SAMPLING AND RESPONSE DIFFERENCES FOR THREE METHODS OF ajOMERATICN OBTAIHED 3N A STUDY' OF CONSUMER POTATO PREFEREL'-ICES By MARCO AOTOJIO M E Z A A TUBS 23 Submitted to the ,School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the require tients for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economies 1950 ProQuest Number: 10008700 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest ProQuest 10008700 Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code Microform Edition © ProQuest LLC. ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 ACKN OALEDGiffliHT The author wishes to extend his thanks to the many persons who nave, through their efforts, made this study possible. Among these he wishes specifically to thank Mrs. Ruth Allen who helped in setting up the study and in the collection of the data; Mrs. Mildred Galloway for her inval­ uable assistance with the statistical analysis; Mrs. Harriet Siegel for reading the manuscript and offering suggestions; and Mrs* Ida Hammond for the typing of the manuscript. The author also wishes to express his gratitude to Dr. Roe Goodman for the help given him in determining the sampling procedure used in the study and to Dr. W. D. Baten and Dr. H.l. Harter for their suggestions about the statistical treatment of the data. Special thanks are due to the members of my Doctoral Committee, Dr. L. W. Witt, Chairman, Dr. M. E. Cravens, Dr. V, E. Smith, Dr, t. H. Combs, Graduate Council Representative, Dr. A. Mauch and Dr. D. C. Ciiae, for their kind consideration throughout my stay at Michigan State College and their helpful suggestions in regards to this study. I especially want to thank Dr, L. W. Witt and Dr. T. K. Cowden, Head, of the Department of Agricultural Economics, for help ana advice which extended far beyond the immediate scope of this study. Their sage counsel shall always be remembered, I also wish to express my gratitude to Dr, K, E. Cravens for his help in planning and conducting the study as well as his comments and suggestions relative to the manuscript. Finally, I would like to thank Dr. Lawrence A. Davis, President cf A, M. and N. College in Pine Bluff, Arkansas, for affording me the oppor­ tunity to study and for encouraging me at every turn. It is ray sincere hope that I may some day prove worthy of all the kind attentions given, r while completing my graduate studies. TABLE CF CONTENTS Page LIST OF T AB L E S ........................................... Ui Chapter I II III INTRODUCTION.............................. . . . . 1 Purpose of the s t u d y .............. . . . . . . . 1 Utility of the stufy 2 ........... . . . . . . . . .................... REVIEW CP LITERATURE. PROCEDURE AM) TECHNIQUES Place and. time of study. • . ............... V 10 Preparation of Questionnaire.................. 10 Selection of samples . . . . . . . . . . . . . . 11 Obtaining interviews . . . . . . . . . . . . . . 14 Tabulation and statistical treatment of data . . 15 Presentation of data . 16 ............. Definition of terms and explanation of techniques IV 3 17 SURVEY FINDINGS - METHODOLOGY Gaqparison of costs, returns and representative­ ness; of sample obtained by three methods. . . . 20 Relationships indicating statistically signifi­ cant differences when tested by analysis of variance. ........... * 24 Comparisons related to method of obtaining infor­ mation where analysis of variance was not used 28 SURVEY FINDINGS - CONSUMER PRACTICES AND QPUIICNS Consumption habits and consumer preference Respondents quality ratings of the po t™toes they had on h a n d ........ ................ .. i 46 TA3IT. OF CCHTMDS (C o a t'd ) Chapter VI Page m jLIGATIOJS OF STUDY Methodology • Marketing ofMichigan potatoes. VII . . . . . . . . . SUMMARY ANDCCWCLUSlOnS. . . . . . . . . . . . . . . 65 ?0 BIBLIOGRAPHY............... . » ........................ . . • 75 APPENDIX.................................................. 77 ii L E T OF TABLES 1 A Comparison of the He turns Obtained for the Three Sampling Methods and the Cost per Completed Ques­ tionnaire........................ ................. 2 Comparison of the Numerical Distribution of the Respondents by Rental Groups ................... . * 3 Number cf Blocks to Stare as Related to Method of Obtaining Data and Rental Groups ................. 4 Number of Blocks to Store as Related to Method of Obtaining Information and Income Level . . . . . . . . 5 Consumption of Potatoes as Related to Time the Telephone Interview was Completed • • .......... . & Number in Family an Related to Time the Telephone Interview was Completed . . . . . . . . . . . . . . . 7 Type cf Potatoes Respondents Use as Related to Time Telephone Interview was Completed • ® a Type of Store Shopped as Related to Time Telephone Interview Completed............................... 9 Distance Traveled to Shop as Related to Time Tele­ phone Interview was Completed. ................. • 10 Store Where Shopping was Done as Related to Time the Telephone Interview was Completed. . . . . . . . . n Consumption as Related to the Number of'Calls Fade to Complete Personal Interview . . . . . . . . . . . . 12 Number in Family as Related to Number of Calls Mad© to Complete Personal Interview. . . . . . . . . . . 13 Type of Potato Used as Related to Number of Calls Made to Complete Personal Interview .......... 14 Type of Store Shopped as Related to Number of Galls Made to Complete Personal Interview.. . . . . . . . . . 15 Number of Blocks Traveled to Shop as Related to Number of Galls Made to Complete Personal Interview. . 16 Consumption of Potatoes as Related to Whether or not Family has Telephone . . . . . . . . . . . . . . . . . iii LIST OF TABLES (Cont'd) Table 32 Page Gcaparison of Methods of Obtaining Data and 'type of Potato to Cost per Pound . . . . . . . . . . . . . 49 33 iype of Store Where Potatoes Were Purchased * • . . • 5° 34 Location of St cere Where Potatoes Purchased.......... $1 35 Income Level as Related to Humber in Family . . . • • 52 36 Percentage of the Respondents Who Gave Their Potatoes a Rating of "Good* for Various Quality Factors. . . . 54 37 Factors Considered Most Important. . . . . . . . . . . 57 38 Percentage of Respondents Who Rated Their Potatoes .......... •Good* When Prepared in Various Ways » 58 y LIST OF TABLES (Gont*d) Table 1? 18 19 20 21 22 23 24 25 26 Ihge Humber in Family as Related to Whether or Not Respondent has Telephone . . . • ........... ♦ . . 37 T^pe of Potatoes Used as Related to Whether or not Family has Telephone* * * 38 Type of Store in Which Respondents Shop as Related to Whether or not Family has Telephone* 39 » Number of Blocks to Stare as Related to Whether or not Family has Telephone * ...................... 39 Type of Area Where Store was Located as Related to Whether or not Family has Telephone .............. 40 Consumption of Potatoes by Respondents to Telephone Interviews Compared with Respondents to Personal ................. Interview that had Telephones 41 Number in Family for the Respondents to the Telephone Interview as Compared to the Respondents to the Per­ sonal Interview who Rad Telephones ......... 42 Type cf Potato Used by Respondents to Telephone In­ terviews. as Compared with those Respondents to the Personal Interview that had Telephones . . . . . . . 43 Type of Stare Where Shopping Was Done by Respondents to Telephone Interviews as Compared to Those Respond­ ents to the Personal Interview that had Telephones. . 43 Number of Blocks Traveled in Order to Buy Potatoes by the Respondents to the Telephone Interview Com­ pared with those Respondents to Personal Interview that had Telephones « « • « • . . . . 44 . 27 Inc cene Level as Related to Number of Pounds per Per­ ........ .............. . son per.Day 28 Answers to Questions Is the Kind of Potatoes on Hand the Same as that Usually Bought?. . . . . . . . . . . 47 29 Type of Potatoes on Hand. 47 30 Answers to Question# Does Grocer Usually have Kind Desired? . . . . . . . . . . . ....... 31 .......... Kind of Container for last Purchase.......... iv 43 48 SAMPLING AND RESPONSE DIFFERENCES FOE THREE IETHCDS OF ENW.IERATIQN OBTAINED IN A STUDY OF CONSUMER POTATO PREFERENCES CHAPTER I INTRODUCTION This study includes two phases of research which are of interest to economists. Cfae has to do with methods of gathering research data. Ibis phase deals with problems such as the validity, reliability, com­ parability, cost and returns obtained when each of the several methods of enumeration are used. It is important not only to economics but to all the social sciences and yet, relatively speaking, not very much work has been done in which two or more of the methods are used simul­ taneously and conclusions drawn concerning the value of each. A second phase of research has to do with a study of consumer preferences. In the final analysis it is the consumer who, in a free economy determines many of the economic relationships. More informa­ tion concerning the characteristics, habits and opinions of consumers is, therefore, invaluable as an aid to understanding the operation of our economic system. Purpose of the Study Ibis study was undertaken as a means of determining the compara­ bility and validity of consumer responses to questions of preference and consumption obtained by the following me thods $ (1) personal inter­ view, (2) mailed questionnaire, and (3) telephone interview. The three methods were tested (1) at the same time, (2) for reporting the same information in each case, and (3) in the same area. The subject ratter around which the methodological study was built had to do with consumer preferences for potatoes, since it is impossible to conduct a study of methodology without reference to same subject* While much has been written an the merits of each particular method, few studies have actu­ ally tried using all three methods simultaneously as a means of drawing conclusions as to the representativeness of their results. A further objective of this study was to compare the results ob­ tained with those reported on previous and current consumer studies of potatoes. Furthermore, this phase had both the advantage of being per­ tinent in Michigan and also the advantage of the availability of results of other studies on consumer potato preference. Utility of the Study The utility of the study is a two-fold one* C&e potential value of this study lies in evaluating the comparability of the results ob­ tained by the use of the various methods listed above in obtaining data on consumer preferences. A statistical analysis of comparable data ob­ tained by the three methods will show the differences in responses due to methods. 2h addition, certain conclusions are to be drawn about the cost, advantages, and shortcomings of each method* The study will also serve as a means of obtaining some inf oxmation about the characteristics, habits and preferences of consumers of pota­ toes. The findings will be tabulated, analyzed, and interpreted and the results used in making certain recommendations. 3 CHAPTER II REVlESf OF LITERATURE Literature dealing with the results obtained by the various samp­ ling methods was examined* These findings, as reported by other work­ ers, were found useful in setting up the present study* U Brown- discusses the three methods of obtaining data and the advantages and limitations of each* The personal interview has the following advantages: (1) it allows better control of the sample, (2) more questions can be asked, (3) field work can be done more quickly than with a mail questionnaire, (4) information can be obtained on sub­ jects not covered in a m i l questionnaire* It has the disadvantages: (1) of personal bias introduced by the interviewer, and (2) that many of the interviews may be given too hastily* Using the telephone interview has the advantage of: (1) being able to obtain, a large number of interviews quickly and at relatively low cost, and (2) spreading the interviews within an individual city in raudctu fashion* The disadvantages are: (1) the fact that it must be lim­ ited to those who have telephones, (2) that only a restoic ted amount and type of information can be secured, and (3) the difficulty exper­ ienced in toying to determine the age, economic condition and occupation of the person giving the information* The mail questionnaire has the following advantages: (1) it avoids the bias of the personal interview, (2) it allows for a wide distribu­ tion of the sample, and (3) respondents may take more care in providing information. However, it has the following disadvantages: (1) the sample 1/ Brown, L* 0«, Market Research and Analysis* Ronald Press, 1337* 4 obtained is not representative of the entire universe, since too great a proportion of certain income levels, educational levels, and respond­ ents interested in the subject distort the representativeness of the sample, (2) there is usually a high cost per return, (3) the question­ naire usually must be short, (4) it requires a longer time to ccmplete the study, and (5) seme types of information cannot be obtained, 2/ cite Hansen and Penning the following as aims in designing sampless (1) operate within the available budget and limitations of time and man­ power, (2) operate within other imposed administrative limitations, (3) produce maximum amount of information possible within aims 1 and 2 above, and (4) give results reasonably sure to fall within a certain allowable sampling error, 3/ Cochran reports that it is possible to use analysis of variance to determine the relative accuracy of various alternative methods of sampling the same population. He illustrates this with an example in which two methods of sampling are compared. y Hochstim and Smith, known universe. conducted a series of experiments using a They concluded that area sampling was more accurate, though more expensive, than quota sampling. After pointing out the advantages and disadvantages of various 2j Hansen, M« II, and W, E, Doming, "Cn Same Census Aids to Sampling," Journal of American .Statistical Association. 1943, 38*353-357. 3/ Cochran, W, G, "Use of Analysis of Variance in Enumeration by Sampling," Journal of American Statistical Association. 1939* 34*492-5X1. 4/ Hochstim, J. H, and H» K, D, Skaith, "Area Sampling or Quota Con­ trol? Three Sampling Experiments," Public Opinion Quarterly, 1948, 12*73-80. 5 types of sampling technique, B r o m questions the value of using the 1940 census as a basis for setting up a sample* He reasons that the very rapid changes that took place right after 1940 reduce the applica­ bility of the data* Eckler and Stuar state that census data for housing and popula­ tion are useful in two ways* (1) to describe a market with which a par­ ticular business organization is in contact, and (2) to provide area and other controls needed for special sampling studies of markets* u Bennett, found that local interviewers, when given sound instruct­ ions, can be relied upon to obtain reliable information* Shapiro and Eberhar tried to determine the extent of interviewer bias by training four professional interviewers and letting them work on the same areas*each having a randan sample* There was a significant difference in answers obtained to 10 of the 34 questions asked by the four interviewers* interviewers* Most of the difference was ascribed to one of the This occurred in spite of the fact that he had been trained with the others* The authors list these different types of in­ terview biass (1) on attitude questions, (2) difference in success in eliciting factual information, and (3) difference in method of classify- 3/ Brown, G* H*( "A Comparison of Sampling Methods,* keting* 1947. 11*331-337. Journal of Mar­ GJ Eckler, A* R* and E, P. Stuart, "Marketing and Sampling Uses of Population and Housing Bata,* Association, 1943* 38*87-92. 2/ Journal of American Statistical Bennett, A* S., *How Good are Local interviewers," 1947. 221*70-72* Printer’s Ink. 0/ Shapiro, S* and J. C. Eberhardt, "Interviewer Difference in an latensive Interview Survey," International Journal of Opinion and Attitude Research. 1947* lsl-1?* ing respondents' answers* 2/ It was also reported. that interviewer bias varies in degree with the type of question asked* Marked bias was found in the following types of questions* (1) those presenting two alternatives, (2) those where al­ ternatives only were partly stated, (3) those containing three alterna­ tives, and (4) those where lists of alternate choices were shown to the respondent* fflie least bias was found in "Yes", "No® type of questions. 10/ Jacoby reports that people are more reluctant to talk about fi­ nances than about social or political questions in answering personal interviews* She reports that better results are gotten when the inter­ viewer thoroughly identifies himself and the purpose of the study* 11/ Franzen and Iazarsfeld compared answers gotten on mail question­ naire with the results of personal interview and found that the results of mail questionnaire showed several consistent and significant biases* Same of these were too small, others were irrelevant and others like education, and size of city, could be overcome by weighting* They fur­ ther found an advantage of mail over personal interview in that the for­ mer contained more detailed answers on cultural questions and greater willingness to admit unusual interests or activities* Bdgerton, Britt and Norman^repcrted using a continuing mailed %/ Calahan, D«, V* Tamouris and H* W* Verner, "Interviewer Bias Involved in Certain types of Opinion Survey Questions,® International Journal of (toinion and Attitude Research* 1347* Is210*230. 10/ Jacoby, Eleanor, *Interviewing Problems in Financial Survey," Inter­ national Journal of Opinion and Attitude Research* 1947* 1 s54-58« 11/ Franzen, R* and P. F* Iazarsfeld, "Mail Questionnaire as a Research Problem," Journal of Psychology. 1945 * 20*^9-320. 12/ Edgerton, H* A*, S. H* Britt and R* D. Norman, "Objective Differences .Among Various types of Respondents to a Mailed Questionnaire,* Ameri­ can Sociological Review* 1947, 12 *435-444* 7 survey in which the same group of male high school graduates were polled each year* They found that the group who had the mo3t interest showed the greatest response* getting returns of 60 - 70 percent on mail questionnaire survey but he used such devices as a carefully prepared and tested letter, an enclosed 25 cent piece* and follow up postcards five to sis days after the original mailing* Collejr^concludes, on the basis of 50 surveys made by General Electric* that the mail questionnaire can give satisfactory results when used properly* He further shows that returns frcta mail questionnaire can be as representative as results from personal interview* While he agrees that interested people return the questionnaire, he believes that in consumer preference studies, interest is a virtue and not a handicap* He also points out that a mail questionnaire can be answered at leisure and costs less than the personal interview* Clausen and Ford^^state that bias can be minimized in a mail questionnaire by securing as large a response as possible and by making corrections for any bias caused by incomplete returns* He suggests the following as a means of obtaining better returns* (1) a letter with full explanation of the survey, (2) air mail special delivery, (3) several follow-up letters, and (4) maximize interest by asking questions that cover several different subjects in the survey* 1*3/ Robinson, Ray, "Five Features Helped This Questionnaire Pull frcm 60 to 70 Percent,® Printerfs link. February 22, 214*25-26. 14/ Colley* R* H*, *Dan*t Look Down Your Nose at Mail questionnaires,* Printer*s Ink* 1945* 210*21* 15/ Clausen* J* A* and R* N* Ford, "Controlling Bias in Ikil Question­ naires," Journal of American Statistical Association. 1947* 42*497- . 511 a personal interview with those who didn't return the mail questionnaire and concluded that it wasn't safe to assume that answers to mail questionnaire can he accepted as indicative of the experience and opinions of those who did not answer* Brooks made an analysis of a study conducted by Nev/ York World Telegram in which the original sample was picked from a list of regis­ tered voters* the sample* He found that the respondents were not representative of There was too high a proportion of certain age groups, cer­ tain incase groups and certain educational groups* He concludes that m i l questionnaires are not reliable to cover a heterogenous universe* recommends the simultaneous use of mail and personal interviews in readership research stating that the personal interview by it­ self is inadequate* 23/ Hansen and Hurwitz suggest that a remedy for the bias introduced in a sample survey might be a combination of call-backs and an increased number cf original interviews* Th&y suggest that a follow-up personal interview on non-respondents to mail interview has greater reliability and lower costs than either one used separately* 3h a Nation-wide survey conducted by the Bureau of Agricultural 16/ Eastman, R* 0«, "Dont's About M i l Questionnaires,9 Printer's Ink. 1943, 202*24-29. 22/ Brooks, V., "Can You Trust Mail Questionnaires,® Printer's Ink. 1947, 220*34-36, 18/ Baxter, R* E#, "Use Both Mail-type and Personal interview in Read­ ership Research,® Printer's Ink. 1947* 219*19-22, 19/ Hansen, M* H* and W. N. Hurwitz, "Problem of Non-Response in Sam­ ple Surveys," Journal of American Statistical Association* 1946, 41*517-529. Economics among 3300 urban homemakers, the following characteristics were mentioned most frequently by the respondents as being desirable in the outward appearance of potatoess (1) smooth skin9 (2) cleanness, (3) few eyes, (4) no spots or blemishes, (5) firmness, (6) no cuts or bruises, (7 ) no bumps, and (8) desirable skin color, The question is raised as to whether those characteristics mentioned most often need be those which the respondents consider most important* The most ecraaon reasons for preferring a particular type of potato were* (1) good taste, (2) cooking quality, (3 ) suitableness for several methods of cooking* A majority of the homemakers said that they considered quality more important than size or price and size more important than price* was some evidence that income had an effect on the answers* There *s income increases, homemakers were less concerned with prices and mare concerned with quality and size* The reverse is true for lower income groups* The average per capita consumption for all the respondents was 2.7 pounds per week* Low inccme families used about 2.8 pounds per person a week while those with a high income consumed about 2*5 pounds* About 60 percent of the homemakers said they bought their potatoes from bulk displays. The methods of preparing potatoes most frequently mentioned by the respondents were* (1) mashing, (2) boiling, (3 ) frying, (4 ) baking, and (5 ) creaming* 20/ U* S* Department of Agriculture, Bureau of Agricultural Economics, Potato Preferences Among Consumers. U, S* Dept* of Agriculture Miscellaneous Publication 667* 11$ p., CHAPTER III PROCEDURE AND TECHNIQUES Place and Time of Study The area chosen for the study was an area encompassing 5°»°0° families in the northwestern part cf Detroit, It included the area bounded on the north by the city line (8 Mile Road), on the west by the city line, on the south by Joy Road and the east by Livemoisj north from Joy Road to Davison, Greenlawn north from Davison to Outer Drive and Schaeffer Highway north from Outer Drive to 8 Mile Road* This area was chosen for two principal reasonss (1) its relative homogeneity in respect to racial make-up, and (2) its size was such that it could be sampled adequately by each of the three methods with the funds available* The three parts of the studies were carried on simultaneously over a 12-day period from the 6th to the 18th of December, 1948* phone survey was completed in four days. The tele­ All the mail questionnaires were placed in the stores for distribution during the period December '6 to 10, The personal interview, handled by nine different local inter­ viewers, took 10 days to complete. Preparation of Questionnaire Considerable time and effort were spent in drawing up the three questionnaires to be used for the study and in pre-testing them* Since each type of questionnaire causes a different reaction, certain types of questions would be expected to be answered much more readily when asked by one method as compared to another* it is difficult to make questions comparable* For this and other reasons 11 Sixteen questions were asked, in some form, on all three question­ naires. A comparison of the replies to these questions served as the basis far the response part of the study in methodology, The differ­ ences between the means of the answers to comparable questions by the three methods was tested by the analysis of variance. Various group­ ing were used* Selection of Samples Personal Interview. The sample used for the personal interview was drawn up with the help of Dr. Roe Goodman of the Survey Research Center at the University of Michigan. The 194® census cm housing for Detroit was the source of sampling information used* In this sampling the census tracts within the area tested were listed as were their average rentals and total dwelling units in 1940* The tracts were then grouped into strata according to average rental and geographic location. Some allowance was made in carder to have ad­ joining tracts within each group (stratum). Each stratum contained a varying number of tracts, a minimum of 800 dwelling units. The strata were arranged according to average rental value from the highest to the lowest. Within each stratum the tracts were then listed in consecutive numerical order. The cumulative dwelling units were then determined and it was found that according to 1940 figures there were 49*494 within our area. corded. For the main sample, each tract’s total dwelling units were re­ Using this figure and the one far cumulative dwelling units, every 400th unit was determined. The block within each area containing every 400th unit was determined. The number of dwelling units within the blocks selected were then recorded. In order to cut down on the expense, it was decided to sample half as many blocks but to take two 12 interviews within a block* 2hus, it was possible to cut down interview­ er traveling time* Within each block the interval for interviews was one-half the total dwelling units in that block, making it possible to obtain at least two interviews per block. A random number table was used to pick the actual unit to be interviewed* The interviewer began at the north­ west corner of the block and counted all the dwelling units until he came to the one listed* This number plus the interval gave the number of the second dwelling to be sampled and so on* No substitution of addresses other than those selected in this way was permitted* For the census tracts that had three or less dwellings in 1940* it was thought likely that seme building had taken place since the census* In these tracts it was decided to sample twice as many blocks as for the rest of the sample* The sample here was to be the block containing every 200th unit in the 1940 census and to sample every tenth housing unit found in this block* The first dwelling was again picked using a random number table with the number of units sampled depending on the total number of dwelling -units. Thus an attempt was made to correct for changes that might have taken place since the census. Telephone Interview* at randan* The telephone interviewees were also selected The names of the exchanges serving the area were obtained from the Michigan Bell Telephone Company* Then every fiftieth page of the Detroit telephone directory was scanned and addresses and telephone numbers.for each one of the ten exchanges found in the area were under­ lined* At the same time all addresses falling outside the area were eliminated. listing. Bach exchange was then sampled separately, using every 7th The sample was drawn in three parts. Tie first part contained 13 200 listings and each of the subsequent ones only 100. fore, really three random sauries* It was, there­ In this way it was possible to terminate the study at approximately 200, 300 or 1±Q0 calls* Furthermore, the first grouping was to be exhausted before starting on the next group* Thus, it was assured that there would be adequate call-backs before an additional sample was introduced* The telephone numbers and addresses were rechecked and placed on the questionnaires which were to be filled out at the time of the tele­ phone interview* Mail Questionnaire. Tie questionnaires for this portion of the study were distributed to consumers through randomly selected retail food stores in the sample area. An individual return-addressed post­ age-paid envelope was furnished with each questionnaire* A group of chain and independent stares were selected and bundles frcm 50 to 1000 questionnaires were left at the cash register* Instructions were given in each store for the cashier to place an envelope inside of each con­ sumer shopping bag until all envelopes were distributed, therefore, it was not necessary for a shopper to buy potatoes in order to be given a questionnaire* Tie stores to which questionnaires were distributed were selected from Route Lists of Retail Stores published by the Detroit News and frcm chain grocery lists. In the Detroit News list the stares are classified into A, B and C, depending on size, retail sales, location and similar factors. It was estimated that A stores and chain super stares did 80 percent of the business, and that the class B and C stores did 20 percent of the business. The questionnaires were then distribu­ ted to stores selected in these proportions at rand cm from these groups. In all, 58 percent of the 11,500 questionnaires were distributed through national and local chains and 42 percent through independent stores. Obtaining Interviews Personal Interview. The personal interviews were taken by a group of nine Wayne University students. briefed. These interviewers were carefully They reported to a central office each day to turn in com­ pleted questionnaires and to discuss problems. Some of the interviews had to be taken at night since the method of selection forbade the taking of an alternate house when the respondent was not at home during the day. Since the interviewers were well instructed and were paid by the hour, it was believed that departure from this rule was at a mini­ mum. Telephone Interview. The telephone interviews were taken by the author and full-time employees of the Department cf Agricultural Econom­ ics at Michigan State College, For the actual interviewing, two private roams with tables were rented in the Michigan Bell Telephone Building in Detroit. Calls were placed from 9*30 a»m. until 5*00 p.m., with no calls from 11*45 to 1*00 p*m. La order to complete some calls, it was necessary to make them from public telephone booths after 7*00 p.m» at night. The date and time of call were noted on the questionnaire as were the number of times the call was tried and the interviewer making the call. All the calls were completed within the four-day period from December 6 to 9« The average length of time to complete a call was about four minutes. This time varied considerably depending on amount of comments that the interviewee added in answering questions. 15 Tabulation and Statistical Treatment of Data After all the various interviews were completed, the information was coded so as to make it suitable for card punching. were listed numerically or given numerical values. All answers Thus, answers to certain questions like "What kind of potatoes do you have on hand?" were coded as follows* 1 for Michigan, 2 for Maine, and 3 for Idaho, Where gradations were indicated, the designations were made frcm low to higi. These numerical designations were used throughout the statistical por­ tion of the study. In order to be able to compare the results obtained by using var­ ious methods, it was thought desirable to test the answers to the com­ pared questions by subgroups based on ccumonly known and used consumer classification characteristics and to answers to other questions. considered were related to; Those (1) income level, as indicated by median rental data for census tracts in the 194® census and by the Detroit News classes, (2) kind of potatoes, (3) reported family rate of potato con­ sumption, (4) number of people in the family, (5) kind of stare, (6) type of shopping area, and (7) number of blocks to store. In the comparisons of the three methods two major sources of vari­ ation were tested. These were* (1) differences in the samples obtained by the three methods, and (2) differences in the responses to similar questions obtained by the different methods cf interviewing, that is — telephone, personal or mail. To test for the first difference above, the chi-square test was used first for comparisons of the three methods si­ multaneously and then used to compare each method with the personal in­ terview sample. This latter test was considered desirable because the personal interview sample was thought to be more nearly representative 16 of the area than either of the other two* Si testing the three methods simultaneously, very few cases were found where the samples taken by the three methods yielded a chi-square test that was significant at the ,05 level and these differences were not considered large enough to invali­ date the results obtained by the use of the analysis of variance* 2h testing the results where no significant differences in sampling occurred, a forn of analysis of variance suggested by Snedecor 21/ for use where the numbers in the sub-classes are not equal was used* When significant differences were shown by the chi-square test or analysis of variance, the *t* test was applied to determine which group averages were significantly different* Two other types of tests were also used. One type of *t" test was used for analyzing differences in weighted averages and another type for testing the differences in percentages. They were used for comparisons 7/here the three methods were not directly involved. For instance, comparisons were made between calls completed during the day and those completed at night* Bresentation of Data For measurable quantities like the number of days, number in family, and the like, the averages used in the analysis of variance were derived directly from the original data. The tables used in discussing these types of data contain these same averages. For the variables that could not be measured directly such as “kind of potatoes bought®, 21/ “name of stare,® and “size preference," arbitrary Snedecor, G* W., Statistical Methods. Applied to Experiments in Agriculture and Biology (hth ed.). Iowa State College, 1948. 17 numbers were assigned when coding these data. The averages used in analysis of variance were derived from this arbitrary designation. In presenting these data, however, the simplest way was to express the actual data in terms of percentage of respondents. Thus, for this par­ ticular type of data, the analysis of variance (shown in the appendix) was run on the averages, using arbitrary numbers to designate nonnumerical data while the tabular comparisons are presented in terms of percentages. Definition of Terms and Explanation of Techniques A few terms and practices used in the study need explaining and clarifying. Inc cine Areas. The Detroit Mews classified the City of Detroit into five income areas by bringing 1940 census figures up to date (1948)* Only a very small area was classed as very high and none in the very low area. 3h the analysis, the two highest Detroit News income areas were combined. All the analysis based on income was made on this basis. The three groups then were named "low" (really medium-low) % '"medium® (same as Detroit News); and "high® (medium-high and hi^h combined). None of the area was classed as low. Method. The term "method® refers usually to comparisons among the three methods used in obtaining the interviews phone interview and mail questionnaire. personal interview, tele­ Thus, any statement to the effect that "there is no difference due to method® means simply that substantially the same results were obtained using each method. Chain Store and Independent Store. Stores were classified in two ways. In one designation, all stores were classified as chain stores when they had a central buying organization and two or more of them were 18 owned and operated by the same company* The second designation further subdivided chain stores into super markets and others* Independent store was used when the same concern owned and opera­ ted only one store* Since it was impossible to visit every store in the area* the independent stores were not classified* Humber of Blocks* The distance that respondents traveled to shop for their potatoes was calculated using a fired scale of eigit blocks to the mile* Cn all three questionnaires, the respondents were asked the name and address (to the nearest cross-street) of the store where they made their last purchase* This location was plotted on the map and the distance between it and the respondents* residence was determined from the scale* Shopping Areas. The Detroit News classification map of major and minor shopping areas was used to designate store location* It was arbi­ trarily decided to consider major shopping areas as extending five stand­ ard sized blocks each way from the intersection along the main street and two blocks each way along the cross-street* Minor shopping areas were considered as extending three standard sized blocks each way from the intersection along the main street and one block each way along the cross-street* Stores falling outside these limits were not considered to be in a shopping district* Kind or Type of Potatoes* This designation refers to whether or not the potatoes in question were from Michigan* Maine or Idaho, While the respondents’ knowledge was probably not infallible, the fact that most potatoes were either in consumer bags or marked on the display is believed to have eliminated most error* 19 Non-reanondenta. la Chapter If, figures are presented to indicate the number of interviews obtained by each of the three sailing methods. All the subsequent discussion deals exclusively with the completed inter­ views, •Non-respondents" refers to the number of these interviews for which no answer was given or determinable far a particular question. The reason for this void information may have been due to failure to recall* incomplete answers, or refusal to give an answer. For instance, where a respondent failed to give his street address (on the mail ques­ tionnaire), it was impossible to determine the number of blocks from his house to the store where shopping was done. respondent, He was classified as a non­ Thus, the category of non-respondents was a heterogeneous classification denoting a lack of classifiable information, whatever the reason. 20 CHAPTER If SURVEY FINDINGS - METHODOLOGY The results of the study will he discussed in two distinct parts* Chapter If concerns the results frctn the standpoint of methodology. This includes first of all a discussion of the representativeness of the samples obtained by the three methods* Secondly, it includes instances where the results, using an analysis of variance in which all three methods were considered simultaneously, varied significantly ac­ cording to the method of obtaining the data* Thirdly, it includes rela­ tionships such as the comparison of day and night calls, the effect of possessing a telephone on certain consumer characteristics, etc* A Comparison of the Costs* Returns and Representativeness of Samples Obtained by the Three Methods Che of the ways in which each of thedifferent methods, telephone, personal interview, and mail questionnaire, differ is in the degree to which information was obtained from all the sampled interviewees. Thus, if a number of families supposedly representative of the area are chosen aa a sample and the number of those actually giving information were to vary according to the method, it would seem to indicate the possibility of one not being as representative as theother. There were 286 families selected andinterviewed by interview* the personal Of this number, 277 or 97 percent were completed* For the telephone interview & 393 family sample was selected for interview* pleted, A total of 308 or 78 percent of these interviews were com­ Refusals were 9 percent and disconnections were 3 percent of the total, the total* Ihose who did not answer their phones made up 10 percent of 21 Of the ll^QQ mail Questionnaires distributed through retail stares, only 480 were returned in the following six week period but mostly with­ in the two week period* This represents about 4*2 percent of the total selected interviewees and is in marked contrast to the preceding two. Part of this low return can no doubt be ascribed to the usual causes of non-personal appeal and lack of interest, Another element to be consid­ ered here is the fact that questionnaires distributed through sane 3tares had very good returns while in others in similar neighborhoods where the same number of questionnaires were left originally, no questionnaires were returned* There is reason to suspect, therefore, that certain stores did not bother to distribute their allotment after agreeing to do so. Table 1, A Comparison of the Returns Obtained for the Three Sampling Methods and the Cost per Completed Questionnaire Type of Survey Humber of Completed •Questionnaires Percent of Sample Cost per CcBcleted Questionnaire Personal Interview 277 97 $2,?2 Telephone Interview 308 78 •44 Mail $iestionnaire 480; 4 •54 Comparison of Coats* An important consideration in comparing re­ search methods is that of the cost involved* ftie figures presented above include only expenses incurred frcm the time that the decision as to the number and method of selection of samples were made until the completed interviews were ready for analysis. Since all the questionnaires were prepared simultaneously and subject to constant revisions, the costs of preparing them were considered equal for 22 the three methods and are not Included in the following figures* The telephone interview was the least expensive, costing only 44 cents per completed questionnaire* The mail quest ianna ire 54 cents and the personal interview cost $2*72 per completed questionnaire* The higher cost for the mail questionnaire than the telephone interview was due in part to the very low returns obtained* Representativeness of Samples Used in Study« In selecting a strat­ ified random sample for the personal interview, census tract rental information and block; statistics from the 1940 census were used* In analyzing the results for comparability, one of the first steps under­ taken was a comparison of the distribution of the respondents by rental groups to see whether or not the three methods gave substantially the same result* result* A chi-square test was made and a value of 14,60 was the This was well within the chi-square value of 21*026 needed for significance at the *05 level* Table 2* Rental Under $29 $30 - $39 $40 - $49 $50 - $59 $60 - $69 $70 - $79 $80 and over Total Comparison of the Numerical Distribution of the Respondents by Rental Croups Personal Telenhone Mail (Number cf respondents) 33 6? 116 42 9 7 20 44 79 39 18 11 ail 221 329 34 5965 61 28 14 10 Total 87 166 281 143 55 32 JH m An interpretation of the results of this ehi-squar© test leads to the conclusion that, insofar as the numerical distribution of respondents 23 among the rental groups was concerned, there was no significant differ­ ence among the three samples. It was this fact that suggested the analysis of variance as a means of comparing the results, There remained the possibility that though the chi-square test, considering all three methods simultaneously, indicated no significant difference, a similar test comparing any two of the methods mi^it yield different results. Table 2. These comparisons were made using the figures in Results of the chi-square test indicated that there were no significant difference between the numerical distribution of the re­ spondents by rental areas when comparing the personal with mail quest­ ionnaire or telephone with the mail questionnaire. However, a compari­ son of the samples used for personal interview and the telephone inter­ view gave a chi-square figure which was significant at the ,Q5 level. This can be interpreted to mean that the two samples are not equally representative of the same population. Since this difference appeared when only two methods were cougared, it did not invalidate the use of the analysis of variance which was based on the simultaneous use of all three samples. The information was used as a cheek on scans of the re­ sults gotten from the analysis of variance. 3h conclusion it can be stated that, when considering all three samples simultaneously, there is no significant difference among the samples obtained by the three methods. The next problem is one of con­ sidering the comparability of the results obtained by the three methods. This presents another opportunity far comparing the reliability of the three methods. 24 Relationships Indicating Statistically Significant Differences When Tested by Analysis of Variance There were a few relationships for which the analysis of variance indicated a significant difference* This section deals with these re­ sults and their interpretation. Number of Blocks to ■Store as Related to Rental Area and Method of Obtaining lata. The average distance between the residence and the re­ tail store varied significantly according to the method of gathering the data, the income level, and the rental area. In other words, when the analysis of variance was used, each of the preceding factors yielded *F® values that were very significant. As far as method is concerned, when considered in conjunction with rental area, the average number of blocks (7 ®2) for those interviewed by telephone was significantly higher than that for those interviewed by mail questionnaire (4 *8), or by personal interview (4 ,3 ) The difference between the averages pertaining to those inter­ viewed by mail and parsons! interview was significant at the ,05 level (Table 3)® As will be noted in comparing Table 2 with Table 3, there is a difference in the numbers of non-respondents. The larger number in Table 2 is due to the fact that non-respondents were made up of those for whom there was no rental information plus those for whom it was impossible to determine the number of blocks to stcre. This latter eventuality came about when either the address of the respondent or the store address were not given. There was no evident trend in the average of the number of blocks traveled when progressing from the lowest rental group to the highest. 25 Table 3* Number of Blocks to Store as Belated to Method of Obtaining Data and Rental Groups. Rental Gt o u d Personal Telephone (Average) Mail Average 5*6 5.2 6.8 6.3 4.3 5.1 5*3 Under $30 $30 - $39 $40 - & 9 $50 - |59 $60 - $69 $70 and over Average 5.1 6.6 3.1 3*3 2.9 5.1 4.3 10*2 7*5 6*1 6*0 Total respondents Nonrespondent s 206 102 204 6.8 10.6 7.2 73 3.8 5.8 6.0 IsQ. 4*5 4.8 5*4 269 211 The average number of blocks traveled was not significantly different when the two lowest groups (average 6.8 and 6*3 ) below $30 and the highest group (average 7*0 rental over $70) are considered* However, the averages pertaining to these three are each significantly higher than the average for each of the other rental groups ($0 - $49* $5° • $59, and $60 - $69)* All differences are highly significant* A chi-square test failed to show any significant difference due to sampling method, when all three methods were considered simultane­ ously* Other chi-square tests ccsnparing two methods at a time yielded a chi-square figure that was significant at the ®05 level when the per­ sonal and telephone or the personal and mail samples were considered* Differences in sampling* therefore, cannot account for the very signifi­ cant "F* figures that were obtained* The significantly higher number of blocks found in the case of the telephone interviewers may have been due to the fact that this method gave the respondent less time to answer the question than did the other 26 two* With little time for thought, it may he that the address (to the nearest cross-streets) of the store where shopping for potatoes was done may have been given too rapidly and thus erroneously* 'Then, too, the difference may be ascribed to not understanding the question as well when posed in this manner as when asked by the other two methods* Thus, the best explanation seems to be that the very nature of the interview, as compared to the other two, may have led to the significant difference. Effect of Income and Method. When considering method in the follow­ ing table, it is found that there was a significant difference among the averages for the three methods. For the personal interview, the average number of blocks was 4 ®9 »f°r telephone it was 7*3 ® and for mail it was 5 *7 * 'The average for each method was significantly different at the *01 level than either of the other two averages (Table 4). Table 4* Number of Blocks to Store as Belated to Method of Obtaining Information and Income Level Personal Telephone (Average) Mail Average Low Medium High Average 5.6 5*2 4.0 4.9 11.1 7*5 6*1 10.5 5*6 Jt*2 5*7 9*3 6*0 6*0 5*9 Total respondents Non-respondent s 237 40 234 74 Uncase ■7*3 327 153 The number of blocks traveled to shop varied significantly by the income level, as measured by the Detroit Mews* This was indicated by the very high "F* value obtained in the analysis of variance. In this 27 case, as one goes fraa low income to high income groups, the average number of blocks decreases significantly. For low income it is 9.3 , for the medium it is 6*0, and for the high it is In each case the difference between any two of these is significant at the *01 level as indicated by the corresponding ®t8 tests# The chi-square figure far the above table was significant at close to the *01 level* Thus, it may be possible that some of the differ­ ence tround may be due to difference in sampling# However, the ®F® value was much too high to be explained entirely by this difference. Attempts to explain the difference between the averages for the methods of sampling led to the same procedure and the same results as in the case of rental area and method discussed previously. There is, however, a discernible trend in regard to income level* As one progresses from low to high inccme the average number of blocks traveled to shop decreases significantly. One explanation for this may well be that people in medium low inccme areas, being more conscious of prices, travel farther in order to shop for bargains* The reason for the noticeable trend in this case, whereas none was found when rental areas were considered, may be explained in this manner* More respond­ ents were classifiable by inccme level than by rental groupings. That is, those who lived on blocks containing three or less houses in I940 were considered non-respondents for rental but were included in the table for inccme groups. Including the greater number that were classi­ fiable into high inccme group decreased the average number of blocks# Thus a trend was discernible# 28 Cggpariaona Related to Method of Obtaining Information Where Analysis of Variance was not Used There were certain relationships thought to have methodological significance, ences, Analysis of variance was not used in testing the differ­ Instead, two types of *t* test were used* Ctae type of "t* test was used to analyze differences in weighted averages and another ■t" test was used to test differences in percentages (See Appendix A). The relationships tested and discussed includei (a) comparison of telephone interviews completed during the day with those completed at night, (b) comparisons of personal interviews completed on first calls and those for which call backs were necessary, (c) the effect of pos­ sessing a telephone on answers given to the personal interview,and (d) comparison of answers given to the telephone interview with the answers given by respondents to the personal interview who had tele­ phones. Each of the above is discussed in relation tos (1) consumption, (2) number in the family, (3 ) type of potato, (4 ) type of store where shopping was done, and (5 ) distance traveled to shop. Comparison of the Answers Cbtained in Telephone Interviews Completed During the Day with Those Completed at Night, Ihen the telephone was not answered during the day, interviewing in the evening was necessary to complete the prescribed sample. This section compares those respondents who were not available during the day to those who were at heme during the daytime. Potato Consumption, Consumption, as measured by pounds per per­ son per day, was little different for persons interviewed during the day and those completed at night. In the first case the average daily consumption per person was 0,33 pounds while for those completed at 29 night the average consumption was Q)«38 pauads. A 181" test indicated that this difference was not significant (Table 5)* Table 5* Consumption of Potatoes as Belated to Time the Telephone Interview was Completed Completed night Completed dav (Percent of respondents) Consuumtion (Pounds per day) Up to .29 0.30 to 0.49 0.50 to O .69 0*70 and over Total 55 25 Average per person per day Total respondents Non-respondents Table 35 11 40 19 100 6 100 0.35 21*4 17 0.38 43 4 Bomber in Family as Related to Time the Telephone Interview was Completed Number in Family Completed night Completed day (Percent of respondents) 2 and under 23 3 -5 6 and over Total 68 — 2 100 2 100 Average 3.6 2.8 Total respondents Non-respondents 256 45 5 2 Number in Family» 51 47 The size of family of those respondents who were interviewed during the evening was significantly smaller than for those 30 interviewed during the day* For those interviewed during the day, the average was 3*6 persons, while for those interviewed at ni^it it was 2*8 persons* The difference is statistically significant at the •01 level (Table 6)* Tvne of Potatoes Used* Both groups, those completed during the day and those during the evening, had about the same percentage of users of each type of potato* In both cases, Michigan users made up about i|X) percent of the responses with Maine and Idaho about j0 per­ cent each (Table ?)• The differences were not statistically signifi­ cant* Table 7* Ttype of Potatoes Bespondents Use as Belated to Time Telephone Interview was Completed Tvpe potatoes used. Completed nirfit Completed day (Percent of respondents) 38 28 33 Michigan Maine Idaho Other Total 100 39 -33 26 2 100 Total respondents Non-respondents 231 30 43 4 I Type of Store. There was no significant difference between the answers from daytime and evening interviews regarding ■die type of stcr® in which potatoes were bought * In both cases, about k4 percent of the respondents shopped in super chain stores, 3^ percent shopped in neighborhood stores, 10 percent in the "other® chain stores, and 9 percent from sources such as roadside market farms and peddlers (Table 8), 31 The differences were not statistically significant. Table 8. Type of Store Shopped as Related to Time Telephone Interview Completed Kind of Store Completed day Completed night (Percent of respondents) Super chain stores Other chain stores Independent stores Roadside, Peddler, Parm Total Total respondents Non-respondents Table 9* 45 12 34 -JSL 41 9 41 100 100 22? 44 3 34 Distance Traveled to Shop as Related to Time Telephone Interview was Completed Number blocks 2 and under 3-5 6-8 9-12 13 - 16 Completed dav Completed night (Percent of respondents) 40 23 14 5 4 43 11 6 11 3 Jii Over 16 Total 10O Average 7.0 9.4 Total respondents Non-respondents 197 64 12 Number of Blocks to Store. 100 35 The average distance traveled by the respondents to the telephone interview was greater for those inter­ viewed in the evening compared with those interviewed during the day. 32 Far those interviewed during the day, the average number of blocks to the store was 7.0 while far the night interviews the figure was 9.4 blocks. This difference is due principally to those respondents who traveled over 16 blocks and is statistically significant at the *01 level (Table 9). Location of Store Where Potatoes Bought. There was almost no difference reported in the location of stares whether the telephone interview was completed during the day or in the evening. In either case, about 56 percent of the respondents shopped in stores considered outside of a shopping area. There were about 35 percent who shopped in stores located in minor shopping areas and about 9 percent who did their shopping in stores located in major shopping areas. Results of the *t* test indicate no significant difference between day and ni^it calls (Table 10). Table 10. Store Where Shopping was Done as Related to Time the Telephone Interview was Completed. Tree Shoot)ins Area Completed night CeusDleted day (Percent of respondents) Major shopping area Minor shopping area Outside of either Total 10 35 -55 Total respondents Non-respondents 195 66 100 6 35 ' -52 100 34 13 Summary of Comparisons Between Day and Night Telephone Interviews, The respondents interviewed during the day and those interviewed at night gave almost identical results insofar as type of store they shop, 33 location of store, type of potatoes they use, and consumption* There were significant differences in the number in the family and number of blocks traveled to shop* Those interviewed during the day traveled about 7*0 blocks to the retail store and had an average of 3*6 persons in the family. Those interviewed at night traveled on the average 9*4 blocks to the store and had an average of 2*8 persons in the family. Both differences were significant at the *01 level. Comparison of Personal Interviews Completed on the First Call with CallBacks. A comparison was made between those personal interviews completed cm the first call and those for which the absence of the respondent from the home necessitated one or more call-backs. The objective here was to try to determine whether or not the re­ spondents interviewed on call-backs were different than other respond­ ents* If the answer was in the negative, one might wonder how important it was to undertake the extra time and cost of several call-backs to complete a selected interviewee. Table 11. Consumption as Related to the Humber of Calls made to Complete Personal Interview Consumption (Pounds per person per day) First call Call-backs (Percent of respondents) Up to 0.29 0.30 - 0*49 0.50 - 0*69 0*70 and up Total 50 29 37 35 11 10 100 14 JA Average 0.36 0.37 Total respondents Non-respondents 191 14 80 0 100 34 Consumption of Potatoes. There was no significant difference be­ tween the two groups in consumption per person# both groups consuming about 0*4 pounds of potatoes daily. There was a slight difference in the distribution of the respondents relative to consumption (Table 11). Humber in Family. The number of people in the family was the same far the personal interviews completed on the first call as for those on which call-backs had to be made. The average for both was 3*7 persons per family (Table 12). Table 12. Number in Family as Related to Number of Calls Maue to Complete Personal Interview Number in family 2 and under 3-5 6 and over Toual First calls Call-backs (Percent of respondents) 22 « 70 — §. 100 30 62 100 Average 3.7 3.7 Total respondents Non-respondents 195 10 76 6 Tvne of Potatoes. The interviews completed on the first call gave almost identical results to those requiring additional calls. cases almost half the interviewees used Michigan potatoes. In both None of the differences were statistically significant (Table 13). Type of Store. There were no differences in the type of store where potatoes were purchased between those personal interviews completed on the first call and those completed in more than one call. In both cases about 60 percent of the respondents shopped in super chain stores, 10 percent in *other" chain stores, and 30 percent in independent stares (Table 14). Table 13* Type of Potato Used as Belated to Number of Calls Made to Complete Personal Interview Tvne of 0otatoes Call-backs First call (Percent of respondents) Michigan Maine Idaho Other Total 1 100 Total respondents Nonresp ondents 187 18 Table 14* 47 27 25 46 32 18 -Jk 100 77 5 T^pe of Store Shopped as Related to Number of Calls Made to Complete Personal Interview First call Call-backs (Percent of respondents) Kind of store Supra? chain stare "Other" chain store independent store Total 59 10 -31 loo 65 6 - 2a Total respondents Non-resp ondents 177 28 68 14 100 Number of Blocks to Store. "316 average number of blocks between the residence and the store was 4«8 Tor those personal interviews com­ pleted on the first call and 3*5 far those for which mare than one call had to be made. The difference is not statistically significant (Table 15 36 Table 15• Humber of Blocks Traveled to Shop as Related to Humber of Calls Made to Complete Personal Interview Number of blocks First call Call-backs (Percent of respondents) 2 and under 3 -5 6-8 9-12 47 42 26 12 26 11 7 4 13 13 - 16 Over 16 Total loo 100 Average 4*8 5.5 Total respondents Non-respondents 171 66 16 34 6 Summary of CcraDarisona Between Interviews Ccmnleted on First Call and Call-backs. There were no significant differences in the responses gotten by the two groups. Both groups consumed an average of about 0.36 pounds of potatoes, and had an average of 3*7 persons in the family. They shopped in the same kind of store, used the same types of potatoes, and traveled about five blocks to shop. Hie Effect of the Possession of a Telephone on the Answers Given to Personal Interview. Hie respondents to the personal interview were divided into those with and those without telephones in carder to compare their responses to selected questions* Consumption of Potatoes*. Hie respondents to the personal interview who had telephones consumed on the average 0*36 pound per person daily. Those who had no telephones had an average daily consumption of 0*39 37 pounds. The average consumption per person was not significantly dif­ ferent between the two groups (Table 16). Table 16, Consmotion of Potatoes as Related to Whether or not Family has Telephone ConsumDtica. (Pounds per person per day) Without telephone With teleohone (Percent of respondents) Up to 0,29 0.30 to O .49 0,50 to O .69 0,70 and over Total 48 30 31 28 11 -11 100 23 8 100 Average 0.36 0.39 Total respondents Non-resp ondents 252 13 4 0 Table 17. Number in Family as Related to Whether or Not Respondent has Telephone Number in Family Without telephone With telephone (Percent of rei?) ondents) 2 and under 23 3 -5 6 and over Total 68 38 54 loo 8 100 Average 3-6 3.2 Total respondents Non-respondents 256 0 13 0 Number in Family. The average number of persons in the family was 3*6 for those respondents to the personal interview who had telephones 38 compared to 3*2 for those respondents who had no telephones (Table 17)* Type of Potato. Michigan potatoes were used by 77 percent of the respondents without a telephone and by 44 percent of those having tele­ phones* Equally great differences can be seen for Maine potatoes* There were 30 percent of those with telephones who used I&ine potatoes while there were only 8 percent of those without telephones who were Maine potato users* Both these differences are statistically signifi­ cant at the *01 level. There was no significant difference in the percentage of those who reported the use of Idaho potatoes (Table 18)* Table 18. Type of Potatoes Used as Related to Whether or not Family has Telephone TVne of notatoes Without telephone With telephone (Percent of respondents) Michigan Maine Idaho Other Total 44 30 23 — 1 100 100 Total respondents Non-respondents 248 8 13 Tyne of Store. 77 8 15 - 0 There was no significant difference between the respondents possessing a telephone and those not possessing a telephone in regard to the type of store where they bought their potatoes (Table 19)* Number of Blocka to Store. The average number of blocks from resi­ dence to retail store was 5*0 for those respondents who possessed a tele­ phone and 3*8 f°r those not having telephones* This difference was not 39 statistically significant, (liable 20)* Table 19* Type of Store in Which Respondents Shop as Related to Whether or not Family has Telephone Tsme of Store With telephone Without telephone (Percent of respondents) Super chain store Other chain store Independent store Peddler, Roadside, etc* Total Total respondents Non-respondents Table 20. 55 67 8 1 100 8 25 0 100 247 9 12 1 30 Number of Blocks to Store as Related to Whether or not Family has Telephone Number of blocks With telephone Without telephone (Percent of respondents) 2 and under 46 30 3 “5 26 11 8 Over 17 Total 100 20 20 20 0 S 100 Average 5*0 5*8 Total respondents Non-respondents 222 10 34 3 6-8 9-12 13 - 16 4 TVpe of Area Where Store located. About one-half of the respond- ents having telephones and only 18 percent of those not having telephones shopped in neighborhood stores* Shopping was done in stores lr. ted in 4o minor shopping areas by 37 percent of the respondents who hah tele­ phones and 73 percent of those who had no telephones* In both case3, the H * test indicated the difference to be significant at hie *01 level* Approximately 10 percent of both types of respondents shopped in stores located in major shopping areas (Table 21)* Table 21* Type of Area Where Store was Located as Related to Whether or not Family has Telephone Tvne of area With telephone Without telephone (Percent of respondents) Major shopping area Minor shopping area Outside of either Total 11 9 37 -32 73 100 100 Total respondents Non-resp ondents 218 38 11 2 Summary - Comparisons of the Answers Given by Those Respondents to Personal Interview Who had and Those Who did not have Telephones* Both groups traveled about five blocks to shop, shopped in the same type of stores, consumed an average of *35 pounds of potatoes per per­ son daily and averaged about 3*3 persons to the family* Hone of the differences found in the above cases were significant. There were significant differences in regard to location of re­ tail store and type of potato used* More than half of those with tele­ phones shopped in neighborhood stores while over 70 percent of those without telephones shopped in stores located in minor shopping area* Both of these differences are significant at the *01 level* 41 Those without telephones contained a significantly hi^ier per­ centage of Michigan users and a significantly lower percentage of Maine or Idaho users than did those who ted telephones. It nay be that the people who did not feel like, or could not afford, to spend money for a telephone also felt that it was not worth while to pay a higher price for Maine or Idaho potatoes. This may eaplain why a significantly higher percentage of them purchased Michigan potatoes. Comparison of the Answers Obtained in the Telephone Interview with Those Given by Respondents to Personal Interview Who had Telephones. The answers obtained on the telephone interview were compared with those given by the respondents to the personal interview who had tele­ phones. This section deals with the results of this comparison for various factors. Table 22. Consumption of Potatoes by Respondents to Telephone Interviews Compared with Respondents to Personal In­ terview that had Telephones Pounds per Person ner dav .... .... _ With telephone Telephone Personal interview interview (Percent of respondents) 48 30 51 28 11 JU 100 12 — 2 100 Average o.37 0*37 Total respondents Uon-respondents 252 290 4 18 Up to ,29 0.30 - 0*49 O .50 - 0.69 0.70 and over Total Consumption of Potatoes. Those interviewed by telephone consumed an average of 0.35 pound of potatoes per person per day while the respondents 42 to the personal interview who had telephones consumed 0*37 pounds daily* Ihe slight difference in averages was not statistically sig­ nificant (Table 22)* Number in Family. The respondents to the personal interview questionnaire who had telephones had about the same number of persons in the family as did the respondents interviewed by telephone* Ihe average for the farmer was 3*5 persons while the latter was 3*6 per­ sons (Table 23). The difference in averages was not statistically significant• Table 23. Number in Family for the Respondents to the Telephone Interview as Compared to the Respondents to the Per­ sonal Interview who Had Telephones Number in family With telephone Telephone Personal interview interview (Percent of respondents) 2 and under 3-5 Over 6 Total 23 68 27 — a 100 100 Average 3.6 3.5 Total respondents Non-resp ondents 256 305 0 3 65 Type of Potato* The respondents to the telephone interview and those respondents to the personal interview who had telephones com­ pared very favorably in type of potato they used. About 40 percent of both groups were Michigan potato users while 30 percent used thine potatoes* Idaho users made up about 27 percent of the respondents. There were no significant differences in the type of potatoes reported 43 used by those with telephones who were questioned by personal inter­ view and those included in the telephone interviews (Table 24 )• Table 24* Type of Potato Used by Respondents to Telephone in­ terviews as Compared with those Respondents to the Personal Interview that had Telephones Telephone With telephone interview Personal interview (Percent of respondents) Tvne of potato Michigan Maine Idaho Others Total 44 30 23 -J. Total respondents Non-resp ondents 247 9 Table 23a 100 39 29 30 2 100 287 21 Type of Store Where Shopping Was Done by Respondents to Telephone Interviews as Gordered to Those Respond­ ents to the Personal Interview That Had Telephones Type of Store Super chain at care Other chain store Independent store Peddler, roadside, etc* Total Total respondents Non-respondents. With telephone Telephone Personal interview interview (Percent of respondents) 55 100 43 11 3* 10 100 247 9 291 17 8 30 -2 Type of Stare. When ccazparing those respondents to the personal interview who had telephones, it was found that 43 percent of them shopped in super chain stores while 35 percent of the respondents to the telephone interview did likewise. This difference was signifi­ cant at the *05 level. 3h both cases about 9 percent of respondents shopped in "other* chain stores, about 30 percent in neighborhood stores and about 9 percent freon miscellaneous sources such as farmers, peddlers and road­ side markets (Table 255. Table 26* Number of Blocks Traveled in Order to Buy Potatoes by the Respondents to the Telephone Interview Com­ pared with Those Respondents to Personal Interview That Had Telephones Number of blocks With telephone Telephone Personal interview interview (Percent of respondents) 2 and under 4,6 40 3-5 26 11 8 22 12 6 6-8 9-12 13 - 16 4 4 Over 16 Total 100 16 100 Average 5.o 7.3 Total respondents Non-respondents 222 232 76 34 Number of Blocks to Store, Those respondents to the personal in­ terview who had telephones traveled on the average 5»0 blocks tc The corresponding figure for those interviewed by telephone was 7.3 blocks* Results of the •t® test indicate that the difference in the averages is statistically significant at the .01 level. Much of the difference can be ascribed to the high percentage of those interviewed 45 by telephone who reported that they shopped in stores 16 or more blocks away frcta their residence (Table 26). Summary - Comparisons between Telephone Interview and those Respondents to the Personal Interviews who had Telephones. swers given by the two groups ccopared very favorably. The an­ Each group averaged about 3.5 persons per family and consumed, on the average, about 0.4 of a pound of potato per person daily. Their habits in regards to type of store they shop and kind of potatoes they use are almost identical, as judged by the percentage distribution of the re­ spondents. None of the differences were significant for any cf the above. Insofar as number of blocks between the store and the residence, the ones interviewed by telephone averaged 7.3 blocks while the re­ spondents to the personal interview averaged 5*0* was significant at the .01 level. This difference The findings here concur with those reported on page 25 where a significant difference was found between results of the personal and the telephone interview. Leaving out the respondents to the personal interview that had no telephones did not ad ter this relationship. 46 CHAPTER V suR var F ii’iDiiiQij - caeuKER p r a c t ic e s a ;d o p in io n s This Chapter discusses findings that are irapcrtant primarily from the standpoint of potato marketing. It is divided into two parts. Part I deals with relationships that, when tested by analysis of vari­ ance, failed to show any significant differences. These relationships are presented in summary tables in the text while the actual tables and analysis are to be found in the Appendix* Consumption Habits and Consumer References for Potatoes Consumption. The average consumption of potatoes by the respond­ ents to the study was about O .36 pounds per person per day. The method of gathering the data, income level (Appendix tables 24 and 25) and kind of potatoes (Appendix table 26) had no significant effect on this average. Table 27. Income level Income level as Belated to Number of Pounds per Person per Day. Number of respondents Low Medium High Total 53 636 m 332 Non-respondent s 144 Average pounds 0.37 0.35 0*31 0.36 Is the Kind of Potatoes on Hand the Sams as That Usually Bought? For eighty-three percent of the respondents the kind of potato they had on hand (Michigan, Maine, Idaho, etc.) were the kind they rep ortad th:. 5 they usually bought (Table 23). 47 Table 28♦ Answers to Question, Is the Kind of Potatoes on Hand the Same as That Usually Bought? Number of Respondents Yes No 843 J2k Total 101? Non-r esp ondents 48 Percent of Respondents 83 JZ 100 A similar ratio was found by each method obtaining data, and for respondents in all income areas (Appendix tables 1 and ?)• Type of Potatoes on Band* When asked what kind of potatoes they had on hand, the answers were distributed in the following manners Michigan, 42 percent? Maine, 27 percent; Idaho, 29 percent; and other, 2 percent (Table 29), Table 29* Type °T Potatoes on Hand Number of Tvne of Potatoes_____________ Respondents Michigan Maine Idaho Other Total Non-respondents 425 269 292 25 1011 Percent of Respondents 42 2? 29 2. 100 54 No significant difference existed in this proportion when deter­ mined by the three methods where they were tested at different area levels of inccsue (Appendix table 3 and 21), type of potato usually 48 bought (Appendix table 1), type of stare (Appendix table 27)* location of store (Appendix table 12)t type of container (Appendix table 2), and size of family (Appendix table 10)* Pogg- Grocer..Usually Have the Tyne of Potatoes Desired? About nine out of every ten respondents said that their grocers usually had the type of potatoes they desired (Table 30). Table 30* Answers to Question, Does Grocer Usually Have Kind Desired? Number of Heso ondents Percent of Respondents 846 100 Yes No 89 11 100 Total Noa-resp ondents 119 This was true regardless of method of obtaining information, in­ come area level of respondents (Appendix table 6)* Table 31. Kind of Container for last Purchase Kind of Container Bulk Packaged Total Total respondents Non-respondents Percent of Respondents 39 Jk 100 1,028 37 Kind of Container Potatoes Purchased in. Sixty percent of the respondents purchased packaged potatoes the last time* Forty percent 49 of the reap ondents had purchased their potatoes from bulk lots (fable 3D . When tested statistically it was found that none of the following had any significant effect on the percentages expressed above; (1) method of gathering data, and area income level (Appendix table 4 and 9)» (2) kind of potatoes used (Appendix table 2), (3 ) location of store (Appendix table 11), and (4) number in family (Appendix table 1$), Reported Average Cost per Pound - Effect of Type of Potatoes* The average reported cost per pound paid for Michigan potatoes was 4*0 cents, for Maine it was 4.8 cents and for Idaho 5*6 cents. Results of the wt* tests indicated that the differences between the averages for any two of the above are significant at the ,01 level. The average cost did not vary significantly between the three methods (Table 32). A chi-square test on the numerical distribution of the respond­ ents yielded a figure close to significance at the *05 level. However, this was not great enough, to invalidate the analysis of variance* Table 32. Comparison of Methods of Obtaining Ehta and Type of Potato to Cost per Pound Tvne of Potato Michigan Maine Idaho Average Total respondents Non-respondents Personal Telephone (Average cents) Mail 4.7 4.2 4.5 5aZ 4.7 3.9 4.9 5,.?.4 4*6 191 86 131 177 334 146 4.1 4.7 5u6. Average 4.0 4.8 5JL 4*6 50 The results obtained are in line with the actual situation usually found in Detroit stores. Michigan fanners, since they have the advan­ tage of location, can sell their potatoes for less and still receive the same or greater net farm price. Maine and Idaho fanners have to transport their potatoes great distances and can only afford to send their very best potatoes. In order to be able to ship their product a long distance, they must get a higher price for their product. The table and analysis bear this out. The average cost per pound of potatoes for all the respondents to the study was 4»& cents. This figure was not significantly affected by method or inecme area (Appendix table 22 and 23), kind of store (28), or number in family (Appendix tabic 29). Table 33* Type of Store Tfliere Potatoes Were Purchased Tvne of Store Chain Independent Roadside, farm, peddler Total Non-respondent Number of Respondents 654 Percent of Respondents 67 250 25 J& 984 _8 100 81 Tvne of Store Where Potatoes Purchased. It was found that 6? per­ cent of respondents made their last purchase of potatoes from local or national chain stores while 25 percent of them purchased from independ­ ent retail stores and 8 percent from other sources such as roadside markets (Table 33)* This relationship held regardless of rental area (Appendix table 2 1 ) , lo c a t io n o f s to re (Appendix ta b le 1 2 ), type o f c o n t a i n e r (Appendix 51 table 9)» number in family (Appendix table 16), or type of store (Appendix table 28)* That is to say, none cif the above variables had any significant effect on the relative percentages stated above. Location of Store Where last Purchase of Potatoes Was Made, Of the respondents to the study, 45 percent of them shopped for potatoes in stares considered to be in a minor shopping area. Approximately the same percentage shopped in a neighborhood store located outside a shopping area, as defined by the Detroit News classification of 1948* Only 11 percent of the respondents had made their last purchase of potatoes fraa a store located in a major area (Table 34)* Table 34* Location of Store Where Potatoes Purchased Number of Respondents Tvne Shoot*in#. Area Major shopping area Minor shopping area Outside shopping area Total Non-resp ondents 92 Percent of Respondents 11 386 45 850 Jtk 100 215 The relationship was not significantly different regardless of the method of collecting data when compared by the following classifications? area income level (Appendix table 20), number of people in family (Appendix table 19), and type of potatoes on hand (Appendix table 12), TflnTiVhttr ip Fflmilv. The average number of persons in the family was 3*6, The following variables did not cause any significant difference from this average: area income level or method (Appendix table 5 and 8), 52 kind of store (Appendix table 16), cost per pound (Appendix table 29)# kind of potatoes on hand (Appendix table 10), and location of store (Appendix table 19), Table 35* Inc one Level as Related to Number in Family Number of Respondents Income Level Low Medium High Total 613 3.8 3.6 251 34 922 3.6 54 Non-resp ondent s Average 143 Summary - Those Relationships Showing No Significance. There were no significant differences between methods of obtaining the data for any of the followings 1, Number in Family. The average number of persons in the fam­ ily was 3 *6. 2. Consumption, Average consumption of potatoes by the respond­ ents to the study was 0«3& pounds per person per day, 3« Type of Potatoes on Rand. Among the respondents to the study there were 42 percent Michigan users, 27 percent Maine users, and 29 percent Idaho users* 4* Purchasing Habits, When asked whether the potatoes they had on hand were the kind they usually bought, 83 percent of the respondents answered "yes'1 and 17 percent "no®, 5 . Satisfaction with Choice of Potatoes in Store* Nine out of ten respondents stated that their grocers usually had the kind of potato uhsy desired* 53 6* Type of Store Purchased from. There were 69 percent of the respondents to the study who shopped in chain stores, 25 percent who shopped in independent stores, 7® Store Location® As far as location of the stare where shop­ ping was done, 11 percent of the respondents shopped in steeres loca­ ted in major areas, 45 percent in stores located in minor areas and 44 percent in stares located outside a shopping area. 8® Kind of Container® Sixty percent of the respondents purchased packaged potatoes the last time they had shopped and4° percent pur­ chased their potatoes from bulk lots. Respondents* Quality Ratings of the Potatoes They Had on Hand The respondents were asked to rate the potatoes they had on hand in regard to certain factors. age", or "poor®. They were to rate them as "good®, "aver­ No attempt was made to define or have them define the three ratings. The percentage of the users of a particulartype of potato who rated it "good," "average® or "poor" wascalculated.A "t* test was used to compare differences in the percentage of respondents who rated their potatoes as "good®. have been compared. Any of the three percentage figures could Comparing all three in turn seemed unnecessary and confusing since each percentage was dependent on the other two. The actual tables, from which the summary table presented in the text is drawn, are included in the Appendix. Cooking quality. Only 66 percent of the Michigan users rated their potatoes as "good" for cooking quality while there were 84 percent of the Maine users and 88 percent of the Idaho users v/ho gave their pota­ toes this highest rating. The percentage figure for Michigan users was 54 significantly lower at the .01 level than that far the other two. Qolor after Cooking. The percentage of users of Michigan pota­ toes who rated their potatoes "good" was significantly lower at the •01 level than that far either the Maine or the Idaho potatoes. Michi gau potatoes were rated "good® for color after cooking by 59 percent of their users compared to 83 percent of the Maine users and 95 per­ cent of the Idaho users. The percentage for Idaho potatoes was significantly higher at the *01 level than that far the other two* Table 36 • Percentage of the Respondents Who Gave Their Potatoes a Rating of "Good" for Various Quality Factors* Type of : sColor : sBlesn- T T tGeneral» Potato &: CookingsAfter t 1ishes &s »Clean- sDesir- ; Rating *QnalitysCookings Taste?Defectss Sizeilineassabilitys Average (Percent of respondents who rated) Michigan Good 66 59 71 47 54 59 55 59 Maine Good 84 83 82 12 63 74 81 77 Idaho Good 88 95 91 78 77 88 87 86 Taste. The percentage of the Michigan users who rated their pota­ toes as "good" for taste was significantly lower at the .01 level than the corresponding percentage for Maine and Idaho users. There were ?1 percent of the Michigan users, 82 percent of the Maine users and 91 percent of the Idaho users who rated their potatoes as "good*. The average percentage for Idaho potatoes was also significantly higher at the .01 level than that of the Maine users. * For a c tu a l d a ta see .Appendix ta b le s 30 * 3 6 55 Ble^a&ea_and Defects* Michigan users again gave their potatoes the lowest rating* Only 47 percent of the respondents rated Michigan "good" while 72 percent of the Maine users and 71 percent of the Idaho users gave their potatoes this hifjiest rating. The percentage for the Michigan potatoes was significantly lower than that for Maine end Idaho potatoes. When asked how they rated their potatoes for size, 54 per­ cent of the Michigan users, 63 percent of the Maine users and 77 per­ cent of the Idaho users rated their potatoes as "good". 'Hie percentage for the Michigan users was significantly lower at the .01 level than that for both the Maine and Idaho users. Cleanliness,. Michigan potato users were least satisfied with their potatoes in regard to cleanliness. Only 59 percent of them rated them, "good" while the comparable figures were 74 percent for Maine and 88 percent for Idaho. Hie percentage figure for Michigan potatoes was significantly lower at the .01 level than that for either thine or Idaho potatoes. Hie difference in percentages between the Idaho and Maine was also significant at the .01 level. General Desirability. The respondents were also asked to rate their potatoes in regard to ^general desirability*. Hie objective, in this case, was to attempt to have the consumers rate their potatoes but not on any one characteristic alone* own evaluation of their potatoes. Instead, they were to give their No attempt was made to define for them or to have them define what characteristics, or to what extent certain characteristics, influenced their rating. Hie result here verified the previous individual findings. Michi- 56 gaa users again gave their potatoes the lowest rating. Only 55 per­ cent of them rated Michigan potatoes "good® compared to 81 percent of the Maine users and 87 percent of the Idaho users. Here again, the percentage for Michigan was significantly lower at the .01 level than that for either the Maine or Idaho potatoes. Results of a Simultaneous Comparison of the Seven Factors Dis­ cussed. The average percentage for each type of potato was derived from the table 36. There percentages were 59 f°r Michigan, 77 for Maine and 86 for Idaho* In analysis of variance was run on the table to see if these differences, in average percentages, were significant. Results of the analysis of variance indicate that the average percent­ age of the users of Michigan potatoes who rated them *good" was sig­ nificantly lower at the .01 level than the corresponding percentage of Maine and of Idaho users. The percentage of Idaho users who rated the potato "good® was significantly higher at the .05 level than that of Maine users (Appendix A). It can be concluded that the users of Michigan potatoes were least pleased while those who used Idaho were most pleased. All this based on what percentage of the users rated them "good". Factor Considered Most Innortaut. The. mail and the personal inter­ view respondents were asked which factor they considered most impor­ tant. Cooking quality, taste, and color after cooking were given most frequently by the respondents to the personal interview. 3h the case of the mail questionnaire over 50 percent of the respondents gave more than one factor as being important. Of these combinations cooking-quality qnd taste? cooking-quality and color after cooking; and taste and color after cooking were given most often. 57 Stable 37* factors Considered Most Important Quality Mail Personal (Percent) Cooking quality Taste Color after cooking Blemishes and defects Cleanliness General desirability Size Combination of above Total ..-Z loo Total correspondents Non-r esp ondent s 249 38 34 34 12 4 4 3 2 27 9 5 5 1 l * -52 100 333 144 * Less than one-half of one percent* Rnting Given bv Resnondents to the Potatoes They Had on Hand When Prepared in Various Ways. The respondents to all three questionnaires were also asked to rat® their potatoes as ■Good** *Average *, or "Poor* when prepared in differ­ ent ways. This section reports the results. Boiled. For this method of preparation, 83 percent of ihe Maine users and 68 percent of the Michigan and the Idaho users rated their potatoes ©s good. The percentage for Maine was significantly higher at the .01 level than for the other two. . The percentage of the users who rated Michigan potatoes •good* was significantly lcwer at the .01 level than either of the other two. For Michigan 71 percent of their users rated them "good** while the comparable percentages were 86 percent for the Maine users and 83 percent for the Idaho users. Fried. For this method of preparation all three types of potatoes 58 were given about the saiae rating by their users* About 70 percent of the users of all three types of potatoes rated them "good". There was no significant difference in the ratings between any two types of potatoes# Table 38# Type of Potato and Rating Percentage of Respondents Who Rated their Potatoes "Good® when Prepared in Various Ways* Manner in which prepared Boiled, Mashed Fried Baked (Percent of respondents v/ho rated) Average Michigan Good 68 71 71 51 65 Maine Good 83 86 73 61 76 Idaho Good 68 83 76 98 81 Baked* Idaho had highest rating for this net hod of preparation since 38 percent of the Idaho users rated their potatoes as "good" while the comparable figures were 6i percent for Maine and 51 percent for Michigan. 'Hie percentage for Idaho was significantly higher at the .01 level than that of the other two. The percentage of Maine users (6l percent) who rated them good for halving was significantly higher at the .01 level than that of the Michigan potatoes (51 percent). Results of a Simultaneous Comparison of the Four Methods of Prep­ aration. Maine potatoes were rated highest of the three for boiling, and got about the saiae rating as did the Idaho when the respondents * For a c tu a l d a ta see Appendix ta b le s 37-40 59 served them mashed* Idaho potatoes were rated tops for baking* All three types received about the same rating when it came to serving them fried* The average percentage for each type of potato was again derived from Table 38* These percentages were 65 for Michigan, 7& for Maine and 81 for Idaho, An analysis of variance was run on the table to see if these differences, in average percentages, were significant. Results of the analysis of variance indicate that the difference be­ tween the three averages was not significant, Thu3, when considering the average rating given the three types of potatoes for the various methods of preparation, Michigan potatoes compared a little more favorably with the other two types* This result was at variance with that obtained when "cooking quality® alone was considered (Table 36)• Actually,the variation in the results is due to the difference in the iwo questions and also the difference in the treatment of the data. "Cooking quality® was not defined in any way and thus became a subjective valuation based on each respondent’s ideas and experiences. Analyzing the results of this subjective valuation, by using a ®t® test on the percentages of the users of each type of potato who rated them "good", it was found that Michigan potatoes were given a significantly lower rating than that for the other two types of potatoes. The result of an analysis of variance, when the ratings for the four methods of preparation were considered simultaneously, indicated no significant difference between the three types of potatoes. What the analysis of variance compared were the averages derived from cca­ bining four separate valuations. Actually, when each method a£ prepa­ 60 ration was considered separately, there were significant differences and these were indicated in the preceding discussion. Considering all four ratings simultaneously allowed far the averaging out of these differences and explains the recurringly contradictory results. 61 CHAPTER 71 IMPLICATIONS CF THE STUDS' Th© ©concxuist is continually searching for better and. less expen­ sive ways of obtaining data, particularly as it relates to the action of consumers or consuming units. It is costly and difficult to obtain representative and reliable data about consumers and their consumption habits, beliefs and opinions* done* More research in this area should be It would increase the value and reduce the cost of a major research need* The present study was undertaken with the twofold objective of comparing the methods and of determining some consumer preferences and actions as they relate to potatoes* The results will be discussed under these two headings* Methodology 3n the present study, a section of Detroit ccn^rising about 12 percent of the dwelling units listed in the 1940 census was sampled* The area, containing 50,000 dwelling units, was racially homogeneous* It was of above average income, as compared to the rest of the city and had a certain degree of economic homogeneity* This is evidenced by the fact that though the range in listed 1940 rentals was from $14 to over $110, 75 percent of the respondents had rentals varying frcsa $30 to $59 * A. chi-square test was used to test for differences between the samples obtained by the three methods* The samples were broken down according to rental groups based on the 1940 census. When all three methods were considered simultaneously there were no significant differences between the sampling methods. It was this fact that 62 suggested the analysis of variance as a way to test for differences between similar information obtained by the three methods* % testing only two of the three methods at a time, it was found that the differences between telephone and personal were significant at the ,05 level* There were no statistically significant differences between any of the other combinations of two methods* The data obtained by the three methods was also comparable* Jh only one case, that of number of blocks traveled to shop, did the analysis of variance indicate any significant differences between the results obtained by the three methods* For such variables as consump­ tion, number in the family, type of store, and type of potatoes bought, there were no significant differences* Thus, it is possible to assume that when sampling a relatively homogeneous area or population, certain types of information can be obtained with about the same degree of accuracy, by any of the three methods* Other factors need to be considered along with the comparability of the results* The most important of these factors is the cost* The average cost per completed questionnaire was $2*72 for the personal interview, 54 cents for the mail interview and 44 cents for the tele­ phone interview* The cost for the mail interview was relatively high because of the very small return obtained by this method* Jh terms of the percentage of the original sample, 97 percent of the personal interview, 78 percent of the telephone interview and 4 percent of the mail interview was obtained* The high return for the first two can be partly ascribed to making call-backs including some interviewing at night. The very low return fcxc the mail questionnaire was due principally to the absence of any special appeal or rewards for answering the questionnaire. Faulty distribution of the questionnaires by some store managers was also a contributing factor• Bren though some stares distributed all the questionnaires, the highest return from any one stare was about 8 percent• The telephone interview was completed in four days and required the full-time services of two persons# The personal interview required 12 days to complete using nine local part-time interviewers* The bulk of the mail questionnaires were returned within two weeks of the date they were placed in the stores* However, same of them were not sent in until three or four weeks after this date* It would seem that the greater speed with which the telephone interview was completed, made it possible to describe a given situation more accurately by using this method* For example, people interviewed on the first day may have been speaking about a situation quite different from the one existing when those interviewed eight days later were contacted* Thus, though they may both be answering the same questions, the answers may be different because each is speaking about a different situation* This difficulty is partially eliminated when the interviewing is confined to a short period* On the basis of the evidence presented above, the writs? feels that under certain conditions, telephone interview is preferable to either the personal interview or the m i l questionnaire* It is less expensive, faster, and more convenient than the personal interview. The percentage of the sample is much greater and the costs less than when using a mail questionnaire* 64 The telephone interview method is most efficient in a situation such as that found in this study where almost all the respondents owned telephones* Assuming the stratified randan sample of the area used for the personal interview to be representative of the area, it was found that 95 percent of the respondents to the personal interview had telephones* Furthermore, for certain variables such as consump­ tion, type of potato and number in the family, the same results were obtained when the telephone interview and those respondents to the personal interview who had telephones were considered* 3h other wards, the emission from the sample of those without telephones did not sig­ nificantly change the results* The telephone interview when used* (1) in an area where almost all the families have telephones, (2) for asking certain types of questions, and (3 ) in connection with a questionnaire that is not too long or too involved, offers a very satisfactory tool for consumer research* However, due to the restrictions listed above, the telephone interview is limited in its application and the other two methods can­ not be summarily dismissed* Both the mail questionnaire and the per­ sonal interview can be used for studies in areas where, using the telephone directory listings would not give a truly representative sample. The limit as to the type of question that can be asked and the necessity for brevity are drawbacks to the telephone interview that are not encountered when the other two methods are used. The higher coats, the longer period required to complete a given number of inter­ views and the bias introduced by using numerous interviewers should not be overlooked in using the personal interview. The mail interview can be used to gather certain types of marketing inform tion for which neither the telephone or personal interview are adequate. Thus, the problems encountered when any one method is considered are many. More research needs to be done using all three methods separately and in combination* More needs to be done in determining the comparability, reliability, and validity of the results obtained when the three methods are used* Until then any conclusion regarding the relative merits of each can only be tentative and quite limited in SOOpGe Marketing of Michigan Potatoes In conjunction with a nation-wide study conducted by the Bureau of Agricultural E c o n o m i c r e s u l t s were presented for three cities, Chicago, Boston and 3-os Angeles* Of these three, Boston has a market situation in regard to potatoes that most nearly approximates the con­ ditions found in Detroit* A ccaparison of the results reported for respondents living in Boston with results obtained in the present study will be made* The results reported for Boston included the followings (1) aver­ age personal consumption was 2*8 pounds a week, (2) 45 percent of the hrwAtimimrg bought the ir potatoes from bulk displays and 51 percent purchased packaged potatoes, (3 ) 72 percent stated that they were usually able to obtain the type of potato they wanted, (4 ) the most common reasons for selecting a potato were good taste, cooking quality and suitableness for several methods of cooking, and (5 ) 92 percent of 22/ U* 3. Department of Agriculture, op. cit> 66 the respondents considered quality mere important than price in purchas­ ing potatoes* In the present study, the average personal consumption was 2,5 pounds a week. There were 60 percent of the respondents who purchased packaged potatoes the last time they bought potatoes* Eighty-three percent of the respondents stated that their grocers usually had the type of potato they wanted. The factors considered most important in buying potatoes were cooking quality and good taste. All of these results compare closely with those reported above. Other findings of the present study included the fact that about 90 percent of the respondents stated that their grocers usually had the type of potato they wanted on hand. Almost as many stated that the type of potato on hand was the type usually bought* About 66 per­ cent of the respondents shopped in chain stares and 25 percent in inde­ pendent retail stares. Shopping was done in stores located in minor shopping areas by 45 percent, in stores located in a major shopping area by 11 percent, and in stores located outside a shopping area by 44 percent cf the respondents to the study* Michigan potatoes were used by 42 percent, Maine by 2? percent and Hah© by 29 percent of the re­ spondents. Average weekly consumption for the respondents to the study was 2*5 pounds a week and there was an average of 3.6 persons in the family. Thus, approximately 9 pounds of potatoes were consumed weekly per family, The average reported price by the consumers for potatoes during the period the survey was taken was 4*0 cents for Michigan, 4*8 cents for l&ine and 5*6 cents for Idaho. Translated to a per family basis, this meant that, on the average, Michigan users spent 36 cents, Maine users 6? 43 cents and Idaho users 50 cents weekly for potatoes during the period of the study* In spite of the differences in these reported prices paid for potatoes, less than half of the respondents reported that they pur­ chased Michigan potatoes, This would seem to indicate that the price of potatoes is not the most important factor in determining which type of potato is purchased. It agrees with the findings of the U, S. Department of Agriculture study previously presented in which the re­ spondents considered quality much more important than price in purchas­ ing potatoes. The average ratings for various characteristics given by the users of each type of potato bears out this relationship. When the respond­ ents were asked to rate the potatoes they used for such characteristics as cooking quality, color after cooking, taste, size, blemishes and defects, and cleanliness, the users of the Michigan potato gave them the lowest rating in every case with the Idaho usually havirg the h i p ­ est rating. The average rating for Michigan potatoes, based on the average of all the above characteristics was significantly lower at the ,01 level than that for the other two, Michigan potatoes compared a little more favorably with the other two types when prepared in vari­ ous ways. These findings point out two ways in which the Michigan potatoes might be improved. potatoes* There is a need for better grading and handling of This is based on the fact that the users of Michigan pota­ toes, as reported in Chapter V, gave their potatoes a significantly lower rating far such characteristics as blemishes and defects, cleanli­ ness and size, than did the users of Maine or Idaho potatoes. Secondly, there is rocaa for improvement in the quality of the Michigan potato. Again, this is based on the respondent’s ratings for the potatoes they had on hand* Michigan potatoes were given a signif­ icantly lower rating by their users than were the ifaine and Idaho po­ tatoes for such characteristics as cooking quality* color after cook­ ing and taste* A program such as that proposed in the Michigan Seal of Quality 23/ Act of 1949» had it been adopted, would have been helpful in injproving the marketing of Michigan potatoes, ftiis act provided for volun­ tary participation by the growers and for establishment of grades at least as high as the Federal grades. It was to be supervised by the director of Michigan Department of Agriculture, Certain requirements made of members were stated as were penalties for infractions. The grades themselves were to be set by members of the commodity co mit tee appointed by the director* There was to be a committee fear each com­ modity. As can be seen, the Act had its shortcomings, Ihese included the voluntary nature of the program, the rather light penalties far infract­ ions and the rather indefinite provisions for the establishment of grades, and fees to be charged. However, as a first effort in the field it had much to recommend it. The writer feels that efforts to initiate such a program should be continued® Kie second area in which Michigan potatoes might be improved has to do with bettering their quality. As was seen in the preceding dis- 23/ Michigan Seal of Quality Act, House Bill No, I92, Michigan 65th Legislature, 1949* 69 cussion, consumers usually consider quality more important than price when purchasing potatoes* Ihere i3 generally believed to be a direct relationship between yield and quality of potatoes. Accepting this premise, the Department of Field Crops at Michigan State College initiated and sponsored the "300 Bushel Club". years* This venture has been carried on for the past 25 Its aims are to encourage the production of high yields and together with this, better quality potatoes. Membership is gained by having an average yield of 300 bushels on five tested acres. Judging by the ever-increasing number of members, the program is a successful one. Another phase, dealing with improving the quality of Michigan potatoes, has to do with the greater use of certified seed. Here again, the Department of Field Crops has taken a leading role, both in certifying seed and in encouraging its use by the growers. The actual and proposed programs presented above are by no means panaceas nor does their adoption mean a complete solution to the problems. Actually, the problems involved in the marketing of pota­ toes are maryand complex. No one study on any one phase can hope to do more than offer limited information and suggestions for future re­ search* Before any very definite answers can be interpreted in terms of a positive policy, research far beyond the scope of this study most be undertaken* 70 CHAPTER VII SUMMARY AND CQNCLUSIOIS The objectives of the study weres 1, To compare data obtained by the three methods, personal inter­ view, mail questionnaire and telephone interview* 2* In conjunction with the above, to obtain information regarding consumer preferences and opinions about the potatoes they use. The study was carried out in the northwest portion of Detroit from 6th to 18th of December, 1948, For the personal interview a stratified random sample based on rental areas according to the 194° census was used. The sample for the telephone interview was drawn from the Detroit telephone directory taking every 7^h listing on every 50th page for the exchanges found in the area. Hie mail questionnaires were distributed through retail stores in the area, selected on the basis of size, sales and location. The percent return for each method was 97 percent for the personal interview, 78 percent for the telephone interview and 4«2 percent for the mail questionnaire. Si the matter of cost per completed questionnaire it was $2.72 for the personal interview, 44 cents for the telephone and 54 cents for the mail questionnaire* Analysis of variance was used for the simultaneous comparison of data obtained by the three methods. For other types of comparisons, two variations of the *t® test were employed, one to compare differences in percentages and another to compare differences in weighted averages. These included relationships which also had methodological importance such as time the interview was cample ted and number of call required 71 to obtain an interview. The following are the most pertinent results obtained in the study* 1* Methodology. A chi-square test on tlie distribution of re­ spondents among the various rental groups failed to show any signifi­ cant difference when the three methods were considered simultaneously* However, there was a significant difference between the samples used for personal and telephone interviews when these two alone were con­ sidered. This did not invalidate the use of analysis of variance which was based on the simultaneous analysis of results obtained by the three methods. 2. The analysis of variance indicated a significant difference, both between methods of obtaining the data and rental groups insofar as nunber of blocks traveled to shop. When considered in conjunction with rental area the respondents to the personal interview averaged 4*3 blocks, tho»e to the telephone interview averaged 7*2 blocks and those to the mail questionnaire averaged 4*8 blocks* The average far each method was significantly different from the other two* Similar differences were found in regard to rental area although there was no visible trend. Those with rentals over $70 both averaged about 7 blocks. Those with rentals between $40-$&9 all traveled a significantly lower number of blocks. The average number of blocks varied significantly according to which of the Detroit News income areas (classified 1948) the respondents lived. Those living in low income area averaged 9.3 blocks, those in medium income area averaged 6,0 blocks and those in high inccsae area averaged 5*0 blocks. 3 * The cost per pound of potatoes varied significantly with the 72 type of potato* Idaho potatoes cost the respondents an average of 5*6 cents per pound, for hhine it was 4*8 cents and for Michigan it was 4*0 cents* The difference in average cost between any two of these was significant at the .01 level. 4* Analysis of variance run on combinations of the following failed to reveal any significant difference between the methods of obtaining the data and* (a) type of potatoes, (b) type of container, (c) whether the type of potato on tend is type usually bought, (d) whether grocer usually has type of potato desired, (e) coat per pound, (f) number in family, (g) consumption, (h) type of store, (i) type shopping area store located, (j) number of blocks traveled to shop, (k) rental area, and (1) income levels, 5* Comparisons were made of the percentage of the users of each type of potato who rated them "good" for characteristics such as color after cooking, size, taste, blemishes and defects, and cooking quality. Almost invariably Idaho ranked first, Maine second and Michigan third. For many of these characteristics the differences in the percentages between the three types of potatoes were significant. Considering all the above factors simultaneously, the average percentage for Michigan was significantly lower, at the .01 level, than that of Maine or Idaho. At the same time,the percentage for Idaho was significantly higher -=t the .05 level than that for Maine. Conclusion. Hie results of the study indicate that it is possible to obtain comparable information by the three sampling methods. Some of this ccomparability can no doubt be explained on the basis of the rather homogeneous group and limited area to which the study was confined. One shortcoming of this study was the Iterated number of similar 73 questions asked on all three questionnaires. Part of this was due to the fact that it was felt that seme questions might be answered on one type of questionnaire but not the other. The varying lengths of the questionnaires in conformity with the manner of obtaining the interview was also a handcap. The necessity for expressing each question in vari­ ous ways according to the type of interview may also be considered a shortcoming. Another shortcoming of the study was the fact that the 1%0 census information used as basis for sampling and analysis was not representa­ tive of the actual situation in 1948 and adjustments had to be made. A third shortcoming was the fact that the limitation of funds made interviewing of non-respondents impossible. The writer recommends the use of the telephone interview far samp­ ling an area where most of the families have telephones and where cer­ tain types of information are sought. This method was found to give results comparable with those derived from using a personal interview or a mail questionnaire. Hie return from the telephone interview was quite satisfactory and it ted the lowest cost per completed question­ naire. The telephone interview was also completed faster than the other two. However, the telephone interview has its drawbacls and the other methods cannot be discarded. Resultg of the subject matter portion of the study indicated that Michigan potatoes were ranked lowest by their users when compared to Maine and Idaho for certain characteristics such as blemishes and de­ fects, size, cooking quality, color after cooking and taste, among others. The conclusion reached from a discussion of the above and other facts is that Michigan needs to improve the marketing of its 74 potato crop if it is to compete successfully on its own m r ’.i-ts potatoes grown in Maine and Idaho* v . ith Two suggestions arc pro-used as means of improving the marketing of Michigan potatoes* with improving their grading and handling. One das to d , The otrier suggestion calls for a continuance of programs designed to encourage increased yields and by so doing, improving the quality of the potato. There is a need for further research dealing with botn phases of this study* Research dealing with methodology or with consumer prefer­ ence supplies information that is useful in understanding some of our potato marketing problems. The data obtained in this study, together with information on both phases supplied by further research, will go a long way toward helping to explain many of our economic relationships. BIBLIGGPiATIlY Baxter, R* E, "Use Both Mail-Type and Personal Interview in Reader­ ship Research," Printer*a 3hk~ 1947, 219*19-22. Bennett® A.5. 221*70-72. "How Good Are Local Interviewers*" Printer's ink. 1947. ------------ Bevis® J, C. "Economical Incentives Used for Mail Questionnaires," Public Opinion Quarterly« 1948, 12*492-493. Blankenship® Albert« How to Conduct Consumer Oiinion ResearchT__Harper and Company, 194&* Brooks® Y. "Can You Trust Mail Questionnaires," Printers Ink. 1947. 220*34-36 Brown, G. II. "A Comparison of Sazanling Methods," Journal of Marketing, 1947. 11*331-337. Brown, L. 0. Market Research and Analysis. Ronald Press, 1937• Calahan, D.» Y. Tamouris, and H, W. Verner. "Interviewer Bias Involved in Certain Types of Opinion Survey Questions," International Journal of Opinion and Attitude Research. 1947♦ 1*210*230. Clausen, J. A, and R. N. Ford. "Controlling Bias in Mail Question­ naires," Journal of American Statistical Association, 1947. 42s 497-311. Colley, R. H. "Don’t Look Down Your .Nose at Mail Questionnaires," Printer’s Ink. 1945» 210121 Cochran® W. G. "Use of Analysis of Variance in Enumeration by Sampling Journal of American Statistical Association. 1939* 34*492-311* Eastman, R. 0. "Boat’s About Mail Questionnaires," Printer's Ink. I943 202*24-29. Eckler, A. R. and E. P. Stuart. "Marketing and .Sampling Uses of Popu­ lation and Housing Data," Journal of American Statistical Association, 1943, 38*87-92. Edgerton, H. A., 3. II. Britt, and R* D. Norman. "Objective Differences Among Various Types of Respondents to a Hhiled-Questionnaire," American Sociological Review, 1947* 12*435-444* Franzen, li « and P. F. Lazarafeld. "Mail Questionnaire as a Research Problem, " Journal of* Psychology,-1943» 20 *239-320 • Hansen® Ivi. H® and W, E. Deming. "On Game Census Aids tc Sampling," Journal of American Statistical Association. 1943* 33*333-337* 76 Hansen, M. H, and W. 21. Hurwitz, "Problem of iTon-Hes~ onse in Sample Surveys," Journal .of American Statistical Assoc loti 1946, 41; 517-529. Hochstim, J. R» and M. S. D* Smith. "Area Sampling or Quota Control? Three Sampling Experiments,• Public Opinion Qpjrtariy. I943, 12s 73-80. Jacoby, Eleanor. *Interviewing Problems in Financial Survey," Inter­ national Journal of Opinion and Attitude Research. 1947# 1*54-58* Michigan Seal of Quality Act. House Bill No. 192, Michigan 65th Legis­ lature, 1949* Mitchell, Walter. "Factors Affecting the Hate of Return on Mail Questionnaires,*..Journal of American Statistical Association. 1939» Moore, Clarence C* "Increasing the Returns frcm Onestionnaires," Journal of Educational Research. 1941* 35*138-140• Ear ten, Mildred. Surveys. Polls and Samples. Harper and Company, New York, 1946* Robinson, Ray. "Five Features Helped this Questionnaire Pull from 60 to 70 Percent," Printer*s Ink. 1946, 214*25-26. Shapiro, 3. and J. C. Eberhardt# "Interviewer Difference in an Inten­ sive Interview Survey,11 International Journal of Opinion and Atti­ tude Research. 1947# 1*1-17* Snedecor, G. W. Statistical Methods. Applied to Experiments in Agri­ culture and Biology (4th ed»)« Iowa State College, 1948. U. 3. Department cf Agriculture, Bureau of Agricultural Economics, Potato Preferences Among Consumers. U. S. Department of Agriculture Miscellaneous Publication 667, 119 P*» 1948. U, S. Department of Commerce» Bureau of the Census, Sixteenth Census of I?. S«. 1940, Detroit Housing Supplement. APPENDIX A Statistical Appendix Includes explanation of statistical techniques used and actual analysis of variance and ®t* where si^iificance was indicated. 77 APTji'iDIX A METHOD OF 03rL.iHJHTG DATA Rental Area Personal * Telephone Under $30 32a 5.1b 22.5c 114.8a $30 - $39 47a 6.6b 292.4a ’ Mail 18a . 22.2c * 24a 29.3c 74-Ar 6.8-3 10.2b 226.4a • 5 .6b l64.1d 505.3-13 « 0 41a 43*9© s 58a 57.8c 146-Ar 6.3-B 922.2-3 7.5b 329.2d * 5 .2b 300.6a 44.3© 80s $40 - $49 3 . 1b & 76.2c 236.2d 77a 6.1b • 75*4© s 94a 99*4© 459. 9a » 3.8b- 377. 7a « 9 $30 . $39 ; 33a 35.5o 117.24 34a 6.0b 35*1© 1 50ft 210.6a 5 5 .8 b 1 3.3b e $6o - $69 5 7a 5 2. 9b * $70 & o ve r5 7a 5 6.7b 17a 34.3© 41.5d * 6.8b t *•. 13*3® J l? a 83a d • 10.6b | Average “ S 14.1© : 23a 95.9a 5 6.0b • 5 « 13*2c j 20a 139.9a i 4 . jb • 4*3Bp Total Average 5 ft : « 1073.8-0 263.5a 117-Ar 5.1-3 596.3-0 18.6c 111.6a 47-Ar 5. 3-3 249.0-D 46.3c * 17.4© 1 44-A r 7.0-B 78.3d j 307.3-0 s 7*2gp 251-Ar 4»3-3 « 4*8 bp - 3653.9-Kl 5 . Correction Tena » (Td)2 » (. 3653*9.)^ » 1 9 6 6 2 .7 T 673 Total Sum Square (1 1 4 .8 )2 -h I292.102; ....- 4 - (7 8 .3 r 2 2 .5 44.3 17.4 s 21830.0 - 19662.7 s 2167.3 Sum Square Method « (8 9 1 .2 ) -I- .... 1300.8 s 2 0 6 2 2 .0 - 19662,7 = 959.3 206 209 Sum Square for Rental j'^rea 303.3 ..... 307.3 s 20373*7 - 1 9 6 6 2 ,7 s 711.0 74 44 Table of Variance : % Corrected t Mean s Sum of Squares 5 Square s: F 3 d. f • 1 Source 678 2167.3 Total 479*6 ?fo.l** 2 Between method means 959.3 142.2 22.6** 7 1 1 .0 Between group means 5 10 497.0 7.9 49.7 M x H 4181 , 5 * 661 6.3 Error APPENDIX A gbcplanation of 3yrr.boIa a - Number o f respondents b - Average number of blocks traveled by those respondents c - Expected number 74 x 206 679 d - Product of b x c Ar - Humber of respondents with rental under $30 Br - Average number of blocks traveled by respondents with given rental - St Dr - Sum of "d’s" for given rental Ap - Humber of respondents to the personal interview Bp - Average number of blocks traveled by respondents to personal interview Dp - Sum of "d's" for personal interview T - Total number of respondents Td - Total of all fld's" Correction Term 3 i m i 2 Error Term ~ T T *t" test for testing difference where F is significant - t = mi « mg "t" test for method - Ccaparison of Personal and Telephone » t s % - mg .25 .25 Sbwp type of *t® test carried on between other method averages and between rental areas. The derived, *t* values were; Personal and mail - 2.I* Telephone and mail - 10.0** Under $30 and $30 $30 - $39 and $40 40 -$49 and $30 $50 - $59 and $60 $60 - $69 and $70 Under $30 and $70 $40 - $50 and $70 $50 - $60 and $70 - $39 - 1*4 - $49 - 7.7** $59 - 3 *0 ** - $69 - 0.46 and over - 3 .2 ** and over - 0.42 and over - 6.6** and over - 4*3** 79 AEPHIDIX A The following indicates how the data for all the relationships tested by use of analysis of variance were handled. hnalys,is of. Variance, for Table 32 in texts Comparison of Methods of Obtaining Data and Type of Potato to Coat per Pound s Personal :Average j Telephone : Mail Tvoe Potato a Cents No* Cases :Cents Ho. Cases:Cents Ho,.oases « Michigan 4a 99 4.2 51 3.9 148 4.0 Maine 4.7 57 4.5 39 4.9 89 4.8 Idaho Average 5*5. -35. 191 5*1 4.7 Ji 54 JZL 5*8 131 kA 334 4.7 Table of Variance i ‘Corrected Totalslvlean j s d.f* jSum of Squares *Square i F Source Total Between method means Between group means M x G Error 655 284.5 2 2 4 647 0*2 274.4 10*9 2.7 2147-8 3.3 Result of «t« tests between types of potatoess Michigan and Maine - 2*4* * Michigan and Idaho - 9*4** Maine and Idaho • 4*0*® Chi-square s 9*32 Need 9*483 for significance at *05* .01 137.2 -- 41.6 80 A PP EN D IX A Analysis of Variance for Table 4 in Text* Gcu^-arison of Methods of Obtaining Bata and Income Level Area to Number of Blocks to Store Personal $ Telephone i flail t ______ Avg, No, ; Avg. No# sAvg. No* sAvg. No. Blocks Cases : Blocks Cases . ’Blocks Oases sBlociag Cases Income Low 5*8 20 11*1 12 io*5 10 Medium 5*2 170 7*5 149 5*6 22? High Average blocks Total No. cases 4*0 JZ. 6*1 .22L k*l 90 7*3 4*9 Source Total Method Income Level M x I Error 237 5*7 234 Table of Variance s Corrected Total* ,t « d#f . * Sum of Squares * 1518*8 797 718.0 2 664*1 2 136.7 4 6181.5 789 *t* test value&s Low and High Income - 9.0** Low and Medium - 7*3** High and Medium - 3*8’* Personal and Mail - 3*1** Personal and Telephone - 9*6** Telephone and thil - 6*1** Chi-square value for table - 14*88 Chi-square needed for significance at .01 - 13*277 9*3 42 6#o 546 i,*o 210 5*9 798 327 Mean * Square * F 359*0 332*o 46.0 42.6 34.2 4*4 7*8 81 APP12JDIX A Illustrations of the two types of Bt® teats used to analyze differences in the study j a) H * teat used to analyze difference in percentages r^ £ k J L 2 k 4- A53Lg^M. r .037 240 378 t « l i L s 4.1** .037 bi) "tBl test used to analyze differences in numerical averages t * ai “ a2 Example - With or Without Telephone to Number of blocks to Store (See Text Table 20). With Telephone 5®^ 222 3^3 Average no. blocks Total no. of cases £x C t r Without Telephone 5*8 11 ,64 12.17 3 an r A 2.173 - Illfil)2 A/ s 222 614 V 30.r * 3 .' v a2 * 1 614 - = ^ 24.2 - 4*9 10 ^ ai " * = ^ = ia7 * ^*8 ", 8.0. r ,2 - 2 / U.17; t (1*48) i«9 ^ a2 = -yn = x^ 8 AEPENDIX B deludes all tables referred to in the text where analysis of variance was used but no significance was found* is not shown* The actual analysis 82 APPENDIX B Appendix liable 1, Method of Obtaining Data to Whether last Potato Purchase was of the Kind Usually Bought by Respond­ ents of Different lac cane Levels. Inc erne Groan Low Total Medium $Whether Kind s Personal s Telephone : Usually sPer* sNo.of sPer- No.of g Bought g i i t t • i 8 s • 8 g g Yes No g 88 : s JL2 * i 100 8 190 «9 8 I t i Yes No S i : % * i 8 Jd k Total respondents Ncn-resp ondent s * P ercen t o f respondents. 92 8 100 s 18 79 X _21 x 100 s 79 -21 * 100 192 » s 86 $ • s 100, i S 8 S 1 t 1 s % Total Yes Ho i i s 78 * .22 8 100 1 High t i 8 77 8 * 23 i $ 100 * 22 % Total t Mail Per­ sNo.of j *cent*«Casea scent* Cases cent* gCasesx Average 52 g 12 g g g « * * 9 e -M l 82 J £ 260 g 100 8 g ¥ e s 86 £ « JU L i loo 100 g • 90 81 s 0 83 s 100 * 111 i 1 « i 264 293 383 13 15 97 87 j a loo 33 APPENDIX B Appendix Table 2. Gaapariaons of Methods of Cfotaining Da i;\ on Kind of Potatoes and Kind of Container t Personal Kind of Potatoes Kind of Container sPer- jNo „of icent*iCases Telephone ..Mali. Per­ No. of Per- sNo.ofi cent* Cases cent*;Cases» Average ; Michigan Total ; * Balk i Packaged s 41 « -51 * Bulk * Packaged 30 70 100 ; 123 42 47 -52 100 -58 184 114 100 43 -52 loo Maine 100 Total * i * 76 16 84 26 -2k 100 84 100 5 23 • 2ZZ S 108 100 Idaho Bulk Packaged Total 43 * 38 t * 62 :100 s s 55 Total respondents Non-resp cndents * Percent of respondents* -2k 100 * 49 * i 30 100 138 * 254 288 429 23 20 51 47 _51 100 84 APPSMDIX B Appendix Table 3* Caspar is ons of Methods of Obtaining Data for Kind of Potatoes used by Consumers at Different lacone Levels* Income tKind of ^Potatoes aUsed t Telephone i I Low & Michigan Maine Idaho Other Total Medium » 62 i 24 * 14 :flOO • « a Michigan Maine < t Idaho Other * 44 s 31 « 22 High i ft ; • 21 £ Total e t ; * t 46 : 31 s s 23 • ; 17 100 1 i S I 41 2 28 « 30 1 slOO 189 * 55 s 22 s 21 '-.-iP s. :100s 53 « * To to.1 res;;-ondents Non-rasp ancients * Percent of reap ondents. 26 3 14 ft * « * 33 1 30 : 36 8 1 .100 ♦ 9 13 ; 100 29 18 - z s ; 5 • : 2 : 43 J * 53 l> 24 j : 5 s * 8 : s £ ft 9 « j Michigan Maine Idaho Other « • I « J ; 47 : 35 » 18 I ilOO 8. ;100 Total Mail Per- iNo*of iPer- :lIo.of: sPer- sHo.of cent*tCases 31 2 I89 100 282 £ 40 8 26 i ; 29 j • 112 ! 81 e .100 • * • ___._| r< __ _ S 28? 387 21 93 1 ■ i . ■ 43 27 28 __2 100 41 26 30 0 100 85 APEEMD1X B Appendix Table 4* Comparison of Method of Obtaining Data and Income Level to Kind of Container. Income sKind of e Law t . * Bull: s Packaged Total Medium i Bulk i Packaged Total High Total s Bulk : Packaged ; s Personal sEer- sWo.of Telephone s Mail Per- illo.of sPer- ilo.of Cases 3 * i i } « J I 9 e a : * 36 s 1 39 £ » 33 3 61 • » 84 5 *100 * 22 *100 • 18 floo £ 12 * «* * * 1 » £ « 0 s 3 41 * 38 * 5 37 * j 62 3 *-52 9 i_ M * J 262 % slOO slOO 136 *100 * 191 • 0 0 « • ? * 40 9 £ 0 * 9 s 40 s 46 : S 48 £ 60 * '-ill 1 . « 52 9 » slOO : 56 $100 83 slOO 3 113 * i * £ • * 6 • t f s • : ; Total respondents Non»resp ondent s *. Percent of respondents 270 297 387 7 11 93 Average : 36 ; 100 * £ 39 3 61 100 6 : kk i : * e «* 100 QG APPENDIX B Appendix Table JJ« Couparison of Method of Obtaining Data and Income Level to Number of People in family. Income Level Low * 9 1 0 Medium • Personal (Average) 3.1 Telephone (Average) 4.4 3*7 3*5 1*1 3.5 3*1 3*5 Mail Average (Averages 4*0 * 3* 8 » 3.6 * 3.6 * High Average Total respondents Non-respondents i % % I 34 376 1 1 * ; s 269 258 395 8 50 85 14 3.8 8? APPENDIX B Appendix Table 6. Gcaaparisons of Method of Obtaining Data by Rental Areas as to Whether Grocers Have Kind Usually De­ sired. Rental Areas Under $30 sDo Grocers ;Have Kind { Usually 0 Desired? * * e Yes s No Total $30 - $39 Total Yes s No « « $40 - $49 Total $50 - $59 $ Yes No i Yes No Yes No Total • e « 5 S « % » t a • Total $6o - $65 s s Personal Telephone t Mail sPer- sITo*of sPer- sNo»ofsPer«* tHo.ofs scent* sCases 5cent*i Cases scent*:Casea Averse e s « « i • ♦ f t s s i « S 87 91 91 94 i t j a I t s-a • • t 6 i _a 100 s 100 i 31 • 100 • 22 % 100 i 30 s s f t s # • • i 8 88 86 s 92 i f t 86 I $ I i 12 i i ft J J l • i I 8 1 J k 9 100 t 100 I 48 ft 100 ft 44 ft 100 i 58 % i i s * S 90 i ft 92 ft % £ 85 i 94 10 ft 8 s 100 108 100 100 s 82 100 89 « $ s ft 0 * e 0 85 l ft f t 88 5 87 • 91 « f t f t 15 f t 12 . « • --9 JL 2 100 t 55 i 100 s 32 • 100 s 40 t 100 f * f t f t f t . f t % s I 81 g 8 100 I 90 1 100 s J f t 10 0 £ * i s i i J& B J ; 26 100 100 f 9 18 f t 100 • 100 ft 8 $ f t J S ft « f c I s S 91 s 92 ft 100 93 • s -2 m i 8 I s i 100 6 s 100 ft 11 ft 100 ft 13 0 * loo s p f t ft ft • 1 « ; 100 : 0 100 s 4 50 94 f t 6 f t » » « 50 s 100 10 100 2 S Too” » 100 i 4 6 . » A « A 1 1 b ft I ft ft $70 - $79 Yes No Total $80 & over Yes No Total ft S « 3 s Total respondents Hon-resp oaden ts, * Percent of respondents. i 216 221 61 87 t 0 ft 300 180 -38 APPENDIX B Appendix Table 7* Comparisons of Methods of Obtaining Lata by R e n ta l Areas to Whether Kind Usually Bought. Rental Areas * Whether i Kind s Usually 5 Under $30 $30 - $39 s Yes * Wo * Total 1 t Yes s Wo t Total $40 - $49 $50 - $59 $60 - $69 $70 - $73 t 1 Yes i Wo t Total t. t Yes s Wo s Total * s Yes ; No 1 Total i s Yea s Wo i Total • $80 & over : Yes s Wo 8 Total e S Personal i Telephone s Mail Per- sNo.of ?Per- sNo.of iPer- aNo. of « ^ 77 s -21 i 100 8 s 86 * _JA s 100 i 1 87 9 • 8 8 8 35 • 9 I i • 52 & 8 i 8 100 i 90 s 8 * 89 8 •JA 8 * 8 100, • 37 • » 8 100 8 8 - 8 8 100 8 3 8 9 a * 75 ft « 8 100 8 5 8 * 5 100 • 8 - S 8 100 « 3 « Total respondents N on-resp ondenis * Percent of respondents. 8 • { 83 : JZ 8 100 8 8 8 86 i 8 JUt i 8 100 s » 9 8 s 80 8 8 20 8 8 100 8 9 8 • • 8 95 8 8 8 100 8 • « I 74 8 $ 26 8 ft * 100 9 6 i 9 8 S3 8 r t tJ Z « 8 100 9 8 £ 8 86 § • jyi • 100 • 4 V i i s 8 0 • 8 78 8 24 52 94 42 19 12 «■» £ 22 1 100 8 8 76 8Jk 8 100 8 8 86 8J A * 100 8 ; 78 8 22 8 100 8 8 86 8- M « 100 ft 5 79 8 21 ft100 8 8 90 8 10 9 100 i' a w d 8 8 32 ft • 8 8 8 63 8 8 8 8 113 8 8 8 • 59 8 8 8 8 28 £ 8 * 8 14 £ 8 O £ 10 S e A « 230 250 315 47 5S 161 79 21 100 82 18 100 84 16 100 86 Jk 100 84 16 100 80 20 100 30 10 100 89 APPENDIX B Appendix Table 8* Comparisons of Method of Obtaining Data by Rental Areas to Average Number of People in Family Personal sAver- sNo.of sage (Cases : : * 3.3 * 35 • Rental Under $30 ..... $40 - $49 • • : 92 * 3.1 1 37 i s # I O i ; 3*8 I I : 3.S * 9 $70 - $79 * 3*2 ( • t 1*0 * 5 3.5 i 6 • • Total respondents llon-resp ond eats ( * 3*6 * 52 i $6o - $69 $80 & over Average • 0 * 3.7 * 26 * 3*9 s t i i 3.6 1 5 2 i 3*4 s $30 - $39 Ifeil ; s Telephone J sAver- (No* of sAver- No.of: Cases: Average .«ase. (Cases sage $ • * 3*4 * 94 * 3*7 * i 3*6 * 45 * 3.5 ( ( * 3.4 : 20 j 4*o I $ * 3*4 s 13 • 3*5 33 1 s 66 s s 113 3*7 3*3 S 3*6 61 g 3.4 s 28 s s 14 ( 3*7 3*4 • 3 i * 3*0 • * 3.5 s i ( : • • 7 *34 ( Its 10 : ( ( • 0 234 237 327 43 31 153 34 3^6 APP3JDIX B Appendix Tab le 9. Comparisons o f Method o f O btaining Data by R en tal Areas to Kind of C o n ta in e r* 1 Rental Areas Kind .of (Container s Personal i Telephone Ifeil (Per- (No*of (Per- iNo*of Per- Momof ;cent*tCases (sent*(Cases cent*(Cases Average ( Under $30 : Bulk ( Packaged s Total • • $30 - $39 ( Bulk (.Packaged Total s 34 «J & * 100 ( 3 46 « J5k * 100 35 52 » $40 - $49 $50 - $59 Bulk Packaged 3 Total ( t Bulk 3 Packaged ( Total * 32 : 68 : 100 92 c 5 49 « -51 ; 100 39 28 * 35 » i 65 ( JZ2 i loo ( 26 ( 100 I 32 ( 8 ( 4^ 8 33 (4k * JZ i 100 ( 52 100 64 0 « ( 36 ( 50 .50 ( * _ik * 8 100 8 94 100 3 113 ( 8 8 54 8 8 46 ( ( 100 ( 43 31 Jl 100 s 45 * ; 55 i ; 100 ; 20 38 62 32 68 100 41 252 100 40 60 100 43 -51 61 100 e * $60 - $69 : Bulk ( Packaged 5 Total ’ * 56 s 44 ( 100 s $70 ~ $79 * Bulk Packaged Total * -M * s 100 3 12 a $80 & over Bulk Packaged Total ' JZL * 5 100 Total respondents Non-respondents * P erc en t o f respondents* 26 100 43 -52 100 c.p, Too 14 •> * 30 ; ? 43 : * 33 100 « 1 54 ; * 17 . 1 » 100 ( 43 -52 J 5 I 1-22 J * 8 100 8 i 7 3 100 : 10 s 236 254 320 41 54 160 100 91 APPENDIX B Appendix Table 10* Comparisons of Method of Obtaining Data and Kind of Potatoes to Number of People in Family. Kind of Potatoes Michigan Personal i p ♦ 3 *1 Telephone i : « * • 3.6 : 3 *5 s 3*6 ! i H * ft • Total reaiDondents Mcn-respondenta * 3.3 t 3.6 ! 3*6 34 3*6 * $ • ft • 3*1 • • Idaho Average 3.9 I • Maine Average Mail t (Average) 5 3*1 i 3*6 I * • • « : * 25k 291 436 23 17 44 3.6 92 APPENDIX B Appendix Table 11« Comparisons of Method of Obtaining Data and Location of Store to Kind of Container Inside and Kind Personal Outs id© Shop­ i of or :Per- iNo.of ping Area Containerscent*;Cases s s sBulk i Major iPackaged % J¥k sTotal : 100 25 s s s i i Minor sBulk i 36 s sPackaged s * sTotal : 100 s 91 i Telephones Ihil t sPer- sNo.ofsPer- sNo*ofs scent*jCasess cent *sCases s Average s i 29 5 32 2 4 i 68 t 100 22 * • *« 39 t 61 i 100 83 t i Neighborhood i sBulk ; 33 i Store {Packaged s M ■ sTotal i loo i 115 i i Noa»respondents * P ercent o f respondents S I s 45 s s i % i i 100 * 130 231 235 46 73 JZi 100 i * i 62 t 1 ; i 36 34 * 66 ; : J4 100 s212 1 100 i 1 I s s « 36 s i i Total respondents 76 100 43 J lL 100 5 i i * * ;121 • S I I i 373 38 62 100 93 APPEHDIX 3 Appendix Tab le 12* Comparisons o f Method o f O btaining Data and Location o f S to re to Kind o f Potatoes inside and s Kind Outside Shop-* of t Major Area Minor Area {Michigan {Maine iIdaho sTotal * Personal ; Telei ilone i Mail {Per- iNo.of :Per­ Ho. ofjPer- No.of: 8 t s { • c • • 20 { 100 { 25 % {Michigan % 44 { & 37 {Maine {Idaho Jd % 100 s 109 sTotal » i • 9 ft 0 • { Total respondents Non-responden ts * Percent of respondents. 36 0 « { : e 0 30 • -21 JO. « 100 { 22 • 100 0 i • : { # s 30 9 39 V 30 0 35 • « -21 1 -21 : 100 * 75 • loo • • : » 1 I 37 36 20 28 • { -3i J£ 100 s 124. 100 * 9 { 1 { » • S { s 37 { : 23 8 t * {Michigan S 52 0 {Maine * 26 9 % 22 i {Idaho {Total 100 % 85 ¥y 9 t Outside of Area s 44 i 36 9 -21 43 : 100 » . 0 i t 1 39 i 28 i _11 » 0 9 ft 40 30 -11 100 202 a 9 36 33 116 • • e » 9 0 219 221 361 58 87 119 100 94 APPENDIX B Appendix Table 13« Relation of Method of Obtaining Bata and Inc due Level to Outside Shopping Area Income Level Low Personal s Telephone Mail sPer» No. of sPer- sNo.ofsPer- sNo.ofj Area seent* Cases scent* sCases scent♦sCases: Average s s 1 •» sMajor 10 1 - s i 27 2 12 ;Minor t 67 1 « 27 • 55 51 ineighborhoods 35 s 46 s -21 * - 22. sTotal s 100 20 100 5 12 5 100- 2 11 loo S « Medium High sMajor i 13 3Minor j 34 ^Neighborhood 1 53 2Total ; 100 sMajor 2 2 sMincr s 50 sNeighborhood: J & sTotal : 100 2 s » 167 s * * 8 46 2 i % i * % $ 12 * s 14 s 27 s s 49 s 61 2 I 37 1002 150 s 100 229 s 2 2 6 S 3 2 76 2 * 45 s : * 2 100 2 71 21 100 J& 100 £ 2 » 2 13 58 « 96 2 w » • 2 * 4» S « « 0 2 V % 2 t * Total respondents Non-resp ondents * Percent of respondents. 235 233 336 44 75 144 4 61 -21 100 95 APPENDIX a Appendix Table 14* Comparisons of Methods of Obtaining Data by Rental Areas to Kind of Store 1. Rental Areas Under $30 $30 - $39 Kind of Stare s Personal i Telephone Mail : :Per- sNo.of Per- silo.of Per- sNo.ofs i cent* %Cases cent*:Cases cent*:Cases: 1 i •» : : £ £ : s : Chain : 70 : 5 85 : 50 : : Independent: j o : JO s * _ il * : Total : 100 • 33 100 •• 20 : 100 : 26 : • £ • « * • s : 79 : I : Chain s 70 s 62 •• : : : _21 : ; Independents j o s J8 : Total : 100 : 45 : 100 ; 58 • 100 • 49 0 s • $40 - $49 : : : £ Chain : 67 s Independent: . 3 1 0ft Total : loo * 82 £ $50 - $53 » 0 £ ; : Chain : 100 s s Independent: - I : Total : 100 : : ; Chain s 60 s Independent: JO : t Total s 100 7 loo • • % * .1 1 * 78 0 5 22 100 t t * * s 0 « 5 33 S JL s $80 & over 5 Chain : 67 I s Independent; j a s s Total s 100 5 Total respondents Non-respondents * Percent of respondents. 3 100 • • 100 * « 0 0 • * 9 : 100 s 12 £ £ ; 100 s • 0 37 17 * 100 8 2 9 : •* 100 s s : : s 0 j : 221 65 87 • 3% J& 100 « « 5 s 212 70 : J£ s 100 • 100 • ; 72 : S 28 * ; 53 * 100 i « : : 85 0 J1 • 100 : 23 % £ 0 1 £ 9 76 * 24 s : ; 5 85 64 J l *J i 100 : 36 : 100 • 0 82 18 : 85 : 100 71 J2 100 0 8 0 58 5 Chain 6* 4 : Independent: s Total : 100 : 33 : • $70 - $79 s « s $60 - $63 60 JO 100 70 JO 100 £ 8 £ • 280 65 J 1 100 ?6 APPITTDIX B Appendix Tattle 15* Comparisons of Methods of Obtaining Bata and Mind of Store to iiand of Container Kind °T Store Chain Kind of i * t Personal t Telephone : Mail i sPer- :No.ofsPer- gNo.ofgPer- sNo.of: s : Bulk Packaged i Total i i • 9 • t i 'Bulk Packaged Total : t: 31 I 1M. i : t : 100 ♦ » • % Independent t • • 167 • * £ : : : - 5 k « * 100 • • • 46 : t 9 0 74 « 9 0 38 62 100 40 60 100 g t 9 9 9 t 34 i : i : i » 159 a : 1 : ; • S • » 0 103 0 : 66 : 100 303 42 a t : • 1 a £ 9 J 5 L 100 S : s s s 5 0 9 • 9 % * £ 0 9 £ g i : $ Total respondents Non-respondents 241 262 362 36 46 118 "* Percent of respondents. Appendix Table 16. '' 42 -Si 100 9 9 59 34 66 100 ~~ Comparison of Method of Obtaining Information and Kind of Store to Number in Family g Kind of Store t & Personal « t » Ifeil t Average 9 *> • « Chain Telephone : (Average) i 3.6 3*6 g 3*6 '• 3*6 1/ t Independent Average 5 ■s _2ii r-> r* 3o : Total respondents Non-respondents 5 3-5 * * 3*7 3^6 ; « t : 0 r; 240 2b2 362 37 46 118 97 APPENDIX B Appendix Table 17• Comparisons of Method of Obtaining Information by Income Level to Kind of Store Kind of Store i Personal Telephone s Mail Per- sNo.of sPer- sNo.of i scent*tCases cent*s Cases scent*sCases i s : * $ % Chain s 91 * * 66 s 69 * % Independent * -3k 1 -3k i * — 2. s i Total i 100 * 21 100 s 13 s 100 s 11 i s s s t : ; s i i i 60 s : Chain s 87 s $ 69 i s Independent * ja. * -M • *_ n « 100 5 168 s 100 : 230 Total s 100 $ 174 •• Lew - : s 1 il A•* .• Income Level Average 73 SL 100 • e Medium e <* High i s Chain t Independent i Total » » • 0 Ilon-resp ondents Percent o f respondents. f s • s * 79 5 38 s s 21 s 100 s 79 s 100 s 98 * l & » £ 26 100 9 i 62 » * 31 : £ 100 £ 48 6 5 O 5 • 9 i » • | % 1 69 s Total respondents * e 74 t s 243 260 339 34 48 141 71 J3 100 98 A P P E N D IX B Appendix Table IB, Comparisons of Method of Obtaining Data and Kind of Potatoes to Whether Kind Usually Bought. Kind of Potatoes Michigan Maine Idaho 2Whether Kind s Personal s Telephone i Ifeil Usually 2Per- sNo. of jper- sNo.of :Per- sNo.of 2 % s *» • « • • * * * s * I 5 s £ 9 f t p Yes No Total Yes No Total » 0 i 0 81 • i 79 : • 71 4 • 2 2 21 2 2 -22 • • 100 2 122 1 100 2 96 2 100 « s 1 • 2 • • • i 2 • • i 89 2 93 84 I 16 4 • 11 J 2_2 2 100 i % £ 100 t 70 0 100 t 0 2 2 I £ 0 0 2 t f 91 * 98 # » • 81 • • • 9 • t __2 0 78 2 100 4 100 2 5k s 100 <* 2 9 a Yes No Total 4 * 0 V * 'Total respondents 2 2 2 2 186 0 0 2 2 » 104 « 2 2 w 138 . 1*. 252 428 6*/' U on-rear,ondeixts * Percent of respondents. 25 100 89 11 100 90 10 • • 76 -St 52 100 APPENDIA B Appendix Tab le 13» Comparisons of Methods of O btaining Data and L ocation o f S to re to Nunbcr in F a m ily . Inside and s. Personal ft ! Telephone Mail Outside Shop- 1 Aver-s Ho. of s Aver­ • No, ofs Aver-; No* of pine Area : age ; Cases • age » Cases ft age j Cases a • ! i 9 t V • O O ♦ Major Area % 23 s . i 22 a 3.3 * * 3.6 45 s i « i • S Minor Area : 3 .b t 1 89 •• 3 . 5 83 s 3»^ s 210 % i s t i % « Outside of ; t * - * • 0 Area s 3.5 > 116 I 3.6 t 130 s 3.6 • 123 8 ft JT6 * 3*6 s Average 5 3.5 ! 4 * « ft 0 0 * Total respondents ITon-resp ondents Appendix Table 20* 232 4 5 235 330 73 100 a • i {Average i i 3.4 *r • 3.3 « ft # e 3*6 i 3^ { Comparison of Method of Obtaining Data and Whether Inside or Outside Shopping Area to Kind of Store Shopping AreasKind of StoretPersonal t Telephone * Mail s (Percent of respondents) 1 s * 96 % • 8? jChain i Major 73 * t Independent 1 -it « * J& S L s 100 100 jTotal 1 100 i i % 5• 78 % 5 Chain j 94 79 ft Minor 22 21 • i 6 {Independent 0 j 100 100 loo {Total s $ a • {Chain {Independent .Total s T o ta l respondents Non-respondents « • -3* 100 5 58 t 69 I J£ * -3i I t i 1 « f t e 11 0 9 86 jyt 100 87 100 % • 61 t Average 100 ; 100 233 235 334 44 73 9b i # 4 63 57 100 Appendix l&fale 21® Comparisons of Methods of Obtaining Ihta by Hental Areas to liind of Pots.toes Used Hental Areas Under $30 - sKind of Personal s Telephone I&il sPer- sNo.of :Per- sNo.of Per- sllo.ofs i ; 9 0 0 9 3 46 0 • s I $ * $30 sMichigan 6 9 55 sMaine s 26 • 3 25 • 9 22 3 e J iIdaho s 18 • 0 25 0 0 22 0 iOther s 1 0 m 3 e 0 -Jl • * 32 I t Totals 100 « 34 s 100 • 24 loo 0 • 41 0 » $33 {Michigan s 55 I 0 £ 49 * • :Maine ; 21 t 3 38 3 $ 14 i & sIdaho s 20 J £ 19 t £ 33 I i * 5Other • 2 i • $ 4 9 • s Totals ,100 51 0 100 3 52 O 100 9 63 • $49 {Michigan t 38 I 41 » O 1 37 • {Maine s 37 S I 27 I $ 28 £ 3 {Idalio s 23 0 i 31 3 £ 33 * £ 0 sOther j 2 O 0 __- 0 I -1 s Total{ 100 { 92 3 100 $ 92 • 100 3 113 £ 0 $59 {Michigan 3 33 • t 50 { 3 46 s 3 sMaine ; 24 1 3 33 0 3 28 • £ {Idaho s 24 0» 0 23 0 31 £ » ; 0— a w tOther 2 0 • £ • 42 s Total $ 100 • 34 0 100 « 100 •« 60 C 0 0 20 S 3 43 £ • $69 {Michigan s 62 0 0 {Maine s 33 £ 9 • 40 14 £ « «» i • ildaho s 3 38 £ • 40 it 11 0 tf - 0* « £ 0 # sOther s —2 i Total s 100 s 8 in 100 3 20 3 100 « 28 • $79 {Michigan • it 43 3 t 50 s 50 * • 8 s 6 22 • 09 {Maine s 17 « 4 • 42 & 0 0 • 28 * sldaho : 33 • » __- t 3. _JL £ • {Other s •• { a 12 O 100 O 0 6 3 100 * % Totalj ioa » 14 0 « 42 % 0 30 • « * 34 « over {Michigan e 20 0 0 23 » 3 {Maine % 33 0 e 0 # 0 * s 40 iIdaho t 33 e 29 < 3 0 3 _____ 10 #0 - s sOther 0 * 3 100 t 100 10 j Total $ 100 7 • 3 9 « 0 0 « 5 . 1 % 56 0 t 1 $40 - $50 - 1 $6o - 9 $70 t - $30 & it 9 a T o ta l respondents Hon-resp ondents 228 245 320 45 mre *j l6o 53 24 21 100 48 24 2^ JL 100 39 30 29 2 100 2^2 29 26 0 100 38 27 32 — 3 100 47 16 34 —2 loo 32 26 — i .100 • 101 AT'PSIDliC 3 Appendix Table 22* Caparisons of Methods of Obtaining Data end Income Level to Cost Per Pound • Inccone Levels i Hon-resp ondents Appendix Table 23, i i • »• 4 .7 £ 210 0 0 00 Total respondents Nail cents sNumber 4 Average Cents 2 • 4.7 £ 11 s 5*1 CO * Telephone s cents sNumber s s. ; 5.8 * 4 « • 3 s 4*9 J 9 0 •s 3 4 . 8 2 38 * «• 4.9 * 3 t 0 s 3 w Personal cents sNumbers 1 s 3 Low s 5 ,0 t 20 5 • 2 Medium 2 i 4 . 4 * 138 0 s ; High $ 4*6 s 44 0 „ Averse 2 • 4.5 • 0 • 0 0 a e s 0 6 0 3 3 3 4.7 * 0 0 s 202 132 30S 75 176 172 4 .6 -4*7 4.7 Comparisons of Methods of Obtaining Data by Hental Areas to Cost per Pound s Personal s Telephone 3 l&il Rental Areas t cents sNumbers centss Number; cents {Numbers Average Cents » 0 0 3 3 « 0 « 3 ♦ 4*6 4.5 3 31 i 4.8 3 7 i 4.6 3 28 3 Under $30 I 3 S 3 3 0 2 4*6 3 $30 - $39 4.3 * 38 3 4.6 3 26 I 4*8 • 47 3 Ui VO 1 v_h 0 $40 - $49 » 0 s 3 0 0 3 4.5 • 61 0 3 * 0 4.8 £ 40 0 0 4.7 £ 29 3 4.3 £ 3 0 3 4.1 £ 6 3 0 $60 - $69 0 4»o 0 s 3 s 5*4 s Average • 4.5 « $70 & over 0 8 3 5 3 21 11 3 0 5*2 3 4*6 j 0 S 3 5 4.8 3 93 3 e 4.7 0 4*5 s 44 3* 4.5 3 3 3 S 0 0 4.5 £ 0 0 7 « • 3 0 0 0 23 3 4.3 0 £ 17 0 4*7 » » 0 0 0 0 i 4 * 1 • Total respondents 172 112 252 Non-respondents 105 196 22S 3 3 3 4.8 4*6 1 r-- 102 Al'PilTDiX B Appendix *Bable 24» Relation of Method of Obtaining Data and Rental Area to Number of Pounds Per Person Per Day s Personal tail Telephone s Rental Areas spounds*numbeis oound3 :number spounds 1numbers Average » 50 * $59 ( ft • A A 1 % J ft *33 • 8 2 .29 £ 19 * •43 £ 26 £ $60 - $69 *39 A i A £ 4t A ft £ *54, * 21 i *31 * 22 © 4 $70 & over •34 9 •as, ^ or „ 0 *39 * 0 • »i.i,o & A A •38 Average ft « » • a «> % $ s 9 9 $ S Total respondents Non-respon&eats 0 0 A * 1 # 0 231 247 309 4.6 61 171 Appendix Table 25* Ccamperison of Method of Obtaining Data and Income level to Ilumber of Pounds Per Person Per Iky, « Income Level Personal * Low Medium High Average i t. • * * • i i Total respondents Ilon-respondents Appendix 'Table 26* •40 *36 s33l .37 2 6 8 9 « ; Telephone • (Ppunds) e V s : s £ : t i 1 i' •40 Mail s Average .3 3 1 •31 *33 * •35 •35 • s • • £ i *36 s •35 a2L .34 t i 291 313 1 7 107 Comparisons of Methods of Obtaining Data and Kind of Potatoes to Pounds Per Person Per Day Mail Telephone Personal pounds i numberi pounds s numbert pounds;numbers Average s i % % d i * •30 I Michigan •37 * 125 i .36 * 115 • .36 s I89 1 0 t • s i $ t £ i .3 6 5 76 t *34 * 84 £ •35 1 109 * Maine •35 « i i » s 9 t i t •31 * 56 • *30 * 92 * s 3 i s 142 S Idaho 1 i i •36 Average t •35 •34 « *37 * • t 1 % s % % S t t % i % Kind of T o t a l respondents Non-respondents 257 291 440 20 1? 40 104 APPENDIX B Appendix Table 27* Comparisons of Methods of Obtaining Data and Type of Potatoes to Kind of Store. • ! i £ Type of PotatoessType of StorejPersonalsTelephone* Mail * (Percent of respondents) • * • Michigan : s Chain 70 5k i 77 « JL i Independent • JSL * _22. 6 100 i Total * • 100 * 100 % 1 s Maine Idaho 1 % Chain 1 Independent % Total t 1 74 % - 2a % 100 • *• s Chain s Independent *-Total s 1 £ : s Total respondents l'lon-respondents i t % 1 1 64 _JL 100 1 92 * 8 * 100 s 62 i 9 0 — SL­ 9 9 £ IGO • i 9 tt I 68 -22 100 I <36 «J L 1 100 * 229 240 366 48 68 114 t Average • « * * 1 1 I £ £ i £ £ : £ £ £ 6? JSL 100 76 _24 100 72 _28 100 10 aiphjdix b Appendix T ab le 28* Comparisons of Methods o f O btaining Data ond TVce o f S to re to Reported Cost Per Pound o f Potatoes Type of Store Personal t Telephone Mail (Cost per pound in cents) Chain 4*5 Independent Average ka.6 IjTT Total respondents Nan-respondents Appendix Table 29* * i « £ 4*8 * s 1 * S £ 2fc*§ 4*8 4*8 Average s 4.7 s juti 4.9 1 i £ 4*9 4*7 i 183 122 285 94 186 195 Comparisons of Methods of Obtaining Bata and Number in Family to Reported Cost per Pound of Potatoes Number in Familyt Personals Telephone s Mail t (Cost per pound in cents) { i i i 4*9 i 4*7 s 3 *o Under 2 3 -5 l 5 4*4 * 6 and over Average « t Non-respondents 4 *6 a 4 *6 4*5 l T o ta l respondents % ! » • i * Average 4*9 4*7 4*6 4*3 4*7 4.2 O « 4*Q 4*7 l • » i 202 138 349 75 170 131 APPENDIX. B Appendix Table 30» Comparison o f Methods o f O btaining Data and Zand o f P otatoes to C olor A f t e r Cooking. Kind of Potatoes Personal Telephone Mail tPercent tNumb er iP ercent sNumber tPercent;Nmaber s Average 9 • • a £ » a • 1 • 0 0 £ Michigan 6 Poor 13 13 J 20 • 29 9 « 11 9 11 * a 23 Average • 26 • 33 • 14 1 16 3 32 • 45 • c;q Good 69 J>S £ 81 s J l • 86 £ J £ 100 : 100 s 125 £ 100 100 « 143 Total • • 5 £ £ 1■ • 2 * • •9 (61 • « S i 5 NumbersPercent tNumiber tPercent tlluaber: Average a » i ft » ft « « » « i S 2 £ 3 t 6 « jr 3 0 • 64 6 © 100 * i © 8 £ 37 • 80 £ 123 * « • 1 » » 8 9 a ft 100 2 2 t 6 » 68 * 76 • « 9 0 2 3 8 i ft ft ft Idaho Poor 4 Average 2 7 Good 5 J32 Total « 100 • A • Other i Poor 0 Average • • Good c Total e * • » 4 JR • 3 5 i f t 2 8 72 1 0 0 £ Total respondents Non-reapondents ft e • 0 ft 2 »_ 9 • 7 s * » 2 « » 17 b 81 « * • 100 1* » « # 4 e 15 • s 81 I 100 f t f t ft ft ft t ft £ 9 4 ft 0 5 1 3 S 68 s 84 ft • f t f t 1 1 9 8 1 0 0 1 S «> 1 8_§2 S 9 1 ft a» « ft 25 1 ft —or Lj, X X 2 ft 20 £ _22 S 115 S • • • « £ 3 1 0 0 $ ft C «> ft 1 f t— 2 ft 9 ; 3 4 263 294 14 14 « • £ £ » £ • • v * ft e © • * • * s : £ t < s « 1 1 52 JR 100 ft 4 46 -ifl 100 «•> 22 73 i l—n 1 0 0 £ ft ft ■9 f t f t 3 f t f t • f t f t f t s ft • 16 • • 71 ; _5 i • 138 « w 3 ft ft ft 3 7 * 5 Jfcfi » ft f t f t ft 1 « ft 22 ft J R f ft t 39 ft f t « ft e e 80 8 0 1 0 0 ft ft « A ¥ t ft 1 1 8 1 0 ft * ft 9 f t ft £ i » * ft 3 2 7 153 7 3 4 . 5 2 100 3 OO 0 6 f t ft ft f t 9 1 0 J « f t J J ; ft f t f t 10 « A f t 100 3 11 8 8 1 0 0 5 1? 76 100 108 APPENDIX B Appendix T a b le 32® Gca^arisons o f Methods o f O btaining Data and land of Potatoes to B a tin g fo r S iz e . Kind of Potatoes Personal Telephone Hail ^Percent tNinabertPercent tNumbersPercent tNumber t Average ; ¥ s • • i 2 ft ft f t 20 t % 2 • a t 6 3 15 f t £ 1 t 30 43 34 40 24 27 67 t 68 76 { Jti -Si .31 -St -5k100 % 123 100 : 114 j 100 133 $ 100 I ; I 1 Michigan Poor Average Good Total Idaho Poor Average Good Total • 9 9 •1 9 - 9 9 * 9 9 9 • i 1 * 9 l a 1 • • 29 70 • loo ' • • < 1 a • $ t < 1 22 - i _51 t 76 i • 9 a 9 9 £ 9 # £ 9 5 % » » 18 82 100 10 9 -1 1 55 t « 9 « & i 31 M 100 2 14 _m 100 ; • • # 9 S 1 » «• 9 • i « • 1 t 26 1 87 83 $ : 2 18 s J .6 91 « « 9 • 9 4 44 -5S 100 3 34 .10 s 77 % » » » 2 f t 9 i s 4 9 27 63 9 94 4 29 £ Jz £ loo i t • • % s 9 £ a » 9 9 f t 9 9 t 9 9 T o ta l respondents 9 & * £ l S 20 2 £ 20 1. » 60 7 » 100 9 9 1 1 5 2 • 4 s 8 £ 9 $ 9 18 s JX £ 100 2 21 JZZ 100 9 9 i ft 5 9 9 -i3 ft 100 9 9 9 2 35 1 £ i 9 9 » Other Potatoes Poor S 14 f t 28 Average Good. 58 Total 100 9 s » £ Non-respondents 9 9 9 * 9 9 • Maine Poor Average Good Total • • 1 2 3 11 2 63 292 315 14 16 165 * $ * .34 9 23 9 JL 9 100 209 APPENDIX B Appendix Table 33. Comparisons of Methods of Obtaining Bata and Kind of Potatoes to Bating for Blemishes and Defects. Kind of Potatoes Personal Telephone Mail ^Percent tNumber Percent >Number ^Percent tUumber* Average i. Michigan Poor Average Good Total Maine Poor Average Good Total Idaho Poor Average Good Total 11 40 49 100 s 13 t 50 . Jl % 124 1 1 20 . Jl 100 2 20 78 100 6 24 I M70 MW 100 3 1 15 20 ; is 100 1 Jl t i i t 1 * 11 « _M s 55. J l 2 16 82 100 24 7 28 80 U5 2 17 65. 84 53 loo 2 s 2 44 i Jk £ 100 1 2 48 £ -52. 100 30 , 14 39 Jl loo Jl 2 s 14 28 61 s 1 £ 23 * -28 £ 52 S : £ I 1 J 30 s , 62 — 2 26 J£ 100 2 27 _2i 100 i s Other Potatoes Poor . 14 Average * 14 Good • ■JS Total . 100 Total respondents Non-respondents t • s 25 Jl 100 1 £ 14 86 _J. . 4 ; 100 I .— L. 1 1 7 262 293 237 15 15 243 17 Jl 100 110 Al'PMBIX, B Appendix Table 34* Comparisons of Methods of Obtaining Data and hind of Potatoes to Hating for Taste Kind of Potatoes Michigan Poor Average Good Total X&ine Poor Average Good Total Idaho Poor Average Good Total Personal Telephone Mail s-Percent tNumberiPercent ;Number 2Percent ^Number iAverage 0 2 • i 2 2 8 i 2 • i 2 2 8 • 2 % 0 6 2 8 a 6 *0 3 t 4 5 • * 17 0 11 2 12 • 46 62 t 24 14 1 -§k j 103 2 J83L 66 t 2Z1 * -21 8 J O s 100 s 123 $ loo 2 115 2 100 134 •0 100 2 i « s 2 • 1 % i 1 a s 4 Jl 2 100 » t 6 0 1 2 2 2 i 1 2 a 3 2 12 2 -22 2 86 i 76 V 100 « 1* 0 £ 1 1 «. s - « 2 0 e 1 % J8 0 100 V 00 e> « s 2 Oihe Potatoes 5 Poor 2 14 Average « 66 • Good 100 • Total s 9 t s 8 2 2 2 » J k 55 S £ £ - 6 7 2 2 4 2 10 30 * -21 2 66 2 83 2 loo 2 2 • a -21 100 2 2 2 4 2 _81 2 83 2 2 2 2 3 21 JiZ 71 9 0 2 c • . 16 J k 100 1 8 2 Ji • 100 » • 2 » * 2 • 2 2 « 2 18 82 - 14 J& S3 ioo 2 100 * _Jl 2 4 2 2 100 - 2 — 2 11 • 0 • 0 « S £ 2 2 i 8 2 0 • 2 Non-respondents 100 0 6 0 Total respondents 3 15 82 2 - « • » • 9 8 2 • a 3 4 - 2 2 2 a 0 1 2 261 231 16 1? 304 l?6 5 3 86 100 111 APPENDIX B Appendix Table 35* Caparisons of Methods of Obtaining Data and Kind of Potatoes to Cooking Quality* $ Kind of Potatoes i Cooking i Quality t Michigan • Poor 4 f t Average ft Good • Total w O * Maine I Poor t S Average 9 Gocd tf Total Idaho Poor Average Good Total • Personal s Telephone * • • m n iPer- ilIo„of :Per- :No*of ;Per- :ITo*of e tcent *tCases scent* sCases :cent*:Cases : Average 9 0 • I s « 2 • 6 • 1 10 : t 1 6 * A 0 S 24 i 0 39 * » 28 • J 17 * * I -51 * » « 74 « 1 66 V 100 s 1 2i| *100 9 115 • 100 i 158 s 100 » * ♦ • 1 0 i i e • A 0 « • 9 i V ft J - 9 2 • 3 • i 1 » A 26 i i 9 £ « : s 6 i 14 « i ft J l * % -22. e i Jk 100 100 i 76 slOO i 83 * 100 : 95 £ i ft • i 0 % 5 « > 1 } 9 :■ « i » I « 2 & e 2 5 s 2 • o - 0 i 10 11 1 • 0 16 5 5 A S 83 it 87 * t 9 84 s_21 f 100 100 : 109 * 91 & S 100 * 55 aoo * £ a ft 5 * * e » 0 « i : 4 it Total respondents iJoa-resp ondents * Percent of respondents. 255 239 3&2 22 19 US 112 APP3NDIX B Appendix Table 36, Comparisons of Methods of Obtaining Data and Kind of Potatoes to Bating for General Desirability. Kind of Potatoes Personal Hail Telephone tPercent Humber:Percent sUumber ^Percent Nunber Average i t 2 8 Michigan « s 1 ft Poor : £ i 8 16 6 3 3 7 17 s 9 f t Average : £ 22 8 i 38 36 37 25 ft 49 51 e a 0 Good 72 • i 0 83 • S L -52 _3Z ft .51 -32 Total • 1 loo 115 0 100 ft 100 100 105 97 5 « s s o f t Maine 1* : f t ft 9 ft 2 s 2 8 2 2 2 Poor l » 9 3 • 13 • 10 ft 17 Average % 2 £ 16 4 29 0 ft t 9 f 82 68 81 Good • 68 • 9.4 -51 £ 62 ft 100 » 83 j 100 Total • loo 54 * 100 f t 8 % f t ft 1 a S Idaho ¥ ft i i « . ~ - $ 1 ft 1 Poor f t 1 i ft f t a 12 £ 11 ft 10 • Average a 17 13 4 e 9 • f t f t s 88 Good a 99 -5Z .22 -82 -32 js * f t 90 : 100 100 71 % 100 100 f t Total i 43 s f t a 0 f t ft 3 • 0 0 i Oilier Potatoes ft v # ~ X f t w # Poor 2 29 22 0 £ 8 £ 2 20 Average I ____£ 0 y g 100 £ 78 It Good 1 ____ i — 4 ft J & •I 100 a 100" 7 a 100 Total » 4 7 S 100 i £ a * a 0 < 8 8 « a Total respondents 1Ton-responden fcs 201 292 245 76 16 235 113 APPENDIX B Appendix Table 37* Caparisons of Methods of Obtaining Bata and Kind of Potatoes to Rating on Quality - Mashed* Kind of Potatoes Personal Telephone Mail {Percent {Number {Percent sNumb er iPercent {Number i Michigan Poor Average Good Total Maine Poor Average Good Total Idaho Poor Average Good Total s £ ft 2 27 ft j i ft 100 t t • • 1 i ft 6 ft J l i 100 e 9 9 % » • 4 3 ft ; J£ £ 100 8. $ 2 2 £ 34 i 88 s 124 » ft j 2 1 ft 4 0 Ja. 9 74 0 e « « i i 6 2 7 2 Jl 2 100 ft S ft 1 e 8 2 ; 2 100 • 2 & « 2 9 • 11 « 4. a a 0 Jk * Jl 2 100 & 52 9 s 9 % e 2 Total respondents Non-respondents a 2 a 6 2 8 2J k ft 108 ft * a a • ft 1 2 6 eJ & « 78 2 « 9 9 » C 9 9 • Jl I 80 2 e a 2 2 ft ft 2 9 ft 29 * 62 ft 100 ft a » 2 ft 1 2 23 * Jk 2 100 2 « 16 ft 50 « 0 105 ft 171 ft • ft s 1 ft- 23 • Jk I 100 V 0 0 9 0 8 a < 21 9 ft J 2 ft 121 £ 2 * « s • 2 7 17 Jk 100 ft ft 250 266 332 27 42 88 6 23 Jl 100 1 13 86 100 4 13 Jl loo APPENDIX. Appendix T ab le 3 8 . a Comparisons of Methods o f O btaining Lata and Kind o f Potatoes to R a tin g on Q u a lity - Bahed. Kind-of Potatoes Michigan Poor Average Good Total Maine Poor Average Good Total Idaho Poor Average Good Total t- Personal i Telephone Mail {Percent tNumb er £Berc ent {Number;.Barcent sNumber i * s $ • • 2 ft • ft ft ft 4» 'ft 0 ♦ ft i « • i i o • ft 4 4 17 37 Jk 100 3 32 Jl 100 « 2 Ji 100 Total respondents Non-respondents s 0 0 s 16 0 35 ft Jk1 ft 94 ft 0 ft 0 * 2 0 21 ft ft 6o « a I f t ft 1 ft ft 51 • • 13 £ 8 s Jl { 100 i 2 £ 8 a 23 s Jl £ 100 2 * ft • 0 a e 2 100 0 100 * * ft ft » ft 9 : 6 *J i » 71 £ £ £ ft 5 ft 15 ft J l « 65 « a $ s i ft J k s 84 £ 5 ft 0 • ft ft 19 40 ft J i • 100 ft ft ft ft 9 ft 38 ft J l 0 100 • 9 ft • ft 4 ft J i * 100 • 0 ft 0 4 : » • ft 27 $ 17 ft 59 i 32 • 9 6o « 0Ji ft ft 100 ft 5 0 ft « • ft ft 8 • 7 ft 35 ft 32 ft J i ft 61 91 9 100 0 < f 0 0 f t ft ft ft 0 - ft * 2 2 5 0 ft ill • J 8 ft 124 ft 100 0 0 O 0 S « 211 220 361 66 88 119 APPENDIX B Appendix Table 39* Comparison, of Methods of Obtaining E&ta and Kind of Potatoes to Rating on Quality - Pried* Kind of Potatoes Personal Telephone Mail iPercent slfmaberiPercent Miuober sPercent sflfumber;Average Michigan Poor Average Good Total 2 14 100 i : Maine Poor Average Good Total t 9 s Jk s 100 i 1 5 i Idaho Poor Average 0-ood Total s * 22 % % 3® loo a ... i Total respondents Hon»respondents « 2 * 14 : _J2 » 98 6 8 86 loo • • 1 i i 6 • 7 » JS l t 92 » s s 2. t 6 i 1 i l : M s 100 £ *Jk 1 J 2 t * i i 6 62 « I iu05—111 i^8 i £ 6 « • 9 •9 -21 * 0 e • 40 3 10 j 2 * 13 ^ -93. 2 100 • • 9 t t 1 8 4Jk 2 50 ■* • » • s 6 i 40 * -5k 1 100 1 27 Jk 100 2 27 -21 100 & 9 * 58 » Jl i W r.' t « 24 J Jk * 100 * s s J • « s £ £ £ £ * • : • j 1 23 61 85 1 16 31 100 £ i 2 2? JO 1 99 9 « 206 214 330 71 94 150 9 i i 2 22 JZ& « 100 • 9 « AEKEHDIX B Appendix Table 40* Comparison of Methods of Obtaining Data and Kind of Potatoes to Rating on Quality - Boiled, Kind of Potatoes Michigan Poor Average Good Total Maine Poor Average Good Total Idaho Poor Average Good Total Personal Tele,.hone Mail Percent tMmabers Percent siJurnber .Percent *Number Average 1 i. i £ • 9 t 9 i « l. $ » » 8 & 9 * • 19 9 1 8 • U 7 5 9 s 9 27 9 12 * 14 9 • 0 22 24 57 33 » 68 « -22 2 -Si « 9 _2i 9 81 9 SI -2i e 124 6 100 » 100 100 • 171 * £ 114 100 • 6 w • t £ V V * 1' » z 9 e • • « c 9 m 2 • 4 ; 1 l % ft 4 c * • 8 ft 8 S 10 I li 23 15 23 p J z • 90 _32 $ .22 9 z 88 SI -22 100 100 100 £ 81 I 76 0 100 t 100 » > 2 $ ft 0 • z I 10 8 2 • » 5 s 7 9 3 3 « 24 I 6 0 19 37 # 12 23 31 68 61 # 61 sJS • A i SI .12 100 120 loo « 82 t 53 ft 100 9 100 ,6 et £ « f 9 9 « • f t 9 J i 9 % 9 Total respondents Non-respondents 253 34 m 31 391 84 APEEfflIX C Includes map of area, and questionnaires used in surrey* ffERMMLl SCALE'- ON£«»ON£CWARTER INCHES - 4-/1H.E5 -" SlONHSAn NAmM33M9 mi*A/WaUi3 WHjS 3 n n a j . i o 118 MICHIGAN STATE COLLEGE East Lansing, Michigan D w e llin g address Date Number of calls f i [““1 1— Hour 1 I f no in te r v ie w why? Does your s to re have th e choice o f potatoes th a t you p re fer? VJhat is the name and address (nearest cross street) of the store where you made your last purchase of potatoes? Store name A ddress_______ At what store do you buy most of your food? Store name Address_____________ _____________ If purchased potatoes at store other than rest of food, i/Aiynot at sameplacet? Did you f '}, husband [" L son f "j, daughter [ j, other | ] make the last potato purchase? Were they from bulk display? P "]? or were How many pounds were there? they already packaged? □ lbs * About how many days will that many potatoes last you? _________ days0 j About hew much did the potatoes cost? Y e s No If yes, about _ Were they the kind you usually buy? howlong have you been buying this kind? « For size, do you prefer potatoes Mixed sizes j } Small | 1 Medium | ) or Large | ] Do you know what state the potatoes which you now have are from?(Would you like to go look on the bag?) (Check as many states as you have on hand) Idaho f 1 Maine 1 j Michigan f | Other [~ What were your reasons for selecting them? on each)« { Don't know j ~) (If from two states indicate reasons H9 We would like to^have you rate the potatoes on hand as "Good", “Average", or "Poor" (Interviewer indicate what state potatoes are from): Use Maine or Michigan whenever consumer has them on hand. Idah0 C Z 3 m i n e t 1 Michigan 1 \ Don«t know \--- 1 { Other { Average Average Good or Pair Poor How was: Cooking Quality^ Size Color after cooking Blemishes or defects Cleanliness \ Taste And how would you rate them for general desir­ ability or quality v&ien considering everything? What factor in potatoes do you consider the most important? How would you rate them on quality, cooked as follows: Method of Cooking Good Average Poor Hot Cooked This Way How rate them when Boiled How rate them when Mashed How rate them when Baked How rate them when Pried Other ways _______ If you vrere to make one criticism of your grocer,s choice of potatoes what would it b e ? ___________________ | _______________ . ______________________________ AID now for some information to compare with potato use. About how many bunches or lbs. of carrots did you us® last week ; stalks or bunches of celery? □ About how much xras your grocery bill, including milk, last week |j_ J- Including yourself, how many people are in your family (eat with you) j About how long have you lived in this neighborhood? Do you have a telephone? *Q Yes j. ears. What is your husbands occupation? In what country was your husbands Father born Family income range per week* Address Under $55 ____ $55”$90 Race: white__ other Over $90 * 120 MICHIGAN STATE COLLEGE Michigan State College is cooperating with your neighborhood grocers in trying to find your likes and dislikes in potatoes. We would like for you to answer these few questions to help us get you better potatoes. Name Address Phone No< Date Hour Does your grocery store have the kind of potatoes that you prefer? _______ What is the name and address of the store where your potatoes were bought? (To the nearest cross street) Did you } j; your husband □ make the last purchase. Were they from bulk? □ J son H i daughter ( [; don’t know ] j; other □ Or were they packaged? |'" | How many pounds were therej f lbs. About how long will this many potatoes last you? □ days. How much did the potatoes cost? Were they the kind you usually buy? Yes ( ' ]; No f j For size, do you prefer your potatoes mixed, small, mediuiq. or large, □ □ Do you know what state the potatoes you have on hand now are from? Maine CHI5 ldal10 1 I? Michigan f" f; Other j (. Don’t know [ What were your main reasons for selecting this particular kind? We would like to have you rate the type of potatoes on hand, (interviewer indi­ cate here if it is Maine Michigan j 1? Idaho [ ^ Or Other 1 1 .< How would you rate them on quality How would you rate them on the Not when: following characteristics? How was: Good Average Poor this Good Average Poor Cooking quality ( ) ( ) ( ) Boiled ( ) ( ) ( ) ( Color after cooking ( ) Mashed ( ) ( ) ( ) ( Cleanliness ( ) Baked ( ) ( ) ( ) ( Size ( ) ( ) ( ) Fried ( ) ( ) • ( ) ( Blemishes and defects ( ) ( ) ( ) Waste ( ) ( ) ( ( ) ( ) General desirability ( ) ( ) ( ) ( ) ( ) ) Including yourself how many people in your family (eat with you) _ About how much was your grocery bill, including milk, last week §_ (ANY COMMENTS OVER) ( ) ( M ICHIGAN STATE COLLEGE BETTER POTATOES This store is cooperating with Michigan State College in a study to find what you want in potatoes. W ill you help them and yourself by giving us your answers to these few questions? We would appreciate it. Don't sign your name, unless you want to. Even if you don’t answer all the questions, please return the questionnaire. What kind of potatoes do you have on hand now? (If none, check here □ and show last used.) Maine □ Idaho □ Michigan □ □ Don’t know □ What size are they? Mixed Size □ Small □ Are they the kind you usually buy? Yes □ No □ Medium □ Large □ Did you buy your last potatoes already packaged? □ Or from bulk? □ About how many pounds of potatoes did you buy last tim e?................................. lbs. About how much did your last package of potatoes cost you?.................................jf. About how many days will a purchase of the amount you have now last?.............................................. What were your main reasons for selecting them? (Check two or three reasons.) Satisfactory experience with them □ Color.................................................................... □ Recommended by someone............................ □ Size...................................................................... □ Convenient package ...................................... □ C leanliness........................................................ □ Good appearance ............................................ Q Someother reason............................................. □ P rice................................................................... □ By the way, do you remember the name and address of the store they were bought at? (To the nearest cross street.) Does your food store usually have the kind of potatoes you prefer? Yes □ No □ Including yourself, how many people are there in your family?................................................... About how often do you buy potatoes?..........................................................................................days. POTATO QUALITY Please C IRC LE the phrases below that describe the potatoes you are rating. Please rate Good, Average, Poor on Quality Factors. Average Poo Good COOKING QUA LITY....................................................................................................................... ....... Cook mealy; firm when boiled; cook soggy; fall apart; cook unevenly; cook slowly; cook easily. COLOR A FT ER COOKING............................... - ................................................................................. Cook white; dark streaks; dark spots; cook dark. CLEANLINESS.................................................................................................................................... ....... Clean; fairly clean; slightly dirty; dirty. ....... Size O.K.; too small; too large; too variable or mixed; too few medium. BLEMISHES OR D EFEC TS................................................................................................................. Rotten spots; soft ro t; bruises; holes; scabs; bad or hollow centers; wrinkled; sprouts; streaks inside; spots inside. ....... Sweet; neutral, tasteless; bitter. GENERAL D E SIR A B ILIT Y .............................................. .......................................................... .. □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ □ Which of the above do you consider the most im portant?....... How would you rate them on quality when cooked as follows: Poor Average Good □ □ Boiled................................. ........................... □ □ □ □ Mashed............................... .......... □ □ Baked................................. ........................... □ □ □ □ Fried................................... .......................... □ □ □ O ther.................................. ........................... The next time you buy potatoes will you try to buy the same kind described above? Yes □ No □ Please give your street address to nearest cross street....................................................................................................... Any comments? ( If you w ish to w rite more, tu rn the page over and w rite on the back.) THANK YOU FOR YOUR HELP. Michigan State College East Lansing JU S T PU T T H IS BACK IN T H E EN V ELO PE AND MAIL. NO STAMP NEEDED.