriculture has pro- vided the consumer with a prodigality of products resulting in an opulence of opportunities beyond anything the world has ever seen. In arriving at this affluent state, there has been such a proliferation of brands and products that the consumer frequently experiences diffi- culty in deciding what to buy, where to buy, and indeed understanding one's own needs in the complexity of a free society, as contrasted to the simplicity of living in a totalitarian state with its paucity of consumer goods. The art of spending money, when faced with so many alternatives, appears to have developed at a much slower pace than production and marketing technology. What confusion that exists in the minds of the consumer may be due in part to the variety of choices available, in part to the slowness of consumers to keep pace with developing technology and to learn their own needs and congruently to develop their buying skills. Competitive advertising may also contribute to the backward art of spending money. The amount of money that a farmer receives for his turkeys depends upon: (1) The economic level of society in general or the amount of money consumers have to spend, (2) The total volume of turkeys marketed, and (3) The cost of getting turkeys to market. Many feel that 67 production costs determine retail prices. Others, including the author, believe that prices are determined by consumers at the retail level. The power which the consumer wields at the retail supermarket shelf is the major key to price determination. Trade demand is derived from consumer demand; therefore, consumer preferences should play a vital role in determining effective consumer grades. To the extent that consumer grades reflect consumer preferences, then it might be said that the problem of developing consumer grades is one of harmonizing acceptable consumer quality with grades. The central problem in consumer grade labeling appears to be one of selecting the system of identifying quality, as the consumer perceives quality, and labeling it in the manner which has real significance to her as a quality label. One might well ask, what is the function of a grading system? Is the function that of describing and identifying quality to facilitate marketing? Or, is the function that of setting standards of quality to which the marketing system should adjust? It would appear that the grading system should be flexible and embrace as its main function that of facilitating marketing by differentiating the product according to factors of quality which the consumers look for in purchasing the product, and labeling these quality factors with such grade labels that adequately reflect acceptable quality and that have meaningfulness as quality labels. Grades, whatever they may be called, imply consumer evaluations of qualities that apparently give different degrees of satisfaction to the individual. It can be said from this study that many consumers are more tolerant of defects causing variations in turkey quality than many 68 quality-conscious experts think they are. It is apparent from this study (as Mills 35. $1., found earlier) that quality and value of a product are what the individual perceives as quality and value, nothing more. Tastes vary. So does the quality image of a product and the value generalizations attached to it. It would be presumptuous to inflict upon others one's own idea of quality and the value which should be attributed to it. For example, who is to say that a turkey which has been classified as B quality because of some missing skin, a broken bone, or one that is lacking in finish is not a high quality turkey? Is it any different nutritionally? Perhaps there might be a few less calories because of the lack of finish. There has been no difference reported in the nutritive value of turkeys that do not conform to the present standards of A quality and those that do. Yet, there are some who would say that such a turkey is of "inferior" quality and place a Grade B label on it. Certainly the consumer did not reject such turkeys in this study. In fact, when such Grade B turkeys were labeled "Good" and sold in compe- tition with Grade A turkeys labeled "Choice", Grade B turkeys outsold the Grade A's (Table 9). As Grade B's, they would not have sold because such a label would have "shouted" to the customer "inferior quality, do not buy!" This was shown to occur in research investigations by Mills, egggl., (1960). The Grade B turkeys used in the above test were actually superior to the Grade A turkeys $2 appearance in the eyes of the consumer as well as the author's. They (Grade B turkeys) had better conformation, were more meaty, and with a grade label which did not cause mental anguish, the consumers bought them and at no reduction in price! 69 Some questions that should be considered when grades are estab- lished are listed below: 10. 11. 12. In Can the factors be measured? Does each factor reflect quality as desired by consumers? Are the factors independent of personal preference? Do the factors approach a reasonably complete measure of quality? Does the emphasis given each factor accurately reflect that factor's importance to a majority of consumers? How definite are the differences in factors between grades? Are these differences clear and no more arbitrary than necessary? Are the tests for quality factors objective? Does the final grade differentiate the product in such a way that the consumer can tell the relative difference in how good the product is? Has the nutritional factor been included? Has the taste factor been included? Can the grade be easily understood and used by marketers? Present System the light of the foregoing grade criteria, one can examine the present A-B-C system. This system is analogous to a chain. It is a chain of technical procedures, and like the old proverb states, no chain can be any stronger than its weakest link. If one of the criteria for the grade standard is in doubt, then there may be doubt in all of them. If the combination of grade factors have little to do with con- sumer quality preference, if the factors add up such that a portion of the product is discriminated against unjustly, then the grading system 70 is in doubt. If one or more of each step is Open to doubt, the whole system is doubtful. It is difficult to visualize how a system.of doubts can be fused into a valid scientific grading system. From the results obtained in controlled experiments, the present letter grading system does not fulfill all the objectives of a sound grading system. It does not measure quality by the same standards as do consumers. The scoring system of tolerances do not combine in a way that accurately reflects consumer preference of the 1960's. It dis- criminates against a portion of the crop unjustly. It does not differ- entiate‘the product in such a way that the consumer can tell the relative difference of how good the product is. It does not take into account the increasing importance of nutrition as a quality determinant. The grade specifications appear to be more arbitrary than necessary. When the A-B-C system is analyzed in light of the proponent's claims presented on page 16, the answer comes out the same. Based upon the results of this study, the scientific reliability and objectiveness of the system as measuring quality is in doubt. Grade A, in the mind of 37 per cent of the consumers (Table 7) certainly does not represent the finest turkey available. If a letter grade label is simple, it is deceptively simple. There is a more pragmatic way of selling by grades that gives higher consumer satis- faction and higher producer returns. Letter grades can be misunder- stood. While it may give the consumer an Opportunity to compare prices, it is doubtful if it enables her to compare value. It does, however, give her a minimum.guarantee of quality. 71 The problem of grading and labeling for consumers requires an utterly different set of requirements and philOSOphy than grading and labeling for the wholesale trade. One often hears the terms grades, standards, and specifications used interchangeably. Although these terms are intimately related, they are different and have different functions. Grading is one thing, and "quality" grading for consumers is a different thing. Simply stated, grading is separating a mass of products into different classifications. Standards refer to the factors which make up quality. Specifications refer to exacting quality standards to which a product must conform. A quasi-simple word like quality may have one meaning, or it may have a dozen, or it may have none at all when applied to comparable grades for the purpose of comparison. It can be seriously questioned whether there is a need for whole- sale grades under today's turkey marketing system. There_i§ a need to have grades more nearly reflect consumers' opinions about turkey quality. Effective grading must be done on factors, but which factors are important? What is important about each factor, and to whom is it important? Is the factor related to value? Whose value? Is the per- ception of the grade label and the character of the product congruent? These questions cannot be avoided in grading and grade labeling. This study showed that the main factors consumers look for in purchasing turkeys are: appearance, meatiness, shape, wholesomeness, serious defects, and finish. In addition, consumers are now thinking in terms of flavor, tenderness, and nutritional value more than formerly. While no objective test for flavor and tenderness (prior to purchase) is available at the present time, they are important adjuncts to grades. 72 Many consumers dramatically demonstrated their preference for better quality differentiation as used in this study by paying premiums of 4 and 6 cents per pound for turkeys labeled with word grade labels. If an item is perceived to have no higher value than another item, or to give no more satisfaction, one does not pay more for the item. Con- versely, if one perceives an item to provide greater satisfaction than another, then that item is purchased even at a higher price. In this study, 37 per cent of the consumers paid the extra price to get turkeys labeled "Prime" or "Choice". The discernable differences in the grades and price would aid in adequately differentiating apparent quality. Both terms are merchandisable, and the turkeys so labeled would repre- sent acceptable consumer quality. Such differentiation makes possible legal price discrimination, that both increases total revenue to the industry and apparently increases consumer satisfaction. Since the "Choice” Grade would constitute over one-half of the crop, it is sug- gested that under the proposed system that the general level of competitive pricing would be the "Choice" Grade. The "Primen grade then could be priced 4 cents or more above the competitive grade. This study has shown that in the general price level of from 35 cents to 49 cents per pound, price premiums are possible when a grading and grade labeling system is used that reflects acceptable consumer quality. It appears to the author that letter grades do not adequately differentiate turkeys as to qualiux and therefore, are not truly repre- senting value as perceived by the consumer. For if such grade labels are accurately representing value, why then would 37 per cent Of the consumers in this study pay a higher price for turkeys? In some cases, ("Good" vs. "Choice") they even bought more Grade B turkeys than Grade A. 73 The answer is that the grade label was a positive quality image creator and communicator. Such word grade labels apparently are more meaning- ful to the consumer. They are better communicators of quality without connotations of inferiority as some letter grades are apt to be. It should be pointed out that while Grade B turkeys are being sold under private brands, they are not labeled as Grade B. This casts a reflection on the whole system. Why have a system if only one grade label from it will be used? Why is the industry not using the Grade B label? The answer to this one is reeking in simplicity. Grade B turkeys labeled as such will not sell! Why label a product as inferior when, in fact, it is not inferior in the eyes of the consumer. At present, a firm has to rely upon brand labeling to sell its Grade B turkeys. Many economists have engaged upon crusades to encourage producers to sell their turkeys by grades as a method by which they could receive a greater return for their turkeys. Under the present system of grading, this has met with little success. Among the reasons for this is the fact that the Grade B turkeys are discriminated against in price by from 3 to 5 cents per pound. Many of these Grade B birds are due to processing damages and no price consideration is usually given to the producer when this occurs. The procurement procedure of buying by flocks results in some producers being overpaid for their turkeys while other producers are underpaid. No satisfactory method has as yet been negotiated by which the farmer can be convinced that he is better off to sell on grade. If the prOposed grade labeling system is adopted, based upon con- sumer preferences as shown in this study, about 50 per cent of the 74 turkeys now being classified as Grade B can be sold at non-discounted prices. This will make it possible for farmers to be paid by grade and would result in increased returns to them. Under this system, only the lowest classification will be discounted in price. Buying by grade is the preferred way of paying the farmer for what he produces. The proposed system will make this possible without penalizing him unduly. Proposed System The author is in favor of grade labeling for all turkeys. He believes in product differentiation by quality factors, grades, grade labels, and price in order to sell turkeys for the most economic benefit to the producers and higher satisfaction for consumers. Product differentiation by grades and grade labels should be based upon current standards. The standards defining turkey quality in 1940 are archaic and unsuited to today's world. They do not represent the quality- conscious consumer's conception of quality and her willingness to pay. This willingness to pay results in part from having more discretionary income than in former years. Word grade labels more nearly represent the true character of the product. The author is not under the illusion that a system of grading and grade labeling embracing the words "Prime", ”Choice", and "Good" is the final answer to the problem of truly representing turkey quality, neither is he naive enough to believe that the A-B-C system is anywhere near today's requirements as a grade labeling system for turkeys. Few, if any, of the strongest advocates for the A-B-C grade labeling system will make claims for its perfection, and they have con- ceded that when a better system is available, which takes into account 75 consumer preferences in a more realistic way, then the system should be changed. Such a system is now available. The proposed system is believed to be a more objective way of correlating quality with con- sumers preferences. This proposed grading system is designed primarily to point out how turkeys can be differentiated into grades in line with expressed consumer acceptance, and with price discrimination with respect to grades as shown in this study. 76 PROPOSED MINIMUM GRADE STANDARDS FOR INDIVIDUAL, READY-TO-COOK HEN TURKEYS (Maximum Defects Permitted) Prime SYMMETRY - Natural Breast - slight curve or dent Back - practically straight Legs and Wings - natural FLESHING - Superior fleshing with a minimum breast width of 5 inches. FINISH - Well finished, no excess fat. DEFECTS - a. Pinfeathers l. Protruding - none. 2. Non-protruding - practically free. Cuts, tears and missing skin - none. Bruises - none. Other discolorations - slight skin abrasions may be present if not on breast. Other1 - None, except wing tip, removed to prevent flying, if properly healed 1No freezer burn permitted. No breeder hens permitted. 77 PROPOSED MINIMUM GRADE STANDARDS FOR INDIVIDUAL, READY-TO-COOK HEN TURKEYS (Maximum Defects Permitted) Choice SYMMETRY - a. Breast - Curved or dented, not to seriously detract from appearance. b. Back - Curved or dented, not to seriously detract from appearance. c. Legs and Wings 4 Slightly deformed, not to seriously detract from appearance. FLESHING - Well fleshed, minimum breast width 3.5 inches. FINISH - Well finished, may have slight excess fat. DEFECTS - a. Pinfeathers l. Protruding - None 2. Non-protruding - few, not to seriously detract from appearance. b. Cuts, tears and missing skin - Minor. 1. Breast - not to exceed 1 inch. 2. Legs - not to exceed 1 inch. 3. Elsewhere - not to exceed 4 inches if clean. c. Bruises - Minor 1. Breast - not to exceed 1 inch in diameter, if slight; % inch if dark. 2. Legs - same as breast. 3. Elsewhere - not to exceed 4 inches in aggregate if none are serious. d. Other discolorations - pigment discoloration permitted not to exceed 1 inch on breast and 2 inches elsewhere. e. Other1 - No broken bones permitted, except wing tips. Dislocated bones - 2 permitted if no serious dis- colorations are associated. Missing parts - none, except wing tip. 1Freezer burn - slight freezer burn permitted. 78 The "Prime" Grade would consist of about the top one-third of the total crop, and command a price of 4 cents per pound higher than the "Choice" Grade. At times of scarcity this price differential could be as high as 6 cents per pound, and the "Prime" Grade should still clear the market. The "Choice” Grade would consist of the lower two-thirds of the present Grade A turkeys and the top one-half of the present Grade B turkeys. The Specifications have been devised so as to approach this division. All other wholesome turkeys would go into the "Good" Grade. This grade, while no specifications are proposed, could be divided into two divisions. One division might be "Good" - Parts Missing, the other simply "Good". This might become necessary if the further processing industry could not absorb all of this grade. This would not appear to be a cause for unnecessary alarm in the future because of the develbping field of cooked turkey products. The United States Department of Agri- culture, Statistical Reporting Service, 1962 (18), in their report of poultry slaughtered under Federal inspection reported a total of 1,379,043,000 pounds of turkeys slaughtered, and 147,549,000 pounds inspected for canning and other processed foods. This shows that during 1962, 10.7 per cent of the total inSpected turkeys was used for canning and other processing. It is proposed that only the "Prime" and "Choice" Grades be sold at retail and the "Good" Grade go into further processing. This is in no way an attempt to push "low quality" turkeys onto the manufacturer. It is simply a suggestion as to the best potential market for the turkeys. Canned and other processed turkey products are the best available methods for utilizing the less acceptable turkeys for retailing. The nutrition, 79 flavor, and tenderness factors; cost, availability, and wholesomeness are the main criteria on which the quality for canned and other turkeys are based. Of course, meat yield is important, but this is offset by price considerations under the proposed plan. At present, many manufacturers probably avoid using Grade B turkeys because of the connotation of this grade in the minds of the consumer should the consumer discover that Grade B turkeys are being used in their products. This would not occur with the use of turkeys labeled "Good" because the consumer has demonstrated that this label is acceptable and does not have the inferior connotation as does Grade B. By restricting retail sales to "Prime" and "Choice", the consumer would have two levels with which to measure quality and value. Table 21 shows calculated gross returns to producers based upon the present grading system and the proposed system. The percentages of A, B, and C turkeys as shown in Table 21 are the approximate average breakdown by the USDAclgl The prices as shown form the basis for com- parison. The percentage of "Prime", "Choice”, and "Good" are based upon a projection of this study and with the price differential of 4 cents per pound for the top grade. The number in the "Choice" Grade was determined by taking the top 50 per cent of the Grade B turkeys and combining them with the bottom 66.7 per cent of the Grade A turkeys. The remainder of the crop would make up the "Good" Grade. 18] Lester Kilpatrick, Acting Chief, Standardization and Marketing Practices Branch, Poultry Division, Correspondence, April, 1962. 80 Table 21. Calculated Gross Returns to Producers for Two Grading Systems, Turkey Hens.* (Letters vs. Proposed) Price Per cent Revenue Grading Number Sold per of Million System Millions Pound Crop Dollars A 41 .24 82 118.08 B 6.5 .20 13 15.6 C 2.5 .15 5 4.5 Total 50 100 138.18 Prime 16.7 .28 33.3 56.112 Choice 27.55 .24 55.1 79.344 Good 5.75 .19 11.6 13.11 Total 50 100 148.566 Gain in Revenue 10.386 *Assumptions: 1. 50,000,000 turkey hens 2. 12 pounds average weight This table shows a gain in revenue to producers of $10.386 million (7.516 per cent increase) as a result of reorganizing the consumer grades of ready-to-cook turkey hens to more accurately reflect consumer preferences as expressed in this sales experiment. Much of this gain would be due to the non-discounting of the price paid for turkeys now classified as Grade B, when 50 per cent of these turkeys would go into the "Choice" Grade. Although under the pr0posed system of grading, the "Good" grade shows a large increase in numbers, this does not mean that the percentage of low quality turkeys will necessarily increase. It means that the quality in this lower grade has increased by virtue of the lower 50 per cent of the Grade B turkeys, being reclassified as "Good". Corres- pondingly, the price of the "Good" Grade has increased. 81 Similar results would be expected with toms although the percentages and prices per pound would be different. Table 22 shows the retailers calculated income from this study, comparing the returns actually received with the incomes that would have been received if all turkeys had been sold as Grade A at the pre- vailing retail price. Table 22. Calculated Returns to Retailers in This Study vs. All Turkeys Sold as Grade A.* Returns Returns Returns from from if City Grade A Prime and Choice All Grade A Gain Detroit $1052.70 $ 729.84 $1752.54 $ 30.00 Grand Rapids 2111.40 1962.12 3836.28 237.24 Lansing 1932.84 711.76 2564.64 79.96 Totals $5096.94 $3403.72 $8153.46 $347.20 *Assumptions - middle of the weight ranges sold. Of the 1660 turkeys sold, the retailers realized an income of $8500.66 compared with $8153.46 if all turkeys had been sold as Grade A. This gives a net gain of $347.20 (4.25 per cent) as a result of price and grade differentiation. During this sales test, in all instances where Grade B turkeys were sold either as "Good" or as private brands, the price was the prevailing retail Grade A prices. If this is a common practice among retailers, then the proposed grading system would in no way be short changing the consumer by either price or quality. Assuming that all retailers make no price difference between Grade Aland Grade B turkeys, there would be no price adjustment necessary by the consumers under the proposed system. 82 Since about one-half of the present Grade B turkeys would now be classified as "Good”, some of the "Good"Grade of turkeys may be sold at retail. If this were the case, then the consumer would be getting them at a reduced price from what she is now paying. However, it is proposed that only the two grades "Prime" and "Choice" be sold at retail and the "Good" Grade is proposed to be diverted into manufactured products such as pies, rolls, etc. Season turkey movement in this study (Thanksgiving, Easter and August-October) was similar with respect to the percentage of turkeys labeled "Prime" and "Choice" sold. Although the sales volume in Charlotte, North Carolina was insufficient for statistical analysis, it is doubtful if the sales of Grade A turkeys would have been suf- ficient to change the total picture. This particular phase (North Carolina) would indicate that the findings of this study would be valid in the South as in the Midwest. It would further indicate that similar results may be obtained in the off season and throughout the United States. These dianoetic extrapolations are predicated upon the assumption that consumers would react in the same manner to the stimuli of grade labels and price. Such extrapolations appear plausible since, in the marketing of products, merchandising techniques are essentially the same in all parts of the United States. The price differential should be used to ration the top birds in a particular crop. By this the author is referring to removing the top one-third of the crop from the bulk of the turkeys and increase the price, thereby, increasing total revenue from all birds. Total revenue will be increased if the trade will differentiate the product by the suggested grades and using comparable price differentials ranging near the 4 cents and 6 cents price per pound differentials used in this study. RECOMMENDATIONS FOR FURTHER RESEARCH It would appear that the next step in researching this particular type of study would be to develop a bag using the proposed grade labels, differentiating the grades at the processing plant and selling turkeys nationwide for a year. From the results obtained in this study, it would appear that regional studies would not be as desirable as nat- ional studies as a source of additional information. During this study it became apparent that one aspect of marketing heretofore considered by many to be solved has not as yet been perfec- ted. PACKAGING! In the opinion of the author this subject deserves immediate attention by the turkey industry. Many turkeys sold in this study would not meet the Grade A standards of quality due to excessive freezer burn and other damage; yet, they were sold under the Grade A label. A large percentage of the heat shrinkable bags used to pack- age the turkeys were coming apart at the front of the carcass so that a ring of freezer burn appeared all across the breast. Excessive free- zer burn also was very noticeable over the legs and wings. It could be Ehat the bags are not strong enough; it could be that too much vacuum was applied during packaging; or, it may be due to careless handling. Whatever is the problem, it is costing the turkey industry money. A missed sale of one turkey does not simply mean that there is one more turkey to sell. It means that this turkey that would have been sold and consumed, if not for this unsightly condition, may have resulted in more sales either to that particular customer or to his or her friends. Instead, this turkey is still in the grocer's free» zer discouraging others who might want a turkey from buying one. A missed 83 84 sale is expensive! Especially one lost from excessive freezer burn. Some of these turkeys are in such bad condition that it is amazing they sell at all. The selling job is not finished when the producer or processor gets his check for a shipment of turkeys. The sale is successfully completed only when the turkey is purchased by the consumer and it is satisfactory. Perhaps too often neglect in the market place has cost the turkey industry much more than is readily apparent to a cas- ual observer. At any rate, the problem of poor packaging is still with us where it does the most damage - the retail supermarket. The consumer is paying more and more attention to taste, fla- vor, and tenderness of turkeys. In addition, she is becoming more aware of the nutritional aspect of her meat purchases. It would appear to this author that more attention should be given to flavor than is being given to it at the present. Many consumers remarked during this sales test that some turkeys just did not have the fla- vor they should have. There may be many reasons why that particular turkey did not have the flavor it should, and one might be inclined to shrug it off as just one of those things. The turkey industry can- not afford to do this. It must be ever alert to the consumer's wishes if it is to prosper. Of what benefit is it to produce turkeys for 18 or 19 cents per pound if the consumers do not buy them? The basic purpose of turkey production is consumption. To lose sight of this fact is to make a grave mistake. Based upon the information obtained from this study, there appear to be more important things to worry 85 about than the incessant pursuit of lower production costs. The con- sumer appears to be more concerned with the eating qualities of tur- keys than she is with a few cents per pound lower price. SUMMARY AND CONCLUSIONS The purposes of this study were to determine consumer acceptance of and preference for grade labels and what were consumers' opinions regarding turkey quality as reflected by purchases. Two primary investigations were made: (1) How much of a price difference per pound above Grade A prices would the consumer pay for turkeys labeled "Prime" and/or "Choice"; (2) Do consumers prefer tur- key grade labels "Prime", "Choice", and "Good" to the labels Grade A and Brand Name (Grade B — no grade label). The experiment was designed to test the following hypotheses: (l) The consumer will purchase turkeys based primarily upon appear- ance and labeled with the grade label which has the highest perceived quality meaning to her; (2) A portion of the consumers will pay a higher price for turkeys labeled "Prime", "Choice", or "Good" than they will for turkeys of comparable quality labeled Grade A or Grade B. Retail sales were conducted in Charlotte, North Carolina, during August-October, 1962; in Grand Rapids and Lansing, Michigan, during November, 1962; and in Detroit, Michigan, during the Easter period, April, 1963. Price differentials of 2, 4, and 6 cents per pound above the prevailing retail Grade A price were used. Consumer preference for grade labels on raw turkey parts, turkey rolls, and cooked and sliced turkey rolls were tested by means of the Detroit Consumer Preference Panel. Price differentials of 2 and 4 cents per pound were used at one section of the panel for comparison with the 86 87 other section where no prices were used. The results of the sales tests were analyzed by the Chi Square method and the "t" test. Confidence intervals for all proportions were obtained. The theoretical sales ratio of Grade A to "Prime" or "Choice" was 2:1. The sales ratio of Grade A turkeys to "Prime" and "Choice" when price differentials of 4 and 6 cents were combined were as follows: Grand Rapids 55:45; Detroit 62:38; and Lansing 75:25. The combined sales ratio was 63:37 and did not differ significantly from the theo- retical ratio. The differences between the theoretical sales ratio and the actual observed store sales in different cities were highly significant. The differences between the several store sales ratios within each city were much less than the differences between cities. The analysis of the panel data showed that the coefficients of concordance were highly significant and that the rankings were un- likely to have occurred by chance. It does not mean that there is, or is not, a difference between any two classes within a particular series. A new turkey grading system was developed incorporating the terms "Prime", "Choice", and "Good" as grade designations and quality specifications for each grade were set forth. The proposed grading system would be another stride toward uniformity of grades for all meat. For years, red meats have been known as 'Prime , 'Choice , and Good . The transition would be simple for the consumer because of the familiarity with red meat grades, and less confusion regarding meat grades would exist. 88 Based upon the results obtained in this study and the statistical analysis of these data it is concluded that: 1. One-third of the turkey hens can be sold for 4 cents per pound more than the current price of Grade A turkeys when adequately differ- entiated and labeled "Prime". 2. Consumer satisfaction was apparently increased as measured by the willingness of some consumers to pay premium prices for turkeys in this study. 3. Some of the criteria by which consumers judged turkey quality in this experiment are: appearance, meatiness, wholesomeness, minor defects. 4. Consumers reacted favorably to grade labels "Prime", "Choice", and "Good" as labels for whole turkeys as well as for raw turkey parts and cooked turkey products. 5. One might expect aggregate sales of turkey hens in the United States to be similar to the sales ratios obtained in this experiment. 6. Consumer grades for turkey hens should be "Prime" and "Choice". 7. All other turkeys not meeting the specifications for the grades of "Prime" and "Choice" should be classified as "Good" and used for canning and other processed foods. 8. The proposed grading system apparently more nearly reflects consumers' opinions of quality than the present grading system. 9. Total industry revenue from a given size of crop would likely be increased under the proposed grading system. This study has pointed the way to possible increased profits for producers, processors, retailers, and greater satisfaction for the 89 consumer. The turkey industry can profit by grade revision, pro- duct differentiation, and proper pricing. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) BIBLIOGRAPHY Anonymous, (1945). Grade Labeling and the Consumer, An Informa- tive Study, The National Publishers Association, New York. Anonymous, (1956). Poultry Grading Manual, Agriculture Handbook No. 31. United States Department of Agriculture, Agricul- tural Marketing Service, Poultry Division, Washington 25, D. C., pp. 30-31. Anonymous, (1961). Regulations Governing the Grading and Inspec- tion 2f Poultry and Edible Products Thereof and United States Classes, Standards, and Grades with Respect Thereto, Part 70, United States Department of Agriculture, Agri- cultural Marketing Service, Poultry Division, Washington 25, D. C. Baker, Merritt W., (1954). "Grades and Grading", The Yearbook 2f Agriculture, The United States Government Printing Office, The United States Department of Agriculture, Washington 25, D. C. Benjamin, Earl W. and Howard C. Pierce, (1937). Marketing Poultry Products, John Wiley and Sons, Inc., New York, 3rd Ed. Brunk, Max E. and L. B. Darrah, (1955). Marketing of Agricu - tural Products, The Ronald Press Company, New York. Dixon, W. J. and F. J. Massey, Jr., (1957). Introduction £2 Statistical Analysis. MeGraw-Hill Book Company, Inc. New York. Farioletti, Marius, (1939). "Food Grades and the Consumer", Yearbook 2f Agriculture, United States Department of Agriculture, United States Government Printing Office, Washington 25, D. C., pp. 364-371. Ferber, Robert and P. J. Verdoorn, (1962). Research Methods 33 Economics and Business, MacMillan Co., New York, p. 304 Kendall, Maurice G., (1955). Rank Correlation Methods, Hafner Publishing Co., New York, pp. 95-105. Kohls, Richard L., (1955). Marketing 2f Agricultural Products, The Macmillan Company, New York 9O 91 (12) Larzelere, H. E. and R. D. Gibb , (1956). Consumers Opinions pf Qua11ty in Pork Chops, Michigan Agricultural Experi- ment Station Quarterly Bulletin, vol. 39 No. 2., pp. 327-333. (13) Makens, J. C., L. E. Dawson, H. E. Larzelere and H. C. Zindel, (1962). Consumer Preferences for Grade, Sex, and Age Lab- eling 2f Turkeys, Michigan Agricultural Experiment Station Quarterly Bulletin, vol. 44, no. 3 pp. 527 533. (14) Makens, J. C., L. E. Dawson and H. C. Zindel, (1962). Grade and Sex Labeling Market Turkeys, Poultry Science, vol. XLI, no. 3, pp. 822- 825. (15) Marsden, Stanley J., and J. Holmes Martin, (1955). Turkey Management, Interstate Printers and Publishers, Inc., Danville, Illinois, 6th ed (16) M1lls, W C., Jr., L. E. Dawson and H. E. Larzelere, (1960) Consumer Preferences for Turkeys with Various Grade Labels, Michigan Agr1cultural Experiment Station Quarterly Bulle- tin, vol. 43, no. 2, pp. 249-260 (l7) Mills, W. C. Jr., L. E. Dawson and H. E. Larzelere, (1961). A Study_ of Word versus Letter Grading of Turkeys, Poultry Science, vol. 40, no. 2, pp. 537- 542. (18) Poultry, (1962). Slaughtered under Federal Inspection and Poul- Used in Canning and Other Processed Foods, United States Department _of Agriculture, Statistical Reporting Service, U. S. Government Printing Office, Washington 25, D. C. (19) Swanson, Milo H. and Carroll V. Hess, (1961). Consumer Accept- ability and Usage _o_f Turkey Fryer Roasters, Departments of Poultry Husbandry and Agricultural Economics, University of Minnesota, Minneapolis, Minnesota. (Mimeograph) (20) Un1ted States Department of Agriculture, Agricultural Marketing Service, (1962), Correspondence. (21) Waugh, Fred, (1953). Readings in Agricultural Marketing, The Iowa State College Press, Ames, Iowa. 92 APPENDIX 93 Table 23. Detroit Turkey Sales by Stores, Grade Designation, and Price. (April, 1963). Grade A Prime Number Price per Number Price per Store Sold Pound Sold Pound l 34 $.38 16 _$.44 13 23 .39 22 .43 Choice 3 16 .39 4 .45 4 15 .39 4 .45 7 4 .39 4 .45 7 12 .39 12 .43 8 5 .39 4 .43 14 15 .39 9 .43 10 5 .39 4 .43 Choice Prime 5 35 .39 2 .45 6 ll .39 7 .43 Good Choice 12 38 .38 26 .42 11 2 .39 12 .45 Brand Name Good 10 6 .53 6 .57 10 4 .39 6 .43 94 Table 24. Grand Rapids Turkey Sales by Stores, Grade Designations and Price. (Novemberl 1962). ' Grade A Prime Number Price per Number Price per Store Sold Pound Sold Pound l 91 $.35 74 $.39 l 101 .35 65 .41 2 88 .35 44 .41 2 ll .43 4 .47 3 5 .43 2 .47 3 3 .43 3 .49 Choice 1 2 .43 l .47 1 4 .43 3 .49 2 59 .35 66 .39 2 0 .43 0 .49 3 73 .35 73 .39 95 Table 25. Lansing Turkey Sales by Stores, Grade Designation, and Price. (November, 1962 . ’ Grade A Prime Number Price per Number Price per Store Sold Pound Sold Pound 4 57 $.39 17 $.43 1 34 .39 17 .43 1 65 .39 14 .45 2 7 .39 9 .45 Choice 3 102 .39 29 .43 3 116 .39 29 .45 4 20 .39 12 .45 96 Computation of Coefficient of Concordance and Chi Square for Panel Data Table 11. Influence of Grade Label on Preference for Cooked Turkey Rolls ,by the Detroit Preference Panel. Afternoon, AprilI 1963 Z lst Rank Symbol Grade Label Place 1 2 3 Persons Z Prime 65.0 39 12 6 * Grade A 20.0 12 21 24 () Choice 15.0 9 23 25 Symbol Rank times number in rank, summed - ranks of all samples Z 81 X = number of ranks - 337 * 126 X = ‘3’ = 112 () 130 M = persons making rankings 57 N = number of samples 3 81 - 112 = (-31)2 126 - 112 = (14)2 130 - 112 = (1822 _ 2 2 1481 - d X?- = M(N-1)W 12Sd§ W M (N -N) M 77976 = .227916 57(3-1) (.227916) = 25.98242** ** = Significant at the lZ level 97 Chi Square for Table l City Grade A Prime (M (12.-192 fi Fi fi-Fi Fi fi Fi fi-Fi Fi Detroit 23 30 -7 1.6333 22 15 7 3.2667 Grand Rapids 107 125 -18 2.5920 80 62 18 5.2258 Lansing 91 83 8 .7711 34 42 -8 1.5238 221 238 4.9964 136 119 10.916; fi = observed frequency Fi = theoretical frequency ratio 2A: 1 Prime x2 = Z (£14122 = 4.9964 + 10.0163 = 15.0127** ** = significant at the lZ level Chi Square for Table 1 \ City Grade A - Prime (w‘ ($211112 fi Fi fi-Fi Fi fi Fi fi-Fi Fi Detroit 23 28 -5 .8929 22 17 5 1.4706 Grand Rapids 107 116 -9 .6983 80 71 9 1.1408 Lansing 91 77 14 2.5455 34 48 ~14 4.0833 221 221 44.1367 136y_:l36 6.6947 x2 = 4.1367 + 6.6947 = 10.8314** ** = significant at the lZ level fi = observed frequency Pi = calculated frequency = Row Total X Column Total Grand Total 98 Confidence Interval for Table l a_ 1162) (.38) S =V/ 357 =y/.000659 .62 - 1.96 (.02567) 5 p' 5 .62 + 1.96 (.02567) .025671 .62 - .05 ‘ p' f .62 + .05 95Z Limits .57 5 p' 5 .67 99Z Limits .55 s p' s .74 p' = probability that the proportion of Grade A Turkey Sales will be between these limits based upon this data. Confidence Intervals for Table 2 S =V n =\// 449 = J.OOO499 = .022338 .66 - 1.96 (.022) 5 p' £- .66 + 1.96 (.022) .66 - .04 5‘p' ‘ .66 + .04 95Z Limits .62“ p' 5 .70 99Z Limits .60 € p' e .72 p' = probability that the proportion of Grade A Turkey Sales will be between these limits based upon this data. 99 Chi Square for Table 2 City Grade A Ag Prime (ii-111‘ (_féfll" fi Fi fi-Fi Fi fi Fi fi-Fi F1;, Detroit 34 33 1 .0303 l6 l7 -1 .0588 Grand Rapids 192 203 -ll .596 112 101 11 1.1980 Lansing 72 63 9 _il;2857 23 32 -9 2.5313 298 299 V1.912 151 150 3.788; X2 =2: SféiFizz = 1.912 + 3.7881 .- 5.7001* fi = observed frequency Pi = theoretical frequency * = significant at the 5% level Chi Square for Table 2 City Grade A ‘ Prime (ft—4‘12? (2:21-12 fi Fi fi-Fi Fi fi Fi fi-Fi Fi Detroit 34 33 1 .0303 16 17 -1 .0588 Grand Rapids 192 202 -10 .4950 112 102 10 .9804 Lansing 72 63 9 1.2857 23 32 -9 2.5313 298 298 1.8110 3.5705 x2 = 1.811 + 3.5705 = 5.3815 Fi = Row Total X Column Total Grand Total fi = observed frequency 100 Chi Square for Table 3 City Grade A . Choice . (fi-Fi) 2 ( fi-Fi) 1' fi Fi fi-Fi E11 fi Fi fi-Fi Fi Detroit 37 44 -7 1.1136 29 22 7 2.1364 Grand Rapids 134 183 -49 13.1202 140 91 49 26.3846 Lansing 114 101 13 41.6733 37 50 -13 3.3800 285 328 415.9071 206 163 31.9010 x2 = Z {fi-Fizz = 15.9071 + 31.9010 = 47.8081** Fi fi = observed frequency Pi = theoretical frequency ** = significant at the 1% level Chi Square for Table 3 City Grade A Choice (EL-£127 (fl)! I__§i Fi fi-Fi Fi fi Fi fi-Fi Fi Detroit 37 38 -1 .0263 29 28 1 .0357 Grand Rapids 134 159 ~25 3.9308 140 115 25 5.4348 Lansing 114 88 26 7.6818 37 63 -26 10.7302 285 285 1136389 206 16.2007 X2 = 11.6389 + 16.2007 = 27.8396** fi = observed frequency Fi = Row Total X Column Total Grand Total ** = significant at the 1% level 101 Chi Square for Table 4 City Grade A Choice <_£1_.-_F..1_1_'2 (5.11112 fi Fi fi-Fi Fi fi Fi fi-Fi Fi Detroit 35 31 4 .5161 12 16 -4 1.0000 Grand Rapids 64 93 -29 9.0430 75 46 29 18.2826 Lansing 136 118 18 2.7458 41 59 -18 5.4915 235 242 12.3049 128 121 24.7741 x2 =Z§fi-F122 = 12.3049 + 24.7741 = 37.0790** fi = obsiived frequency Fi = theoretical frequency ** = significant at the 1% level Chi Square for Table 4 City Grade A Choice (mint)? (.fi-F—ilz fi Fi fi-Fi Fi fi Fi f1:£i Fi Detroit 35 30 5 .8333 12 17 5 1.4706 Grand Rapids 64 90 -26 7.5111 75 49 26 13.7959 Lansing 136 115 21 3.8348 41 62 -21 7.1129 235 235 12.1792 128 128 22.3794 x2 = 12.1792 + 22.3794 - 34.5586** ** = significant at the 1% level fi = observed frequency Fi = calculated frequency = Row Total X Column Total Grand Total 102 Confidence Intervals for Table 3 2 - £1.22 .58 042 S = p’ n = 491 =v’ .0005 = .022361 P‘t1.§_(S)-p'-p+t§(8) 2 2 m m .58 - 1.96 (.022361) .58 + 1.96 (.022361) .58 - .04 s p. s .58 + .04 95% Confidence Intervals .54 5 p' 5 .62 99% Confidence Intervals .52 5 p' 5 .64 p = observed proportion of Grade A sales p' = probability that observed sales of Grade A turkeys will fall within the range shown based upon this data. S = variance of proportion Confidence Intervals for Table 4 s = ual—.21 = 4.6536335) =m = .025 .65 - 1.96 (.025) 5 p' 5 .65 + 1.96 (.025) .65 - .05 5 p' 5 .65 + .05 95% Confidence Intervals .60 5 p' 5 .70 99% Confidence Intervals .59 f p' 5 .71 103 Chi Square for Table 7 City grade A Prime and Choice (fi-Fi)z (fut-Pi)? fi Fi fi-Fi Fi ‘ fi Fi fi-Fi Fi Detroit 129 139 -10 .7194 79 69 10 1.4493 Grand Rapids497 603 -106 18.6335 407 301 106 37.3289 Lansing 413 365 48 6.3123 135 183 -48 12.5902 1039 1107 A 25.6652 621 q 553 51.3684 _ 2 X2 =z(-f-%l) = 25.6652 + 51.3684 = 77.0336“ fi observed frequency Fi theoretical frequency 1 highly significant at the 1% level Confidence Intervals for Table 7 = ‘/ (.6166037 =\/--T—2§§0‘ =,/‘.0'—TZ'001 0 = .01185 CD .63 - 1.96(.01185) é p' *- .63 + l.96(.01185) W. 063 - 002 p. 5 .63 + 002 95% Limits .61 5 p' 5 .65 IR 99% Limits .60 p' 5 .66 104 Differences between Stores - Grand Rapids Store no. 1 n1 = 340 p1= .58 Store no. 2 n 272 P2: .58 M II $91 = 155854042) - ¢’.00071 $92 = v/fg.58;§é427 = .7’.00089 8(91-02) v’-00071 +-.00089 = .04 p1 -P2 = 2 = o t(p -p ) 1 2 Sprl-pZS .0 Store no. 3 n3 = 291 p3 = .43 Store no. 2 n2 = 272 p2 = .58 P3 291 \/.000857 Sp(p2-p3) =./7000857 +-.00089 = .0418 t 3 .58 -.048= W 2039 Store no. 1 Store no. 3 S(p1-p3) =./700071 + .00086 = .039623 t = 058 ' 048 = 2.53 .0396 t = 2.58 at 5% level of significance 105 Differences between Stores - Lansing Store no. 1 n1 3 130 p = .76 Store no. 2 n2 = 36 p = .53 Spl=/ (.76)(.24) .00 130 Sp2= g.53gg.47) = ‘/.OO691 S(p1-p2) = /.0014 + .00691 = .09116 p1'P2 .76 -.53 t(131132)“ Sp(p1-p2) = .09TI6 = 2-52 Store no. 3 n = 276 = .79 3 p3 Store no. 2 n = - 2 36 p2 .53 Sp =/ '79 '21 = \/.0006 3 27 S(p3-p2) = VCOOO6«+ .00691 = .0867 079 - 53 = ______:__.= 2. .0867 99 Store no. 1 Store no. 3 S(p1-p3) = ¢20014-+ .0006 = .0447 .79 - .76 .0447 t - 2.58 at 5% level of significance 106 Differences in Sales Ratios of Grade A Turkeys (between Cities) Lansing n1 = 548 p1 - .75 Grand Raplds n2 = 910 p2 = .55 (.75)g.252 _ Vflffifififif SP1= 548 ' _ (.55)(.45) _ sz - 910 - «.000271 S(p1-p2) = ¢C00034-+ .000271 = .024698 t=p1-P2_o75‘055 s - .024698 S = variance * = reject hypothesis of no difference if the absolute "t" value is i 1.98 at 5% level or i 2.58 at the 1% level Confidence Intervals for Real Difference in Grade A Sales, Lansing and Detroit m .20 - 1.96(.0247) Real Difference (D) 5 .20-+ 1.96(.0247) uh .20 - .05 D 5 .20‘+ .05 95% Limits .15 5 D 5 .25 'h 99% Limits .16 D 5 .26 107 Differences in Sales Ratios of Grade A Turkeys (between Cities) Continued Detroit n3 = 208 p3 - .62 Grand Rapids n2 = 910 p2 = .55 Spy/062%?” = 7.001132 .62 "' .55 t = .0374, = 1.87 Confidence Intervals for Real Difference in Grade A Sales, Detroit and Grand Rapids .07 - 1.96(.0374) e 0 €— .07 + 1.96(.0374) 007 ' 007 f D f .07 + .07 957. Limits 0 5 D 9 .14 997. Limits 0 5 D f .17 108 Differences in Sales Ratios of Grade A Turkeys (between Cities) Continued Detroit n3 = 208 P3 .62 Lansing n S(p1-p2) = [00034 + .001132 = .0165 _ .75 - .62 _ 7 ‘ . 01—65" ' Confidence Intervals for Real Difference in Grade A Sales, Detroit and Lansing .13 - 1.96(.0165) 9 D £- .13 + 1.96(.0165) 13 - .03.], é.13+ .03 95% Limits .10 5 D 5 .16 997. Limits .09 f D £— .17 "c {”1 '7}; ". uni?» " ‘ [PE 0;“ i'aii"'2! :3 M '5.- g.‘ g