IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIILIIIIIIIIIIIIIIIII*L, 3 1293 1042 This is to certify that the thesis entitled A SURVEY OF CONSUMER CHARACTERISTICS AND PERCEIVED RISK AS THEY RELATE TO GENERIC SHOPPING presented by Beth Axelrad has been accepted towards fulfillment of the requirements for Master of Arts degreein Advertising Major professor jf/M //_///%n {@75/4 I Date claim é} km 0-7 639 LIB R A R ‘1” Michigan State Universfiy OVERDUE FINES: , . 25¢ per day per item , 1"?“ Eff , RETURNING LIBRARY MATERIALS: m Place in book return to remove ‘ ,L‘ 3..-, . I "ll 1.: "‘9‘wa "2' charge from circulation records W :5}: JVQ ‘ l' a @1ch FALI984 g 221% W; ‘ Specia'Ermission _NO renewal $93 a 1'99“: A SURVEY OF CONSUMER CHARACTERISTICS AND PERCEIVED RISK AS THEY RELATE TO GENERIC SHOPPING By Beth Axelrad A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Advertising 1980 ABSTRACT A SURVEY OF CONSUMER CHARACTERISTICS AND PERCEIVED RISK AS THEY RELATE TO GENERIC SHOPPING By Beth Axelrad This study was conducted to better understand the nature of generic shoppers and how to better market generic goods. A survey was conducted among 200 adult shoppers to determine if generic and non-generic shoppers were significantly different on selected demographic and behavioral dimensions. The study also sought to determine if the psychological variable "per- ceived risk” would help explain generic shopping behavior be- cause generics do not offer consumers the traditional risk- reducing information (i.e., labels, brand names). Data was gathered by personal interview. The questions covered brand behavior and brand perceptions. Perceived risk was measured by asking respondents to indicate how much risk they perceived when buying generics by summing responses on an ordinal-level—scale for six types of risk. Results show that generic shoppers view generic products as equal in quality with national brands. Also, a significantly higher percentage of generic shoppers perceived generics to be low-risk items. Therefore, marketers of generics need to consider more than just the money-saving benefit of generics. They must expand their thinking to include consumers' perceptions of quality and risk. ACKNOWLEDGMENTS To my thesis director, Dr. Bruce Vanden Bergh, I wish to extend my sincere gratitude for his constant patience and assistance throughout the entire ordeal. Knowing how much he had on his mind, I appreciated the time he always managed to find for me. I would also like to express my thanks to Dr. Martin Block for his continual support throughout my entire graduate education. In addition I want to thank Dr. Dean Krugman for listen- ing to me every time I came bursting into his office with my computer printouts. Lastly, I want to thank my special friend Greg Stropes for making the worst times at MSU seem much better. ii TABLE OF CONTENTS LIST OF TABLES. I. INTRODUCTION. The Generic Industry What Are Generic Grocery Products?. History Of The Generic Line The Future Of Generics. Why Study Generics Thesis Overview. II. LITERATURE REVIEW . Generics Studies Purchasing Behavior. The Price- ~Quality Relationships. Perceived Risk . . . . Information Handling III. HYPOTHESES. Differences Between Generic and Non-Generic Shoppers . . Perceived Risk and Generic Shopping. IV. METHODOLOGY Sampling Data Collection. Questionnaire Design Questionnaire Construction Stage 1. Focus Group Results Stage 2 . . . Questionnaire Construction Stage 3. Operationalization Of Variables. Validity And Reliability . Logical Consistency iii TABLE OF CONTENTS (Continued) Transferability. . . . Measurement Validity . Measurement Reliability. Internal Validity. . Statistical Significance Strength Of Association. External Validity. V. RESULTS. Treatment Of Data . Data And Results. VI. CONCLUSION APPENDICES A Unbranded Items Currently Available B Questions Asked Of The Focus Group. C Questionnaire Code Sheet. D Occupation And Social Status Scale Code Sheet E Questionnaire F Focus Group Manuscript. REFERENCES iv Page 76 76 77 79 79 8O 81 81 84 109 116 118 120 123 124 129 171 Table 10 11 12 13 LIST OF TABLES A Product Category Breakdown Of the Frequency Of Trials Of Generics By Members Of The Boston And Chicago Sample. Household Income As Reported By Progressive Grocer in March 1979. . . . How Generic Products Move A Typology Of The Criteria For Evaluating Research. . . . . . . . . . . . Difference Between Generic And Non- Generic Shoppers Based On Spending From A Food Budget. . . . Different Methods Generic And Non-Generic _ Shoppers Use To Budget Their Food Bill. Qualities Looked For In National Brands By Generic And Non-Generic Shoppers Benefits Of Generic Products As Per- ceived By Generic And Non-Generic Shoppers. Differences Between National And Generic Products As Perceived By Generic And Non— Generic Shoppers. Active Search For News Of Sales Or Coupons By Generic And Non—Generic Shoppers. . Information Sources For Brands Purchased Of Generic And Non-Generic Shoppers. T—tests: Demographic Differences Between Generic And Non-Generic Shoppers. Differences In Income Of Generic And Non- Generic Shoppers. 17 19 55 75 84 86 87 88 90 91 92 93 96 LIST OF TABLES (Continued) Table 14 15 16 17 18 19 20 21 22 23 Frequencies Of Degrees Of Risk Associated With Generics. Differences Between Generic And Non- Generic Shoppers Based On High And Low Risk Perception Perceived Risk By Product Category And Brand Purchase Behavior. . . Generic Products Tried And Perceived Risk. Difference Between High And Low Risk Per- ceivers By Repeat Buying Of Generic Pro- ducts. . . . . . . . . . Difference Between High And Low RiSk Per- ceivers With The Decision Of Never Con- sidering Buying Some Generic Products. Perceived Risk By Generic Products Never Consider Trying. Differences In Active Search For Informa- tion By High And Low Risk Perceivers How Often High And Low Risk Perceivers Look For Information . . . . . . Summary Of Significant Findings Of Present Generic Study. . . vi Page 98 99 100 102 104 104 106 107 108 110 INTRODUCTION The Generic Industry What Are Generic Grocery Products? Far beyond expectations of the forecasters in the gro- cery industry, generic products have been quite successful (Coyle, 1978; O'Neill, 1978). Also known as "plain labels" (Burck, 1979), ”plain wrap" items (Sales & Marketing Manage- ment, 1975), "no names" (Heller, 1978), and "no frill labels” (Selitzer, 1978), they are easily recognized by their two color label, either a combination of black and white or olive green and white. Now some manufacturers are introducing red stars on their labels. Although there is less fancy labeling, grocery industry sources report that savings by using a two color label are only a small percent of the penny per unit savings (Coyle, 1978, p. 77). In most stores the generic products are shelved in a separ- ate section. Taking the Jewel Tea Company's lead, many stores house them close to the beginning of the store traffic pattern. According to Darrell Schmuker, Vice President, Foods Merchan- dising at Meijer Supermarket, Inc. placement of generic pro- ducts has been tested. Results indicate that consumers prefer to shop for generics in one section. Contrary to Mr. Schmuker's industry sources, the focus group conducted for this research 1 generated the opposite feelings. One individual said, "I find it very aggravating at Meijer's to separate" (Appendix IF,P- 139) The reason for this is that women liked to compare prices. They found it inconvenient to go back to the generic section if they forgot an item. Several women felt the Meijer's and Shop-Rite sections were "totally” out of the way (Appendix1Ffi'P- 139%.However, after a recent trip to Meijer's (June 28, 1980) it seems a few products had been integrated into the grocery pattern for comparative pricing. A third characteristic that sets generics apart from other products is the content of the package. When Jewel originated the marketing of American generics, their philosophy was ”smart shoppers don't always need top grades in the grocery items they buy, and they certainly don't need to pay for the advertising and promotional costs of national brands” (Coyle, 1978) Therefore, besides the savings from no additional pro- motional expenditures or advertising from a national manufac- turer, the product is of a standard, less fancy grade. When packaging generics, the company considers these products ser- viceable, but not extra standard or fancy to be appropriate for the generic label. This means cracked peanuts go into the peanut butter, paper toweling lacks absorbency, and dish detergents have less perfume (Coyle, 1978), Meijer brand generics are manufactured under the same philosophy. For example, their detergent contains less soap so more liquid needs to be used. In the food line, their generic corn may have been harvested later than planned, thus it is a bit tougher than the standard or fancy grade. Therefore, based on label, shelf placement, no promotion, and product quality, generics are an unique concept in the grocery industry. History of the Generic Line Although Jewel instigated generic marketing in the United States, they originated in France in 1976. The chain that introduced unbranded merchandise was Carrefour. The unbranded items were called ”produits libres.” Literally ,translated it means free products. The free connoted free- dom from paying for expensive promotion and packaging. Pro— duits libres were introduced with the support of a $3.3 million advertising campaign. The prices were up to 30% lower than branded merchandise. A Fortune magazine article reported they were "a phenomenal success" (Burck, 1979). Following Carrefour's good luck, two other French chains came out with their own versions. They also reported growing success in their generic areas (Coyle, 1978). However, the one catch that probably contributed so heavily to their suc- cess is that in French supermarkets there were no private label goods which offer any savings to the consumer. With knowledge of this, Jewel introduced their ”no- names” in August of 1977. Jewel noticed a trend among con- sumers to "tell it like it is" (Coyle, 1978). so Jewel offered a non-advertised, lower cost package and lesser quality product that met nutrition standards. Some shoppers felt they were really low end, or second grade private labels, without the colorful labeling. Consistent with the new straight forward outlook, comparing generics to private labels was not necessary. When private labels appeared in the stores, they were intended to look like national brands but offer a savings. The reason being that the middle class American did not want to be embarrassed by having to purchase a lower priced brand (Burck, 1979, p. 71). However, according to Burck of Fortune magazine, during today's inflationary times there is no stigma attached to saving money. By reading the market well, Jewel tapped a successful oppor- tunity for growth. According to early 1980 Selling Areas-Market— ing, Inc. figures, generic sales are increasing. Based on the study of the market, ”the effect of increased availability, both as the result of more operators handling generics and more lines being added,” generic tonnage has increased by 60.4% in one year (Dietrich, 1980, p. 127). At the same time private labels dropped by 6.6% and advertised brands dropped by 2.7%. The Future of Generics The future of generics is a cloudy subject. Various sources have differing forecasts for the permanency of this phenomenon. One debator who questions the sustained success of generics is Robert O'Neiflq editor of Progressive Grocer. He thinks that consumers are going to become dissatisfied with the inferior quality. His major concern is that there I may not be a 'sufficiently large segment that is willing to settle for second and third rate products" (Food Product 222., 1979, p. 46). In 1979 Nielsen looked at the sales rates of generics in the stores which had just picked them up. Their results showed less growth than in the generics offered earlier. Despite some pessimistic views of the future of generics, a good many sources see sustained or increased growth. Accord- ing to Marsh Blackburn, president, Sales Force Cos. it takes a good marketing ability to keep generics on demand. Those retailers that have done good marketing have had outstanding success with their generics (Edwards, 1979, p. 88). As of April 1980, 43% of all United States supermarkets stocked generics (Bell, 1980, p. 60). In 1979 10,300 stores were carrying them . One year later that increased by 41% so now 14,600 stores carry them. Again, Clark Bell, marketing columnist for the Chicago Sun Times, states that the problem of sales erosion is weak marketing. However, in the short term the opportunity for generics remains substantial in light of continued inflation. The inflation rate, as posted by Washington in May, showed the average increased cost to consumers for the first five months of 1980 as 15.3% (State NEWS, 1980). Food and beverage prices rose slightly .3%, but average take home pay fell .9 percent. This means that for the past year take home pay has been on the decline. Furthermore the Labor Depart- ment predicts additional increases in food prices this summer. A ten year forecast done by a Wharton econometrics group predicted inflation to increase between 5 and 7 percent, With inflation as it is and its projected growth, saving money is going to continue to be important (Business Outlook, 1979). Why Study Generics Even though generics boast of no advertising or promo- tion, it does not mean that it excludes a good marketing pro- gram. By looking at what others have forecasted for the life cycle of generics, marketing becomes a very important factor in relation to their staying power. As mentioned earlier, Marsh Blackburn at the 1979 Food Marketing Institute Conven- tion is quoted as saying, "Generics require marketing ability... and only a few retailers have handled them well so far. ...Generics are around for awhile as are the marketing challenges that go with them” (Edwards, 1979, p. 88). In his article on generic products, Bell (1980) provides a basic rationale as to why retailers neglect a strong marketing program for generics. Brand name manufacturers as well as the outlets selling private labels are threatened by pushing generics. The pattern most retailers follow is a heavy intro- duction when the store begins to carry them. Sales climb then decline until a plateau is reached in about a year. There they stay. Bell concludes his article by saying that they represent an opportunity for the aggressive merchandiser. However brilliant the marketing plan is, the ultimate success of the product is determined by the consumer. Accord- ing to Myron Glassman (1979), consumer soverignity is the major reason why generic groceries have boomed. He attri- butes this mostly to the consumerism movement. They recognize that lower grade products with less esthetic appeal contain the same nutritional value as the fancy grades. This aware- ness has turned the informed consumer away from advertising. He is willing to give up the fancy packaging and advertising for lower prices. Glassman believes that the consumer no longer wants to be fooled by extrinsic product characteristics. One traditional extrinsic product element that is included in marketing strategies is a brand name. Kotler states the purpose of branding may be to connote quality or provide an opportunity for endowing your product with a unique story for differentiation (Kotler, 1976, p. 191). Cunningham, as early as 1956 researched the area of branding to provide executives with the answer to the question, "is it worthwhile to spend hundreds of thousands of dollars in an effort to identify its products in the buyers' minds with a brand label” (Cunningham, 1956, p. 116)? The results of his study indicated that indeed brand loyalty does exist within individual product groups which justify the expense. This phenomenon is mentioned here because generics intentionally counter the status quo. That is one reason why continued study is necessary. What variables are then related to draw— ing consumers to generics? It has been a confounding area for marketers. When generics were launched into the supermarkets in 1976, the advocates saw them as a "price-break for needy families” (Selitzer, 1978, p. 1). Their differentiation rested on their low price. Yet, as studies seem to indicate, the generic purchaser is both the lower income, blue collar individual as well as the upscale professional. Middle in- come moderately educated individuals purchase fewer generics (Food Product 921., 1979, p. 53). Farley in 1964 developed a theory of brand loyalty based on information search; His theory predicted that lower income families are most likely to switch brands. Therefore, they seem most likely to try generics. Thus, the difficulty arises in creating an effective marketing plan based on income due to the nature and polarity of generic shoppers. Frank in his 1967 article evaluating brand loyalty as a segmentation tool, supported segmentation strategy as a way to market a product on the basis of consumer needs and wants. The eventual goal is to increase profits by segmenting. The results of his study led him to conclude that socioeconomic or personality characteristics do not help explain brand loyalty. He suggests future researchers look at sociological and psychological factors as well. The focus of research done on generics in the past has been basically demographic. (Coyle, 1978; Murphy, 1979; Zbytniewski, 1979). The studies show small differences between groups. Therefore the major goal of the following research was to look beyond just the socioeconomic characteris- tics of generic shoppers and introduce perceived risk, a 10 psychological variable, as a basis for segmentation. Accord- ing to Raymond Bauer, almost all consumer behavior involves risk (Bauer, 1967). He mentions that the greatest risk occurs when buying expensive items. However, any action a consumer takes may produce unpleasant consequences of which he cannot anticipate. Since generics are packaged with little informa- tion on the label and no advertising, it was suggested that there is more risk involved in their purchase than in the purchase of branded items. Since generics are relatively new on the market an opportunity still exists for developing solid marketing plans. As many marketing specialists have noted, the need for improved strategies exists.‘ Due to the fact past research has examined the traditional segmenting variables, a new approach taking into account current trends is necessary. As Glassman stated, the ultimate success lies with the consumer. Therefore, a clear picture of who that consumer is can help the retailer segment better, improve his marketing strategies which keep generics on the shelves and allow consumers money saving products. As will be addressed in the literature review in greater detail, the earlier generic studies did not cover a broad range of variables as they related to generic shopping. They did not attempt to go beyond simple descriptive research. Percentages were reported and some chi—square analyses were 11 performed on demographics. The purpose of this study is to extend this limited body of knowledge, by looking at variables that have not been studied in relation to the generic consumer. Some of the same demographic variables will be looked at using interval—level statistics. The new variables introduced to explain this phenomena will be analyzed using the appropriate statistical techniques in order to determine the significance of the variable as it relates to the generic shopper. The major research questions the study attempted to answer were divided into two broad areas. First, differences between generic and non-generic shoppers were examined, followed by differences based on perceived risk levels. In summary, the basic question asked about generic and non- generic shoppers was: Are there significant differences between generic and non—generic shoppers in regards to: (a) their use of a budget for groceries and how they budget (b) the qualities they look for in national brands (c) the primary benefits of generics (d) the differences they perceive between generic and national brands (e) their use of information sources (f) their need to search for news of sales or coupons (g) their occupation (h) their education 12 (1) their income (j) the size of their household (k) the ages of their children living at home The question asked concerning high and low risk perceivers was: Are there significant differences between high and low risk perceivers in regards to: (a) brand behavior (b) first generic products tried (c) generic products never considered trying (d) search behavior I (e) overall generic shopping 13 Thesis Overview The rest of the thesis will be divided into five sections. Review of Literature Hypotheses Methodology Results Conclusions The review of literature will cover a broad area. Only one academic study has been conducted on the generic consumer. Two retail studies have also been published. However, they lack well defined measures. Their weaknesses will be dis- cussed in light of the thesis research. Due to the small amount of literature on generics, topics from other areas in marketing were reviewed and inferences were drawn about generic consumers. Studies about private brand consumers were reviewed along with purchasing behavior articles that grouped consumers by social class, income, or personality. The literature also examines perceived risk research. Much of this has been studied in light of information handling (Cox, 1967; Cunningham, 1967). This section, in addition, discusses the price-quality relationship since the generics' prime benefit is low cost. The hypothesis chapter then takes each one of these variables and predicts its relationship to generic shopping. 14 The major variables studied were: income, education, size of family, age of children, budgeting, information handling, and perceived qualities and benefits of generic and national brands. The hypothesized relationships of perceived risk to product trial and most common brand category bought is also included in this section. The methodology section describes the Operationalization of these variables, sampling technique, and data collection. Since a perceived risk scale was administered, discussions of the reliability and validity of the methods are also found in this section. As in all research there were certain limita- tions to keep in mind. The problems of convenience sampling and control of questionnaire administration are presented. In addition, the steps leading to the development of the ques— tionnaire with the results from the focus group are included. The results section contains descriptions and inter- pretations of the data. The raw data from the SPSS computer runs have been tabled and presented for comparative analysis. A discussion of the significance of results follows the data. Lastly, the results are examined in light of their impact of future marketing actions and research. Since the orientation of this research is to provide additional informa- tion to the ”practical” body of knowledge, results will be discussed concerning their contributions to the retailing area. LITERATURE REVIEW The following review covers five areas of literature. Due to the fact that published generic studies are scarce other variables from different areas were researched to pro- vide additional background information. To begin with, the generic studies are analyzed. This section is followed by a discussion of purchasing behavior in regards to private brands and the characteristics related to various groups of purchasers. The price~quality relationship is next. A variety of literature that relates price to perceived quality is the focus of this section. The last two sections review perceived risk and information handling as a way of coping with risk. The discussion of risk looks at the constructs of risk as they apply to the generic study. Although the various risk relievers are mentioned, information handling is emphasized due to its importance in risk reduction and its applicability to marketing planning. 15 l6 Generic Studies Currently there are only three published generic studies available as public information. Although they touch on the same topics as the research-in this thesis, the usefulness of their data is questionable. Two of the studies used fre- quency counts as their only measure of results. The other study conducted chi square tests on some of the data. These provide an introductory insight into the area, even though their analysis is not rigorous. The first study which came out was conducted by Progressive Grocer, an industry magazine, following Jewel's launching of the new generics into the market (Coyle, 1978). A total of 400 telephone interviews were Conducted in two cities, Boston and Chicago. This research occurred when only a limited number of generics were in the market on a trial basis. Table 1 reports the percentage of individuals in the sample from this study who reported having bought at least one generic item. From this information, it can be observed that in both samples, paper or plastic products, canned vegetables and fruit and canned juices accounted for most trial pur— chases. Most of the other products showed a much lower rate of trial, ranging from 6% to 37%. As reported here, this infor- mation provides no insight on who is the generic shopper. However, 17 TABLE 1 A Product Category Breakdown Of The Frequency Of Trials Of Generics By Members Of The Boston And Chicago Sample Boston Chicago Paper or plastic products 83% 59% Canned vegetables 45 52 Canned fruit 44 47 Canned juices . 44 33 Soap/laundry products 37 24 Peanut butter or preserves 27 18 Canned fruit drinks 20 27 Canned tomato products 20 24 Mayonnaise or salad dressing 20 15 Canned tuna 19 6 Canned soups ll 6 Pet food 8 6 Canned softdrinks 3 6 Source: Coyle, J.S. ”Why Jewel Did It, How Consumers Respond, What the Risks Are, Where It All Goes From Here." Progressive Grocer (Feb. 1978) p. 75-84. the comment in this article about the research was that paper products are "often perceived as a safe first step" (Coyle, 1975, p. 80). From this product analysis and casual comment by the author the theory of perceived risk was considered as a possible explanation of generic shopping. This study also looked at demographic characteristics. The general conclusions were that full time housewives with limited incomes and large families tended to try generics. However, the Boston and Chicago samples did produce a few differences. 18 Two very interesting areas the study researched were what the consumers felt makes generics different and cheaper. In both Boston and Chicago, most felt it was low price which made generics different from other branded goods. The absence of advertising and fancy labeling is what was seen most often as what makes generics cheaper. These questions seemed to be appropriate as discriminators of generic or non-generic shop- pers. Perhaps each group looks at generic products differently. Therefore,these questions were employed in this thesis' com- parative study. In December of 1978,Progressive Grocer conducted a second study. This time they concentrated on who was the generic shopper not on the product categories. The sample size was 595 shoppers at a major East Coast grocery chain. The results were reported in percentages. Anyone who had ever bought generics was included in the generic group. Their logic may be faulty in this area. It would seem that the consumer who buys generics habitually is different than someone who tried them in the past but no longer buys them. For one reason or another they stopped purchasing generics. Thus, it is question- able whether their characteristics are similar to the loyal generic shoppers. Secondly, their results, reported in percentages, do not show major distinctions between generic and non-generic shoppers. For example in the demographics, the results for income indicate small percentage differences. Table 2 reports this data. 19 TABLE 2 Household Income As Reported By Progressive Grocer in March 1979 % Who Ever Non-Generic Household Income Bought Generics Shoppersa Under $10,000 49% 51% $10,000-$14,999 53 47 $15,000-$19,999 50 50 $20,000-$24,999 58 42 $25,000 or more 48 52 aResearcher derived these figures from data provided. Source: Zbythiewski, T. and Heller, W.H., "Rich Shopper, Poor Shopper. They're All Trying Generics." Progressive Grocer, (March 1979) pp. 92-110. ‘No statistical significance tests were performed on this data either. Whether income category $20,000-to-$24,999 really describes generic shoppers is debatable because the other percentages are so close. Although they went beyond demographics, to attitudes towards quality, there were no statistical tests performed in order to judge whether non- generic shoppers have significantly different perceptions about the quality or characteristics of generics. Probably the most descriptive study on generic consumers in literature was conducted by Murphy and Laczniak. (1979). They administered a telephone questionnaire to 429 consumers 20 during April 1978. Their results of purchase frequency of nine goods studied were similar to the ones found earlier. They also had the respondents designate themselves as generic shoppers, by the various product categories. Of the 347 who answered the question, 34.8% considered themselves regular generic purchasers. In addition they probed into price and quality perceptions. The results were reported in terms of the entire sample. In other words, they did not look at perception differences by generic and non-generic shoppers. They found two thirds of their sample believed the price was "slightly lower,‘ while 27% found generic prices "very much lower," and 7.3% found them the same. With regards to quality, generics were seen as being of "average” quality compared to other brands by 74% of the sample. Only 7.3% felt they were ”very much below average” (Murphy and Laczniak, 1979). Furthermore, Murphy and Laczniak tested the demographics they examined by using chi square tests. They categorized age, income, occupation, number of persons in household, and education into nominal groupings. These are all variables which can be looked at intervally, which is a higher level of measurement. Of the demographics analyzed, number of persons in the household and level of education were significant. The major weakness of Murphy and Laczniak's study is that is does not take advantage of statistical methods to finely analyze the data they have available to them. Being that they did have a question which differentiated generic 21 from non-generic shoppers they could have looked at their data comparatively. The research conducted in this thesis looks at many of the same variables they studied. However it takes the data one step further. In the final section of the article, they conclude, ”the analysis of the demographic features of the survey group does not suggest an identifiable market seg- ment for these products" (Murphy and Laczniak, 1979, p. 14). The step taken in this thesis is to compare the demographics again using a more powerful statistic (t-test), and compare perceptions of generic products by generic and non-generic shoppers using a chi square test of significance. The goal of the research is to discover if the generic shoppers have any differentiating characteristics from non-generic shoppers. Past research has already demonstrated that few demographic differences exist between generic and non-generic shoppers. Therefore, in addition to reexamining demographics, and internal and external product quality perceptions, by generic and non- generic shoppers, perceived risk was introduced to explain the various purchase frequencies of the different products. Thus, the following research was intended to provide more information and explanation about generic shoppers as an iden- tifiable market segment. 22 Purchasing Behavior As stated in the introduction, many different sources of information were consulted from which inferences about gen- eric shoppers were drawn. One of the areas studied was private label customers. Private brands or store brands, as they are commonly called, were introduced around 1958. Their introduction was triggered by the same mechanism that inspired Jewel to create generics, increasing competitive pressures (Frank and Boyd, 1965, p. 28). At the time Jewel began selling generics they had an 11% share of the market (Murphy and Laczniak, 1978). Since private labels as well as generics were introduced in an effort to differentiate from the offerings of national brand manufacturers, parallels may be drawn between the type of consumer who may look for these brands. In 1965, Frank and Boyd asked the same question about private-brand-prone customers that had been asked about generic customers; are they an idenfi- fiable market segment? They discussed that prior to their re- search a study sponsored by Good Housekeeping showed that only slight differences exist between the socio-economic status of private label and manufacturer's brand consumers. Frank and Boyd used consumer panel data. They did a pro- duct by product analysis examining the extent of association between "the expected level of a household's private-brand- proneness and its socio-economic, consumption, and store shopping behavior" (Frank and Boyd, 1965, p. 30). By using 23 partial correlation coefficients to judge the degree of association between the demographic variables and private- brand proneness they found no or small differences. Four variables which did reveal some significance were: number of persons in family, education, number of cars, and total consumption weights. Private—brand proneness increases as do these var- iables. Based on economic theory, John Myers in 1967, looked at private-brand proneness as an attitudinal construct. The attitude is centered around the theory of differential price elasticities. He assumed that consumers judged private brands on the basis of price and quality. The combinations of high, low, or same price, with high, low or same quality in regards to national brands created a typology. Each combi- nation may identify a potential market segment. The objectives of the research were to look at private brand attitude in relation to price consciousness and price quality comparisons along with store preference. Consistent with Frank, Massy and Boyd's (1967) research, Myers (1967) found that socio-economic variables were not strong predictors of differences in private- brand attitude. Due to the fact that Myers checked the validity of this attitudinal scale by examining the degree of distortion which arises from differences in store preferences, it is assumed the private-brand attitude is a sensitive measure. Yet, as with previous research, he found modest significance. The social, 24 economic, and psychological criterion variables logically fit as possible determinants for private brand proneness, however they predict little. Therefore, these variables may not identify meaningful differences. Based on this assumption, studying consumption rates, may lead to a better profile of segments of consumers. Frank, Messy, and Boyd (1967) analyzed socio-economic and demographic factors as well as consumption figures for 57 grocery prOducts. A consumer panel provided the data base, as was the case in the previous studies mentioned. A multiple regression model was the methodological test for significance. The first set of results on socioeconomic and demographic variables were consistent with past findings. The research- ers state that the fourteen variables are "poor predictors of consumption” (Frank, Massy, Boyd, 1967, p. 189) in most of the 57 product classes. The other part of the study examin- ed the stability of purchase rates over time, in which they found moderate stability for a majority of grocery products. For forty-eight percent of the cases, between 50 to 81 percent of the variance is explained. Thus the researchers concluded that consumption rates in some product categories (toilet tissue, tuna, canned vegetables, smme canned fruits, dry noodles and spaghetti, cleansing agents, pancake mixes, short- enings, peanut butter, syrup, packaged desserts) may be a better predictor than the traditional socio-economic variables. 25 This study is important in light of the product cate- gories which produced the highest R2 coefficients. These products do have generic counterparts in the market. Thus, in relation to the generic study, demographics may provide limited explanation, along with education and income. As a result, a wide variety of socio-psychology variables were introduced to augment the degree of explanation from data ana- lysis, as well as questions concerning which brands were pur- chased most frequently. Demographics were not ignored though, because Blattberg, Peacock and Sen (1976) found them closely related to general characteristics of the household. Blattberg, Peacock, and Sen studied buying strategies between similar and dissimilar product categories. They look- ed at 16 out of 18 brand choice segments based on combinations of brand loyalty, type of brand preferred, either national or private, and price sensitivity. They found that similar pro- duct categories often are subject to the same brand choice strategies. However, it was a small proportion of households which used identical strategies (Blattberg, Peacock, Sen, 1976, p. 154), They concluded that buying behavior may be more closely related to "general characteristics" of the household, such as demographics. They suggested further research in this area. With the variety of conclusions drawn from past research on variables which affect grocery shopping, it became a ques- tion of what exactly are the most reliable approaches to 26 studying generic shoppers? If demographic variables provide insight into segmentation strategies then it is important to review the results of past studies. However, the consump- tion rate study (Frank et al., 1967) warns that demographics may not be adequate to explain shopping behavior. Thus, per- sonality variables were introduced to explain more about generic shoppers. Since the overt differential advantage offered by generics is lower price, intuitively income would seem to be related. Families on restricted incomes would be attracted to the generic product. The original generic manufacturers also thought their main appeal would be to this group. Contrary to their original intentions, generics are purchased by those with middle or higher incomes (Zbytniewski and Heller, 1979). In a study conducted by Kunreuther in 1973, he attempted to analyze why the poor may end up paying more for food. He found that many of his low income sample members perceived they had less storage space which forced them to buy smaller quan- tities more frequently. In addition, a budget restriction puts a constraint on purchasing the larger, most costly size. Many of the generic products do come in large sizes but are comparatively less expensive than national brands‘ large sizes. Size and storage may possibly explain why less low income shoppers buy generics. Another dimension about low income shoppers recognized by Goldman (1976) was knowledge level. He wanted to find out if low income shoppers have a more limited shopping 27 scope because they are less mobile geographically and psychologically. As a result they are aware of a smaller set of stores than higher knowledge level consumers. He also tested the hypothesis that lower income consumers have a higher knowledge level because they shop around more in order to find the best deal. In the product category of furniture stores, low income shoppers did not investigate all the alternatives available to them. Goldman classifies them as ”nonthorough" shoppers, which remains consistent with past research. Due to Goldman's findings, it may be related to the generic phenomenon so as to postulate low income con- sumers may not be aware of generics or their benefits. Besides income, social class has been a popular variable for determining consumer behavior. Based on occupation, sources of income, and housing type, Martineau (1958) predicted social class from a weighted score. From the scores he divided respondents into five classes ranging from upper class, upper middle class, lower middle class, upper lower class and lower lower class. His study demonstrated that ”class membership is an important determinant of the individual's economic behavior" (Martineau, 1958, p. 125). He also mentions that the food re- tailer must adapt to the social class status of the neighbor- hood. Interestingly he notes that in Chicago, the Jewel chain's largest consuming group is the middle class. The paradox in this, is that when Jewel first launched generics, as researched by Progressive Grocer, the group they originally appealed to 28 were the lower income, less educated shoppers, Jewel's smaller segment (Coyle, 1978, p. 82). This may have been a sound strategy to increase market share by offering something that appealed to one of the smaller segments to increase patronage. However, in 1979 Progressive Grocer reported that middle in- come families tended to buy generics more than low income families. Therefore, if this data is correct, the market shift- ed in the two year period. In the next section, consumers will be examined by their perceptions of the price quality relationship. 29 The Price-Quality Relationship One of the early questions attacked by researchers was, does price have more of a meaning than economic sacrifice? (Leavitt, 1954) This area was pondered in 1954 by Leavitt. He was interested in determining whether price implies quality, good value or social propriety. This being the case, then, when consumers chose the higher priced brand it could be explained in terms of quality. Leavitt was one of the original experimenters in this psychology-of—pricing field. He gave a sample four product categories, two brands per product. They had to make a pur- chase decision for each product based on price information only. He checkedhow many times the higher priced brand was purchased. His sample was Air Force officers, majors, and lieutenant colonels and graduate students. The sampling was good in the sense that these people probably had little pre- vious information on cooking sherry, moth flakes, and floor wax. Razor blades were also used. From the sample, Leavitt stated that when people believe brands are different they frequently choose the higher priced brand for some products. It occurs more often when the price differential is large. With his sample and the products tested, it is questionable whether the results can be generalized. Do female graduate students produce the same results a housewife would produce? 30 The reason the price-quality relationship is discussed in this thesis is that the primary benefit of generic products is their lower-than national and store brand price. Part of the thesis questionnaire concerned the quality-price perceptions consumers have of generic products. In the case of generics, the 1979 Progressive Grocer survey found that besides lower quality some consumers attributed the lower price to cheaper packaging, cheaper labeling, and no advertising. Thus, price- quality relationship is irresolute. Leavitt looked at high and low priced brands. However, there is also the question of differentials. When it comes to price differentials, some consumer groups may be more sensitive to price changes than others. Gabor and Granger suggest that price determination should be gauged according to the price sensitivity of socio-economic subgroups (Gabor and Granger, 1971, p. 41), They postulate that high price sensitivity is related to price awareness. Thus awareness of the price con- scious segments allows the retailers to set a price right for the market. Price consciousness can also be thought of in terms of deal"proneness. A deal is a reduction in the standard retail price of the product (Webster, 1965, p. 186) Those indivi- duals interested in deals may be the ones attracted to generics for the savings they afford. A deal~proneness index was for- mulated and measured against 45 variables relating to purchasing. Using a regression analysis on panel data, Webster found that 31 the older housewife is more prone to take advantage of deals. Also, high deal prone consumers reveal less brand loyalty, and purchase in smaller quantities. This price-quality relationship has also been tested under experimental conditions. In these experiments, price has been operationalized as a surrogate indicator (Engle, Kollat, and Blackwell, 1968, p. 433). In this research, interest has been in determining if price is used as a surrogate for quality. McConnell (1965) examined the relationship using beer. All three brands of beer were identical, with unknown brand names. A high, medium, and low price was bestowed upon each beer. McConnell found that with a physically homogeneous product the highest priced beer was perceived to have a_better quality. He related this finding to dissonance theory, in that subjects felt that ”you pay for what you get." In 1971 McConnell's findings were supported and extended by the introduction of new evidence. Jacoby et a1. (1971) experimented with price in relation to other cues. Again using beer tasting, they examined the effects of price, composition differences, and brand image on the perception of quality. Their results indicate that when price is used as a quality cue in conjunction with the other cues, it does not significantly influence the perception of product quality. Brand name did have an effect on quality perception, and when it was combined with price information, more significant results were obtained than with price alone. Thus Jacoby et a1. concluded that quality perception results from the interaction 32 of many variables. Price has a significant effect only when it is used alone. Gardner (1970) would partially agree to these conclu- sions based on his study of toothpaste, shirts, and suits. His finding showed that price does convey "some sort of quality information" (Gardner, 1970, p. 34). It surfaced most as imply- ing quality about toothpaste, not so much with the suit. There- fore Gardner concludes, in a generalized sense, that price is most useful when comparisons of the other product's attri- butes are difficult for the consumer to observe. The psychology of pricing is imbued with many character— istics. According to Monroe (1978) in his review of the liter- ature on perceptions of price, prices supply cues to the con- sumer which facilitates the differentiating process between items. It also may indicate quality as well as sway the price conscious individual. Monroe states that whether subjective and explicit infor- mation will be used as part of a consumer's perceptual set depends on his dispositions and past experience with the pro- duct. Part of this process, credited to Bruner by Monroe, is symbolic value which leads to a perceptual assessment. Price is one of the symbolic variables. Monroe then divided up the literature in his review by those studies involving price alone as the single cue and those studies which added other types of information. The conclusion he draws which has re- levance for the generic study, is that "brand name is 33 important and possibly dominates price for relatively in- expensive grocery products” (Monroe, 1973, p. 73). However, one important factor to consider is the variety of research designs and products that were tested. In relation to generics, this may imply that their low cost is not associated with a lower quality, thus providing some insight into why they sell so well. 34 Perceived Risk In the previous section the implied price-quality relationship was discussed. Closely tied to this area is the theory of perceived risk. Risk becomes involved when the individual balances the monetary cost of paying more against the chance of paying less for an assumed lower quality product (Shapiro, 1968, p. 24). Returning to Leavitt's statement quoted earlier, since a lower priced product may imply lower quality, a prospective buyer may judge the higher price to be of a better quality and purchase it (Leavitt, 1954, p. 207). The risk involved is the chance of lesser qual- ity, therefore the higher priced brand becomes attractive. The areasof pricing and risk have been addressed by Horton (1979) who examined risk in light of personality var- iables. In his discussion of risk, Horton mentions brand loyalty as a way to reduce risk. It affects the price-quality risk dichotomy as such: if consumers turn to brand loyalty to reduce risk, they may be relying on nationally advertised brands. Basically these brands with "reputations” ask a high- er price. Thus, this hypothesis evolved into presuming that the more "visible manufacturer with the higher priced brand would be viewed by consumers as a less risky choice" (Horton, 1979, p. 234). He tested six product classes by thirteen personality variables. His results confirmed his first hypo- thesis that subjects low in self confidence and/or high in anxiety tend to select the high identification and price brand. 35 Although perceived risk was not tried in the previous studies on generic grocery products, it has been used to explain attitudes towards generic drugs. Generic drugs afford the consumer one of the same benefits of generic grocery products, that is low cost. As of 1978, generic drug practices had only been moderately accepted. The Food and [Drug Administration does assure that all prescription drugs contain the same ingredients and perform in the same manner, the only difference being priceg(Bearden & Mason, 1978, p. 741). As with generic products, generic drugs do not have heavy pro- motional expenses, as in brand names. (In order to identify generic drug preferences,Bearden and Mason conducted research using risk. The risk model employed was that used by Peter and Ryan in 1976. Individual preference in regards to the purchase of generic drugs equaled the probability of loss from that purchase multiplied by the importance of the loss from that purchase. The relationship was tested by six dimensions of risk: (a) financial, (b) social, (c) performance, (d) psychological, (e) physical, and (f) convenience.fi Corre- lations were calculated between a person's risk and attitude of performing the act. Based on the means reported, it seems those individualsgunfavorable to generic drugs perceived more risk across the six dimensionsr They also attributed more importance to sustaining that loss than those favoring generic drugs. Since demographics have been weak discriminators of generic and non—generic grocery purchasers, perceived risk 36 seemed as a plausible explanation for the differences Bearden and Mason found. The level of perceived risk also has bearing on the degree of commitment to the product. That may partially ex- plain why some generic products sell better than others. Robertson (1976) introduced this under the premise that a product or brand may not be closely tied to a belief system. fTo further this he states that some consumption is trivial and non-ego involving, thus, less of a commitment; In rela- tion to risk, Cunningham (1967) correlated the relationship of perceived brand commitment to perceived risk. The three products evaluated were headache remedies, fabric softeners, and dry spaghetti. Based on the results,he concluded that "”perceived risk is positively related to perceived brand com- mitment” (Cunningham, 1967, p. 513). If, as Cunningham believes, risk is the nucleus of buying behavior, then perceived risk probably is involved with the introduction of a new brand in the market. Since generic products are relatively new, risk should be looked at concerning product trial. Cunningham used data from a test marketing project involving fabric softener. He tentatively hypothesized that people high in perceived risk would be slower in trying the fabric softener. Looking at it from a market share perspective, during the first few weeks of intro- duction the market share would be comprised of low risk 37 perceivers. \He also suggests that high risk perceivers tend to be more brand loyal; which means it is harder to get them to switch. When discussing perceived risk, Raymond Bauer (1967) the originator of the theory deserves attention. In his ini- tial article on "Consumer Behavior As Risk Taking” he referred to risk as being only a perceived effect. The rationale be- hind this is that an individual can only deal with risk as he subjectively perceives it. If the possibility of a real threat exists and the individual does not perceive it, then he cannot be influenced by it. Bauer's theory would also seem to imply that individuals perceive it differently as well as cope or reduce it on a personal level. One extension of Bauer's theory, that individuals per- ceive risk differently can be explained by the various com- ponents of risk others have identified (Cunningham, 1967; Milburn and Billings, 1976; Peter and Ryan, 1976; Peter and Tarpey, 1975). To begin with, Cunningham views perceived risk as the culminating effect of self-confidence. He has this divided into three categories; generalized self confidence, specific self confidence,and intermediate self confidence. The component of generalized self confidence was related to the way the individual feels about her capabilities in rela- tion to others. Cunningham tested this by asking respondents how confident they were of their abilities and if they were bothered by what other people thought of them. Low perceived 38 risk was hypothesized to infer high generalized self confi- dence. However, none of the tests proved significant. On the other hand, the specific self confidence com- ponent did show some significance. This question involved consumer confidence in choosing a fabric softener, headache remedy and dry spaghetti. The results indicated that high specific self confidence was associated with high perceived risk for fabric softener and low perceived risk for dry spaghetti. Cunningham interprets this finding as people high in self con- fidence are more able to recognize the inherent riskiness of a product that is considered high risk as well as recognizing low risk products. He concludes then that "perceived risk is a product specific phenomenon" (Cunningham, 1967, p. 108). He mentions that natural groupings for the type of risk perceived for products may exist. However, he is inclined to believe that the perception of risk is, as Bauer believes, an indivi- dual occurrance. Based on research conducted by Peter and Tarpey (1975) perceived risk may not be so product oriented as it is related to expected loss, return, or a combination of loss and return, by brand. These researchers identified six components of risk which had been used in another study and found conceptually independent of each other. The dimensions were: (a) financial, (b) performance, (c) psychological, (d) physical, (e) social, (f) time. These may sound familiar, as this scale was used in the generic drug study mentioned earlier. (1975). The three 39 models tested were the overall perceived risk in purchasing a brand j, the overall perceived return for purchasing a brand j, and the net perceived return for purchasing a brand. j. By using factor analysis on the six components, they found that the net return model explained more variance in relation to automobile brand preferences. Although risk is considered in a brand choice, so is the expected return. However, of these two dimensions, risk did explain more than perceived return. Related to Peter and Tarpey's interpretation of choice as a combination of loss and gain is Milburn and Billings‘ (1976) discussion of decision making on the basis of risk and subjective expected utility (SEU). The components of SEU are the probability of the alternative and utility for the consequence. The authors state, ceteris paribus, perceived risk increases when the probability of the negative conse- quence increases, and the positive probability decreases. Due to this classification system, risk can be judged by its degree, as utilities and probabilities change. Milburn and Billings also posit that individuals consider different dimensions in assessing risk. Based on past studies,they (concluded that risk is not a general trait, but varies with the task and situatiofig This finding acts as a justification for Peter and Tarpey's multi-dimensional scaling of overall perceived risk. 40 Another study which has become a classic in perceived risk was conducted by Roselius (1973) on risk reduction methods. His rationale for this study was that retailers often face a trade-off of the cost incurred in offering a risk relieving outlet and the benefit of higher sales volumes afforded from its successful implementation. Rather than the six risk components mentioned earlier, Roselius identified four losses on which to evaluate the risk relievers, all of which were included in the previous study by Peter and Tarpey. They were: (a) time loss, (b) hazard loss, (c) ego loss, and (d) money loss. The relievers selected were: (a) endorsements, (b) brand loyalty, (c) major brand image, (d) private testing, (e) store image, (f) free sample, (g) money back guarantee, (h) government testing, (1) shopping around, (j) expensive model, (k) word of mouth. The findings have interesting implications on what has already be done. For example, for all types of loss mentioned, (buying the most expensive model was consistently the least favorite method of risk reduction. That questions previous studies which postulate that higher price means higher quality and is less risky (Shapiro, 1968) Confirming other studies, Roselius found that brand loyalty was consistently sighted as the most favorable method of relieving risk, while purchasing a major brand was second. The other methods were given some- what more neutral rankings. Due to the mixture of rank ordering within each loss, the study showed that different relievers are preferred depending on the loss involved (Taylor, 1974, p. 57). 41 Other risk reduction methods have been studied in addition to Roselius'. The most common of these is informa- tion seeking either through formal or informal channels. This will be the topic of the next section. 42 Information Handling Beside the eleven risk reduction strategies Roselius postulated, information can be one of the most effective relievers available. Information sources have been divided into informal or personal sources (Cunningham, 1967; Cox, l967).and formal or market dominated information such as advertising (Cox and Rich, 1964; Houston, 1979; Thorelli, 1971). Past research has demonstrated that different channels seem to be more important in providing certain information than others. Therefore,sources of information effect specific risk components individually. One of the earliest studies on risk reduction methods involved telephone shopping. Although it is a convenient mode of shopping, most women do not pursue over the phone shopping. Cox and Rich hypothesized that the degree of un- certainty of this form of shopping acts as a barrier which prevents more frequent use of telephone orders. They saw the shopper had only two means to relieve or lessen the uncertainty: (a) relying on her past experience with the store, product or brand,and (b) relying on a newspaper ad which may picture the article. The economic cost of a bad decision is at stake. Also,other elements of risk involved are: time loss from having to return the item, and ego loss from the frustration of making a poor decision. 43 The results displayed differences in the way phone and non—phone shoppers view information. Phone shoppers, 50% of the sample, ordered items which had been advertised in order to reduce uncertainty. The results also pointed to non-phone shoppers as being "unwilling or unable to use news- paper advertising as a reliable" source of information for re- ducing some of the risk. Those women who did find the advertis- ing very helpful, were five times more likely to shop by phone as women who did not find newspaper ads informative sources for the products. In conclusion, it seems that referring to newspaper advertising is the most frequently used method of re- ducing risk for telephone shopping, used by 58% of the sample. Reliance on past experience was cited less frequently, by only 18% of the sample. With phone shopping, there is less prepurchase contact with the product. Therefore, it is considered a "highly risky venture” (Cox and Rich, 1964, p. 33). Similarly, a new brand on the market carries with it an inherent risk, since personal experience with it is low. Cunningham (1967) looked at informal word of mouth communication regarding old and new brands. In the case of headache remedies and fabric soft- eners, both high and low risk perceivers did discuss old and new brands. However, those individuals in the high to medium perceived risk group discussed new brands approximately 40% more frequently than old brands. 44 In addition to informal sources of information,Sheth and Venkatesan (1968) studied prepurchase deliberation and brand image or loyalty, as risk reduction methods. Their basic hypothesis was that these three variables change over time as the consumer learns more and creates decision rules or heuristics for repeat purchases in the same situation. Using hair spray as the product with college women divided into high-and low—risk groups as the sample, they found, prepurchase deliberation decreased with both high and low risk groups. Information seeking also decreased over time. However, early in the study both groups sought information, with the high—risk group turning to personal sources more fre- quently. Brand loyalty increased over time, but there were no specific differences between risk groups. The data from the study on prepurchase information seek- ing, as Sheth et al. reported was derived from weekly question- naires filled out before the next hair spray selection was made. Only personal sources significantly differentiated between high- and low-risks groups. According to more current research, Sheth et a1. may have understated the actual importance of non-personal sources by relying on the survey method to collect this data. Newman and Lockeman (1975) believe that recall abilities, even on the day of the purchase can be less complete than actual observation. They advocate more retail store observation for better measurement of whether 45 consumers are more active searchers of information than results have suggested. Other research directed at information seeking behavior postulates that consumers can handle only so much information when making a choice. Once the capacity is full, the consumer becomes confused and choices are less optimal (Jacoby, Speller, and Kohn, 1976, p.'63x Jacoby et a1. (1974) studied the relationship between amount of information and number of al- ternatives for decision making. Providing more information to reduce risk showed to be favorable by the sample. However, there was no statistical significance between amount of infor- mation desired and correct decision making. Furthermore, this desire for information may be different for the various sociO*economic or age groups in society. Schiffman (1971) studied the sources of information for the elderly. For this group, past experience substituted for exter- nal sources of information. Thorelli (1971) examined educa- tion and income as variables which determine information use activities. His Norwegian sample elicited data that showed consumers with higher income, and more education (and house- holds subscribing to a consumer information magazine) turned to information sources more often and thoroughly than less edu- cated,low income (non-subscribing) households. In explanation of this phenomenon, a quote from Donald F. Cox seems most appropriate, 46 The nature and amount of risk will define consumer information needs and consumers will...turn to the information sources whose perceived information characteristics seem most likely to satisfy their particular information needs. (Cox, 1967, p. 613) When discussing information sources, advertising is indeed one of the most visible and easily obtainable sources. Past research indicates that the use of advertisements varies among products as well as consumers (Engel et al., 1968, p. 398). Basically,ads communicate product attributes to pros- pective buyers. Galbraith labels this a "simple communica- tive purpose" (Galbraith, 1967, p. 16). However, the role of advertising is a much more complex issue. As studied by Houston in 1979, different sources provide various degrees of useful information in the purchase of certain products. In his text, advertising, be it televi- sion or magazine, was not always rated highest for providing information on a particular attribute. Advertising was con- sulted most frequently for style information. For price, durability, extra features and dealer reputation, personal conversation or dealer visits was cited most often. Thus, Houston concluded when consumers need information on an attribute which is specific to a product, for example styling information, a prejudiced source such as advertising does not impair its usefulness. When objectivity is important, for example when judging durability, advertising is not consulted as highly as other individuals or rating magazines. However, 47 Houston implies he does not want to belittle the role of advertising by his findings. Bucklin (1965) would agree with Houston because he stated that advertising is merely one source which consumers use to improve their purchasing ability. It'S use depends on past experience and the need for better product information as Cox's quote implied. In conclusion, advertising's role can be thought of informing consumers of product benefits of which they may not have been aware (Bucklin, 1965, p. 52)- HYPOTHESES Due to the fact there has not been a great deal of research on generic consumers the major goal of this study was to provide some exploration into the possible character- istics related to generic shopping. However some relation- ships were speculated upon based on the findings of past studies. Those results which had implications for the study of generic shopping have been discussed in the literature re- view. This chapter provides the rationale for the inferences drawn to describe the generic phenomenon. 48 49 Differences Between Generic And Non-Generic Shoppers Returning to the original studies conducted on generics, variables surface that can be used as potential delineators between generic and non-generic shoppers. Both Progressive Grocer studies surveyed demographic variables as did MDrphy et al. For these studies, using the percent represented in each group or a chi square, few demographics were found to be significant. In addition, in the articles on private-brand-prone consumers, Frank (1965) found that the number of persons in the family, and educational level were related to private brand proneness. Also,income and occupation and their relationship to generic buying were studied. Goldman (1976) who researched the knowledge of low-income shoppers found them less aware of the variety of alternatives available to them. Lastly, in light of the demographics undertaken in this research, Martineau (1958) who examined social class believed that occupation was one factor that influenced a person's economic behavior. Eco- nomic behavior may mean income level or just the values placed on certain activities, since not all lower income individuals save money in buying the less expensive brand. In fact, Kunreuther (1973) suggested they pay more due to budget constraints, therefore, they have to buy smaller sizes more frequently. Thus, the use of a budget in doing grocery shopping may influence whether a consumer shops generically. Generics often are packaged in large sizes, yet do offer a savings. 50 Budgeting may not matter then in this case. Regardless of the package size, this variable was included. With this informa- tion in mind, the following variables were examined in order to see if indeed there was a difference between generic and non-generic shoppers: (a) occupation, (b) number of peOple in houshold, (c) ages of children at home, (d) edu- cational level, (e) yearly total family income, and (f) use of a food budget. As mentioned before, the main purpose of introducing generic products was to offer consumers a third way to save money. Thus,consumers' perceptions of the price-quality dich- otomy of generics was explored. Past research in this area has indicated in the absence of other marketing cues, price tends to connote the quality level (Leavitt, 1954) (McConnell, 1965). Others have found that when combined with a variety of marketing cues, brand name and brand image for example, the use of price for judging quality is cited less (Jacoby et al., 1971), One factor that can be assumed is that generics are gen- erally free of promotional inputs. There is no distinctive label which a brand loyal consumer can rely on to imply quality. There also is no manufacturer advertising to provide generics with an added value. They are only promoted by the retail outlet, usually in newspapers, praising their low cost. There- fore, without the traditional marketing cues, price may be the only signal generics have to connote quality. 51 Progressive Grocer was interested in a similar type question and so in their 1978 study they asked why consumers thought generics sell for less. They.assumed all individuals in their sample were aware of this benefit. However, since one of the main objectives of the thesis research was to try to distinguish generic from non-generic shoppers, this topic was explored by seeing if the differences between generic and national brands was perceived differently between the two groups, assuming they were not aware of price. Thus, the Progressive Grocer response categories were used with price being added to the list. Along the same line of reasoning, generic and non-generic consumers may perceive the benefits of generics differently. Generic shoppers may see the price as indicative of a lower quality but still acceptable (Coyle, 1978). Whereas, non- generic shoppers may see the only benefit as lower cost which may imply lower quality. Thus, the differences in the benefits that generic and non-generic shoppers perceive was also ex- plored. As Norris (1960) implies there is an added value that adver- tising imparts to goods. Some consumers may choose national brands because of this. In order to find out if generic and non- generic shoppers placed different values on the national brands in their shopping cart, the qualities that shoppers bestow on their national brands was researched. 52 Although partially related to perceived risk, the search for information was included because generic and non-generic shoppers may look for information differently. As Cox states in his article, information sources are consulted as consumers identify the need to use them (Cox, 1967, p. 613). Therefore, the need to check a newspaper may vary as well as the number of days the shopper reads it to find out about ways to save money. The types of information sources consulted by individuals varies. Cunningham (1967) reported personal sources tend to be relied upon more by high-medium risk perceivers. For tele- phone shopping,Cox and Rich (1964) found that newspaper adver- tising was consulted most for ordering products over personal experience. Houston (1978) concluded that for information about certain product attributes different sources were con— sulted. Therefore, not all consumers seek their infor- mation from the same sources. Again, this area may show that generic and non-generic shoppers rely on different sources for product information. Based on the literature review and focus group results,the following hypotheses regarding the differences between generic and non-generic shoppers were developed to test in the study. Major Hypothesis 1: Generic shoppers are significantly more likely to: (a) Shop on a food budget (b) Have a lump sum or percent of paycheck as their budgeting method (C) (d) (e) (f) (g) (h) (i) (j) (k) (l) 53 Look for taste and consistency in the national brands they buy Perceive the primary benefit of generics to be their low cost and acceptable quality Perceive the major differences between generics and national brands as less cost and no promotion Actively search for news of sales and coupons Use of shopping experience as a source of informa- tion Have a middle to upper level occupation Have a higher education than non-generic shoppers Have an income under $15,000 or over $30,000 Have larger households than non-generic shoppers Have younger children at home than non-generic shopers 54. Perceived Risk And Generic Shopping Progressive Grocer's two city research study included details about which products sold the most frequently. There appears to be a definite distinction among the different product categories. Table 3 shows this breakdown. By the per- centages displayed, it seems that shoppers prefer some generic products over others. The Progressive Grocer's explanation as to why the paper and plastic products are doing best is because they are viewed as safe (Coyle, 1978, p. 80). In the theoretical sense, less perceived risk may be associated with an inedible such as paper napkins than peanut butter. As a result of this evidence on the movement of products,_ the theory of perceived risk was introduced to help explain gen— eric shopping behavior. Bettman (1975) devised two constructs to explain risk more precisely. One is called inherent risk. This is attached to the latent risk a product class holds for a consumer. Handled risk, the second construct occurs after purchase. This paper is concerned with inherent risk. Inherent risk also varies with the level of salience associated with the product class. Betteman experimented with a linear and a multiplicative model for comparing inherent risk across product categories. With both models, he offered an explanation for one set of results that perceived price may be related to perceived quality, thus a higher price leads to lower risk. 55 TABLE 3 How Generic Products Movea In Boston In Chicago Product % Buying % Buying Paper or plastic products 83% 59% Canned vegetables 45 52 Canned fruit 44 47 Canned juices 44 33 Soap/laundry products 37 24 Peanut butter or preserves 27 18 Canned fruit drinks 20 27 Canned tomato products 20 24 Mayonnaise 20 15 Canned tuna l9 6 Canned soups ll 6 Pet food 8 6 Canned softdrinks 3 6 aBased on buyer of at least one generic item. Source: Coyle, J.S. Why Jewel did, how consumers respond, what the risks are, where it all goes from here. Progressive Grocer, February 1978. The variation may be due to the fact that a consumer's price sensitivity varies again by product class (Gabor et al., 1964), The sensitivity may be due to a heightened price aware- ness, such as with those who watch their budgets. There also may be a range of prices that a consumer has before a slight decrease in price would encourage him to try a new product. As Peter and Tarpey (1975) identified it, personal risk can be divided into six categories. Each category may have a varying relevance to generics. The category of finan- cial risk would include the price versus performance or quality 56 expected from the generic. The physical risk would include the health hazard perceived by eating the edible generics. The performance risk would apply to the item tasting good, cooking properly, or doing the task expected. Since generics are shelved together in a single section, convenience risk would be the extra time expended going to the generic section. The social risk would be what friends and relatives thought of buying the lower priced goods while psychological risk involves inconsistency with the consumer's perceived lifestyle. On these six categories a risk index was calculated and used to judge relationships between high and low risk per- ceivers. This classification system was looked at to see if high and low risk perceivers buy more national or generic brands in the product categories suggested by the Progressive Grocer study. Perceived risk was also correlated with first generic brands tried and generic brands shoppers said they would never consider trying. It is assumed that since generics do not offer any risk relievers Roselius supported, that generics would attract low risk perceivers. In summary, the basic hypothesis tested for perceived risk was: Major Hypothesis 11: Low perceived risk individuals are significantly more likely to: (a) Be generic shoppers (b) Have tried more generics when they first started buying generic grocery products 57 (c) Have considered buying more generic items than those of high perceived risk (d) Actively search for news of sales or coupons than high risk perceivers METHODOLOGY Sampling The sampling technique used for the study was a non- probability method called convenience sampling. It involved the personal judgement of the interviewers for subject selec- tion in the sample. When subjects are not selected proba- bilistically, an estimate of the degree of sampling error cannot be determined. However, due to the location of the study, East Lansing, Michigan, home of Michigan State University, special care was taken to escape oversampling university students, something random sampling may not have been able to avoid. Another probability sampling method considered before the convenience sample was decided upon was systematic sampling. First, systematic sampling using the East Lansing Telephone Directory for either a mail questionnaire or a phone interview presented problems. Since the respondents of concern were non- college students, with over 42,000 students listed in the dir- ectory, the sample may have been composed entirely of students. Although names could have been cross checked in the student directory, not all students are included in the book. It is 58 59 voluntary information. So,in the case of systematic random sampling the sample may not have been representative of the target population. Secondly, a goal of the research was to look at an approximately equal number of generic and non- generic shoppers. An excess of telephone calls would have to have been made using a screening question to get ample representation. The cost of sending mail questionnaires to enough respondents to get about the same number in each cate- gory would have been prohibitive for this researcher. With other probability techniques such as cluster, the same pro- blem exists of contacting too many students. Although con- venience samples are not recommended for causal research, Churchill states that they may be employed with exploratory designs, which is the purpose of this research (Churchill, 1976, p. 264). Once the sampling technique was planned, the next step was to find a local grocery store that would allow a univer- sity research team to administer interviews to shoppers. That turned out to be a problem because many store chains have policies which do not allow solicitors or researchers ex- ternal to the company in the store. Fortunately,a grocery store which would allow academic research was discovered which fit the major requirement of being an adequate distance from the MSU campus for sampling purposes. The research was conducted at the Eberhard Food Store on Michigan Avenue next 60 to the Frandor Shopping Center in East Lansing, Michigan. The information was collected May 23 to May 28 between 10:00 a.m. to 5:30 p.m. Two hundred and twelve (212) interviews were completed. Of the total, 200 were acceptable for the sample. Due to the restrictions placed on us by the store manager, only four interviewers were allowed in the supermarket together. Two interviewers were stationed in the generic section and the two other interviewers stood slightly beyond to catch those who did not place generic items in their shopping cart. The interview staff consisted of the thesis researcher and eighteen undergraduate students at Michigan State University in the advertising research methodology course. As Babbie (1979) states in his text on social research, ”the interviewer should be a neutral medium through which questions and answers are transmitted" (p. 338). Therefore, an interview training session was held in order to acquaint the students with the questionnaire and interviewing style. The original bias in a convenience sample is that only those individuals who happen to be at that place at that time are sampled. In addition, the interviewers for this study were instructed not to question students for data which intro- duces another bias. However, one bias not intended by the researcher was the oversampling of elderly and retired indi- viduals. Close to the Eberhard store is a retirement build- ing. Thus, a large number of people over 65, both men and women shop at Eberhard's. 61 Data Collection The data was collected by a questionnaire via personal interview. The use of personal interviewers allows for sampling control, the key to the convenience technique. The sampling control permits the researcher to direct the inquiry to a designated individual and helps to secure cooperation from that respondent (Churchill, 1976, p. 177), The personal interview provides the interviewer with the most sample con- trol over mail and telephone survey methods. He knows who he is speaking to, meaning there is no mystery as to who filled out the mail questionnaire. Also, refusal to participate is lower. Because the respondent has to put forth minimal effort by just answering questions verbally and is in a face to face situation, he/she is less inclined to refuse to answer. A personal interview also affords more information con- trol. The personal interaction between individuals allows the interviewer to graphically show pictures, scales, or other stimuli. Furthermore, interview surveys are more effective in dealing with complicated issues that may need additional explanation. Lastly, there is the consideration of administrative control. These elements include speed, control of the reply situation, and cost. In relation to speed of gathering the data, personal interviews require the least time to complete. The data for this study was collected in three days. 62 Unless the same individual conducts all the interviews for the research there remains the problem of interviewer bias. Moreover, the problems of controlling the reply situa- tion increase with the number of interviewers doing the research. Field errors, or observation errors result from interviewer bias. It causes inaccuracies in the data. This type of data collection error assumes that there is a true value for a variable. Thus, the observational error is the difference between the true value and the value reported by the interviewer. Often it is undetected or undetectable. According to Churchill, (1976) the three response biases to which interviews are subject are as follows: 1) Errors in asking questions and in probing 2) Errors in recording the answers 3) Errors due to cheating In order to reduce the problems of interviewer bias, first an interviewer training session was held and secondly, all the student interviewers administered their questionnaires under the researcher's supervision. As most research text's state, personal interviews are reportedly the most costly. However, since this study was conducted at the university level, the supply of interviewers was sufficient with only the expense of granting extra credit points for the research course from which the students volun- teered. 63 Questionnaire Design Questionnaire Construction Stage 1 The questionnaire for the data collection was developed in three stages. The first stage involved deciding on what information to acquire for the study. The answer to this question came from the literature review and the objectives of the research. In order to proceed to the actual construc- tion of the questionnaire, a focus group was conducted to elicit content and response information. The group consisted of eight middle class housewives from the Ingham county area. Five major categories were probed: 1) Shopping habits 2) Information sources 3) Consumer brand perceptions 4) Perceived risk 5) Demographics Focus Group Results Stage 2 To begin with, under the heading of shopping habits questions concerning grocery budgeting, routinized versus extensive shopping, and store patronage were asked. The re- spondents saved money by purchasing store brands, or generic brands and clipping coupons. They also reported that many of them now do a lot more comparative price shopping within one store. With the gas shortage,several women now stick with 64 one grocery store. By becoming familiar with that store it reduces their shopping time. One woman said in regards to her supermarket, "I know the store inside and out and that's why it's very easy to shop." It seems to encourage brand loyalty too. Again when pressed for time, the women know where to go in the store for the product, and as one member of the group remarked, ”you just look at the can and say, oh, that looks familiar and throw it in the cart." From this topic the discussion moved on to sources of information. Two—thirds reported they use newspaper adver— tising as a source of product information, while the majority cited shopping around or experience with different products and brands as their primary sources of information. Those who said they read labels for nutrition and content information were the mothers of children with special allergies. When asked specifically about their perceptions of generic products many had definite ideas concerning their differences. Poor quality was cited most often in connection with the paper products. Others felt the prepared foods, such as ketchup, or jellies, were not Very good. The basic differences between generic and national brands were perceived as taste, texture, cooking time, labels, and consistency. Generics were perceived as having less attractive and less informational labels. National brands were perceived to be of a higher quality and part of their higher price was due to the label. 65 The main reason they tried generics when they first appeared on the market was to save money. However, none of the group members continued to buy generics because they found the quality unsatisfactory. In the area of risk, many women expressed a hesitancy towards purchasing certain branded products. For example, meats, fish, and specifically tuna were mentioned. One woman would only buy her one brand of tuna regardless of what was on sale. In the area of detergent and paper goods most found the store brands satisfactory. However, with peanut butter, and softdrinks, and pet food the national brands were pre- ferred. The session ended by asking a few demographic questions. For a complete review of the questions discussed by the focus group see Appendix B. Appendix F is the manuscript. Questionnaire Construction Stage 3 Based on the information from the focus group, the first draft of the questionnaire was written. The form 0f most of the questions was open ended. The response categories were listed under the question for the interviewer to check off. After six revisions following reviews by three adver- tising professors, the questionnaire used for the study was prepared. It is found in Appendix E. 66 Operationalization of Variables As discussed earlier, four major topics were covered in this study. From the broad categories of shopping behavior, consumer perceptions of the qualities of branded goods, per- ceived risk and basic demographics, specific measurable var- iables were extracted. The two discriminatory variables were the levels of per- ceived risk and the state of being a generic or non-generic shopper. The questionnaire included two possible ways to be considered a generic shopper. Question three was a list of fourteen products which the interviewer read one at a time. To each product the subject responded by indicating whether they bought mostly the national brand, the store brand or the generic. If they never purchased the product, a "don't buy” category was checked. If the individual indicated he/she purchased any generics on the list of products, they were given generic shopper status. Due to the fact that the product list was not exhaustive of the total generic offerings, the in- terviewer also looked into the respondent's cart. If they had placed other generic products in their cart, the interviewer checked the response ”yes” to question #22. The first com- ponent of the screening question was as follows: As I read each of the following names, please tell me whether you buy mostly the national brand, store brand, generic brand, a combination or do not buy the product at all. 67 The list included the following products: toilet paper peanut butter napkins mayonnaise paper towels spaghetti canned vegetables canned tuna canned fruit canned soups soap food for your pet laundry products cake mixes Question 22, which was directed to the interviewer, asked, ”Does the shopper have a generic product in her/his cart?" It was a behavioral measure to supplement question three. Since the product list in the screening question was rather small, the cart check was very valuable in identifying a generic shopper. The second explanatory variable was perceived risk, first discussed by Bauer in 1960. As he believed, consumers act to minimize or reduce the amount of expected negative utility associated with a purchase. In 1965,Peter and Tarpey studied three multiplicative risk models. In the models, they identi- fied six separate dimensions; financial risk, performance risk, psychological risk, physical risk, social risk, and time or convenience risk. Peter and Tarpey's scale quantified the probability of each risk occurring when a person purchases a particular brand (1975). It asked the respondent to rate the probability of each risk occurring for a particular product, and the importance of sustaining a loss as a result of purchasing the product. Since the perceived risk scale in this study was administered orally, it had to be simplified. Therefore, the scale was converted to a summated instrument which measured 68 the degree of risk an individual feels when purchasing generic products. The scale generated ordinal-level data and measured the degree of risk for each of the six components. A sum was calculated for the total items for each individual. The SPSS program "Compute” command was used to divide the respondents into high and low perceived risk groups. A frequency list of all scores was segmented and a median was also calculated. Then, SPSS ”Recode” was run which put those individuals from the median (10.05) and up into the high risk group and those below the median in the low risk group. The following are the operational definitions for calculating the risk. Financial Risk Because of such things 4 great financial risk as Its poor quality, if 3 moderate financial risk you bought a generic pro- 2 slight financial risk duct, would you consider 1 no financial risk at all 1ta Social Risk Because your friends and 4 great social risk relatives would think less 3 moderate social risk highly of you if you 2 slight social risk ‘bought a generic product, 1 no social risk at all would you consider it a Physical Risk Because it would not be 4 great physical risk very safe, if you bought 3 moderate physical risk a generic product would 2 slight physical risk you consider it a 1 no physical risk at all Convenience Risk Because it would take 4 great convenience risk extra time to buy the 3 moderate convenience risk prgduct, 1f ygu boughtl:d 2 slight convenience risk eneric pro uct wou 1 no co ' ‘ you consider it a nvenience risk at all Performance Risk Because it would not cook 4 ‘ . great performance r1sk properly, if you bought 3 moderate performance risk a generic product would 2 slight performance risk you con31der it a 1 no performance risk at all 69 Psychological Risk great psychological risk moderate psychological risk slight psychological risk no psychological risk Because it would not be consistent with your life style if you bought a generic product would you consider it a 14wa Since the major benefit of generics was originally to help shoppers save money, it was hypothesized that generic shoppers are more price conscious. Therefore the following question, ”Do you have a food budget?" aimed at identifying those consumers. Question two acted as the validator of a "yes" response to number one. It asked the respondent to describe how their budget works. Therefore, those who re- sponded, ”I don't know,” or ”I guess I really don't have one" after they thought about it were not classified as budgeters. In order to find out which generic items the generic person tried originally when they first began using generics, the following question was asked: When you first started buying generic items, which were the first products that you bought? The space provided under the question allowed the interviewer to list all possible responses. Current generic shoppers, as well as past ones may have tried generic products and discontinued use for one reason or another. In order to find out if there are differences between current generic users versus nonusers on the basis of satisfaction,the following question was asked: Are there some generics that you have bought in the past and will not buy again? The question was followed by "yes” or ”no" responses. 70 In this study, it was important to look at risks's influence on purchasing. As seen by industry reports, certain generic products sell better than others. In order to find out if there were certain products which consumers felt were too risky to try in the generic section, the follow- ing question was asked: Are there some generic products you would never consider buying? This question was followed by asking ”which ones?" The inter- viewer was instructed to list all items. In order to find out what the qualities are that shoppers look for in the nationally branded goods they buy, or why they purchase them, the following open ended question was asked: What qualities do you look for in the national brands. that you buy? There were five (5) response categories following the question plus an ”other” category. The response categories were based upon the focus group's responses or previous studies that ex- amined similar variables. Those included on the questionnaire were: consistency higher quality familiarity reduces shopping time advertising brand recognition Again, with the intention of discovering what consumers regard as benefits of the generic products the following open ended question was asked: Could you please tell me the primary benefits of generic products? 71 Some response categories were also provided. They ori- ginated with the focus group. An ”other” opportunity was also present to cover the range of answers not covered by the focus group. The interviewer checked as many as applied. The responses included in the question were: Low cost Same quality as national brands less quality but acceptable for my purposes I don't need advertising to sell me a good. In order to find out what consumers felt were the basic differences between generic and national brands, the follow- ing question was asked: Now, what makes generic items different from national brands? The response categories came from a previously administered questionnaire by Progressive Grocer (Coyle, 1978). They included: lower price low production/distribution costs lower quality no difference no advertising don't know cheaper packaging other One aspect that surfaced in the focus group discussion was that consumers save money by shopping with coupons. Often they are clipped from the newspaper. Thus, some conscientious shoppers may make special efforts to look for savings informa- tion. This information was obtained by asking: Do you make a special effort to look for news of sales or coupons? This question was extended by looking at how much effort is put into the search. The following was asked in order to 72 find this out: Can you tell me about how often you look for this information? Again, following this open ended question there were re- sponse categories to aid the interviewer in faster recording. As generated by the focus groups, the categories were: Everyday Twice or three times a week Food day in the newspaper The day I do my shopping Weekends Other The variables in the demographic section were treated as interval-level data except for income which was categorical. The first one was occupation of head of the household, operation- alized by asking: What is the occupation of the head of the household? The interviewer recorded the occupation. For coding purposes an interval level socio—economic index was consulted from Occupations and Social Status by Albert J. Reiss Jr.(l96l). Reiss ranked each occupation based on occupational status from 0 to 100. The scale cannot be accepted as the "exact representa- tion of the status of individuals in each occupation or as an exact representation of the stratification system of society” (Hall, 1969, p. 274). However, the advantage of using this scale is that it allows a measurable scaling of occupations. The scale has been attributed to O.D. Duncan. He constructed the scale by combining measures of education and income. Al- though he noted, the two measures taken individually indicate 73 social status, the combination of factors produces a "multi- dimensional scaling of occupations" (Hall, 1969, p. 274). Thus the socio-economic index accounts for both the social and economic positions in the stratified society. Appendix D has the listing and rank of all the occupations of respondents in the survey. To find out how many people lived at one household, the question was asked: How many people including your spouse and children live at your home address? The actual number was recorded. The ages of the children were sought by asking: If you have children, what are their ages? Again, the raw numbers of the ages were recorded for all children. In order to lead into the sensitive question of level of education, first, the respondent was asked: What was the name of the last school you attended? The answer was recorded but there was no computer tabulation of the responses. This question was then followed by: What was the last grade you completed in this school? The actual years of attendance were recorded. A comparison based on income was another objective of this study. Since today many families are two income families, the following question was asked: 74 Please look at this sheet and tell me if your yearly total family income falls in category 1, 2, 3. On the sheet the respondent held, the income categories were coded: l = below $15,000 2 = $15,000-$30,000 3 = above $30,000 After the interviewer thanked the respondent for participating, she/he recorded the respondent's sex. 75 Validity and Reliability The following section on reliability and validity of the research will be examined based on Morris Holbrook's (1975) review article on advertising research. The adequacy of the test will be discussed in the areas presented in Holbrook's typology, in Table 4. TABLE 4 A Typology Of The Criteria For Evaluating Research Adequacy Of The Test Adequacy Of The Theory Validity Reliability Syntactics Logical consistency Transferability Semantics Measurement validity Measurement reli- ability Pragmatics Internal validity Statistical Signi- description ficance Prescription Strength of Associa- External validity tion Logical Consistency Holbrook defines logical consistency as the degree of fitness between theoretical sources and its derived empirical test of theory. In his article, Holbrook's major concern is with the fallacy of assuming all linear relationships. He objects to ignoring a hierarchy of effects situation, or feedback situations. 76 This study would probably come under scrutiny by Holbrook due to the fact it does not have its own theore- tical background from which to draw. Therefore,the generic topic was covered by inferring variables to test from other areas of marketing research. Transferabiliry Transferability is the application of concepts from other related fields to the hypothesis under study. This becomes a reliability problem when the borrowed theories are not consis- tent with the effect being studied. In doing this research on generics, the variables studied are intimately related to mar- keting. Risk and information handling were exhaustively studied together by Cox et a1. (1967). Shopping habits as a basis for segmentation has also been studied extensively (Frank et a1. 1967; Farley, 1964). Therefore, the concepts posited here have long been involved with researching brand behavior. Measurement Validity The two components of this criterion are definitional val- idity, the extent to which the operational definition actually elicits the true meaning of the variable under study; and instru- ment validity, the amount of independence among the variables between tests. Questionnaire data in relation to definitional validity, may side step the exact issue. This problem was over- come by relying on previously administered questions used for 77 practical or academic research. Since the results from the generic questionnaire in this study had similar findings as the ones from which many questions were borrowed, a retest type validity was established. The instrument validity problem in survey research, accord- ing to Holbrook, is in respondent bias. Three checks were made on the issues usually considered sensitive on questionnaires to avoid this problem. First, when interested in the level of education of the respondent, a buffing question of the name of the last school attended was asked. This question was then followed by the question about the highest grade completed in this school. Secondly, the issue of income was handled by hand- ing the respondent a sheet with income categories from under $15,000, between $15,000 and $30,000 and over $30,000 coded by 1, 2, 3. The respondent merely indicated into which category, either 1, 2, or 3, their yearly total family income fell. The third area that might encourage ”yeasaying" as Holbrook calls it, was the question concerning budgeting. The question was ”do you have a food budget?" Consumers may feel it is socially desirable in these inflationary times to answer yes. In order to make sure that indeed they had a legitimate budget, the next question asked them to describe it. The ”I don't knows” or "unsures" were then considered non-budgeting individuals. Measurement Reliability The importance of measurement reliability is a special case when using an attitudinal scale. In this study it was 78 necessary to check to see if the perceived risk scale actually tapped the construct of interest. There are a variety of statistical methods available to estimate the reliability of the scale, to check for consistent results. Since the questionnaire was administered only once, the most appropriate measure of reliability was the Guttman split-half. The split— half test is applied to unidimensional scales. This means the items are summed into a single attitude scale. Thus, the total score is a reflection of perceived risk. In order to compute the reliability score,the total items were divided into two equal halves with alternating items in each half. Items one, three and five were in the first group, with two, four, and six in the second. The test then corre- lates items with each other. According to Nunnally, a reliability score of .70 or higher is satisfactory (Nunnally, 1978, p. 245), By using the SPSS program "Reliability.n a Guttman split-half measure of .78827 was calculated. Thus, the perceived risk scale is a significantly reliable measure. Internal Validity Holbrook interprets internal validity as "the extent that an empirically established relationship permits a causal inter- pretation” (Holbrook, 1975, p. B-38). One reason he sees it as a problem, is that the groups studied are systematically different to start with. He recommends contingency tables as one way to control statistically for extraneous factors which 79 this research employed. By using a chi square test of sig- nificance, the variable relationships which occur by chance can be detected. Statistical Significance As previously discussed, statistical measurements allow the researcher to make judgements about the probability of the relationship occurring due to chance alone. Since there are random fluctuations, measurement irregularities, and variance in every sample, a test for statistical signifi- cance predicts the reliability of replicating the results in reproducing the same test. For the nominal- and interval-level data in this study, a .05 significance level was set. All data that did not exceed the 95% confidence interval was rejected on the basis of occurring by chance variation. Strength of Association Holbrook's philosophy concerning significance is that given a sufficiently large sample, any relationship will end up being statistically significant at a given probability level. Therefore,before any test was run an alpha level of .05 or less was established apriori as the acceptable level for the chi square and T-test analysis. In deciding on the significance level, the problem of committing Type I and Type II errors was realized. Since this is basically exploratory research,moderate to strong esti- mates of possible relationships were sought. Blalock (1979) 80 cites that .001 to .05 are considered conservative levels (Blalock, 1979, p. 161). However .05 was established as the maximum level at which to reject the null hypothesis. This would minimize the Type II errors (failure to reject assump— tions when they are false) if the more conservative estimate is used. A conservative alpha level increases the risk of a Type I error (rejecting a set of assumptions when they are true). To maintain the quality of the findings, this re- searcher chose to be more conservative. External Validity External validity deals with the extent of generaliza- bility of the results to real life occurances. One advantage to survey research, as conducted here, is the fact that there is no artificial manipulation of the variables. Also,the sur- vey was conducted in a grocery store where the respondent had shopping on the mind and was relaxed. Furthermore external validity is judged by the repre- sentativeness of the sample. The members of this sample were a broad mixture of individuals, from truck drivers to college professors. However, due to the location of the supermarket where the questionnaires were administered, students and re- tired adults were highly represented. Also, a non—probability sample was drawn. Thus, this may slightly affect generalize? bility of the results to the population at large. RESULTS Treatment of Data The main objective of this study was to compare generic and non-generic shoppers and the relationship of risk per— ceptions to generic shopping. Therefore, the statistical methods employed to compare the groups were chi-square tests on the nominal data and t-tests on the interval data. Appendix C is the code sheet for the questionnaire. It contains all the nominally coded categories. Question fourteen was the six-part perceived risk variable. In order to convert this ordinally constructed scale into nominal categories of high and low risk perceivers for chi-square ana- lysis two steps were taken. First an SPSS frequency run was done to find out the aggregate median value. The scale was computed by summing the six items for each individual, and arriving at a median value for the total sample. Next, all re— spondents in the sample from the lowest value (5) to the me- dian value (10.05) were categorized as lowerisk perceivers, while those respondents from the median value to the highest scale value (24) were categorized as high-risk perceivers. This was formulated by using SPSS programs "compute" and "recode." There was another item on the questionnaire that also had to be summed for t—test analysis. Item seventeen asked 81 82 for the ages of the children living at home. First, all two hundred questionnaires were scanned. The highest number of children at home was five. Therefore,10 columns on the computer card were allowed. Since it was important to preserve all data,each child counted. However, mean scores for the groups generic and non-generic shoppers needed to be calculated for the t-test. Therefore,SPSS "compute" was run to sum the ages of all the children per group. Means were caluclated from the new computed variable. One additional technique used for data analysis found in the supplement guide to SPSS is the "multrresponse" pro- gram. This program helped analyze the data in questions 3a and 7. Question 3a was a list of the generic products that the respondent had first tried. Question 7 was a list of generic products that the respondent would never consider try- ing. For both questions in all two hundred questionnaires the maximum amount listed was six. So six columns per question were allotted. Due to the fact twenty-six product categories were found, the responses were coded with the alphanumeric characters of the alphabet. The coded categories can also be found in Appendix C. The six columns per question needed to be summed to perform a significance test based on high and low perceived risk. However, alphanumeric characters cannot be aggregated by a ”compute" program on SPSS. As a result, the'mult—response" program was used. This program provides no statistical tests 83 of significance. It merely tables the combined data into one aggregate table. 84 Data And Results The Differences Between Generic And Non-Generic Shoppgrs Intuitively, it would seem that price conscious indi- viduals who save money by purchasing generics would be on a food budget. In Table 5, it was expected that cell four, gen- eric shopping on a food budget would have the highest frequency. On the opposite side it was expected that most people not on a food budget would be non-generic shoppers because they are not as price conscious. As Table 5 results show, this relationship was not as strong as expected. TABLE 5 Difference Between Generic And Non-Generic Shoppers Based On Spending From A Food Budget Non-Generic Generic Shopper Shopper Not on a food budget 67.4% 53.8% On a food budget 32.6 46.2 chi square significance ld.f. .0532 Although the chi square was not significant at .0532, more non-generic shoppers reported that they were not on a budget. The next highest percentage was generic shoppers not on a budget. Thus, as Kunreuther (1973) suggested budgeting does not always imply the lowest priced brand will be selected. 85 There is also the possibility as Goldman (1976), reported, that the non-generic shoppers on a budget may just not be aware of the low cost benefit of generics. Had they known, they may have taken advantage of generic foods. At present no marketing literature reports differences in how people manage their food budgets. Thus, a question was asked as a check to make sure those who responded positively to being on a food budget actually had one. This information did on its own have important ramifications. Table 6 reports a comparison between those generic and non-generic shoppers on the ways they budget their money. The results presented in Table 6 show there was no significant difference between the groups regarding their budgeting methods. However, of both groups, setting aside a lump sum for groceries was the most popular. Also, generic shoppers use more budgeting methods, such as looking for price specials and buying these goods plus using coupons, neither of which non-generic shoppers reportedJ By combining the focus group's responses with the research findings from past studies, the variable, qualities looked for in national brands, was devised. As Norris (1960) implied adver- tising confers an added value to products. So it was expected that this factor would show up as a differentiator between gen— eric and non-generic shoppers. Coyle (1978) states that the Jewel chain originally felt that the generic shopper believed he/she did not need the advertising or labeling. However, results showed 86 TABLE 6 Different Methods Generic And Non-Generic Shoppers Use To Budget Their Food Bill Non-Generic Generic Budget Method Shopper Shopper Lump sum 78.6% 64.6% Coupons 0 6.3 Stick to shop- ping list 10.7 14.6 Percentage of paycheck 3.6 4.2 Experience 7.1 2.1 Price specials O 8.3 Chi-square significance 5 d.f. .3118 that in both groups advertising was not considered an important quality for national brands to have. The most important national quality that both generic and non-generic shoppers considered important was the high quality associated with a national brand. Again, both groups considered consistency and familiarity as close seconds for national brand qualities. As reported in Table 7, there were no statistically significant differences between the qualities generic and non-generic shoppers see in the national brands they buy. The price of generics in relation to their quality is another area where generic and non-generic shoppers differ. 87 TABLE 7 Qualities Looked For In National Brands By Generic And Non-Generic Shoppers Non-Generic Generic Chi-Square National Qualities Shopper Shopper Significance Yes No Yes No Consistency 28.9% 71.1% 21. % 78.4% 2424 Familiarity 23.3 76.7 21.6 78.4 .7698 Advertising 1.1 98.9 2.9 97.1 .3756 High quality 53.3 46.7 48.0 52.0 .8493 Brand Recognition 8.9 91.1 2.9 97.1 0768 Taste 18.9 81.1 28.4 71.6 .1221 Price special 11.1 88.9 9.8 90.2 .7673 Color texture looks 2 2 97.8 2.9 97.1 .7549 Ingredients 1.1 98.9 1.0 99.0 .9291 Low sugar 1.1 98.9 0 100.0 .2858 Freshness 2 2 97.8 2.0 98.0 .8993 The theory of Jacoby et a1. (1971) was applied to generics. His position was that without a brand name or brand image, price alone will affect perceptions of quality. Therefore, when consumers were asked what they felt were the primary benefits of generics, "low cost” was assumed to be the number one answer of both groups, while generic shoppers would indi- cate ”less quality but acceptable for my purposes” most often. The results, reported in Table 8, highlight some inter- esting differences. As expected,the majority of the shoppers view generics as less expensive items, 83.7% of non-generic shoppers and 89.3% of generic shoppers. However, very few 88 TABLE 8 Benefits Of Generic Products As Perceived By Generic And Non-Generic Shoppers Non—Generic Generic Chi-Square Benefits Shopper Shopper Significance Yes No Yes No Low cost 83 7% 16.3% 89.3% 10.7% .24 Same quality 9.8 90.2 24.3 75.7 .0078 Less quality but acceptable 4.3 95.7 2.9 97.1 .5907 No advertising 4.3 95.7 1.0 99.0 .1364 Don't know 7.6 92.4 2.9 97.1 .1378 shoppers in either group commented that generics were of inferior quality but acceptable. Therefore it seems that gener- ic shoppers view the benefits differently. As was discovered, almost 25% of generic shoppers view generics as comparable to national brand quality, versus only about 10% of non-generic shoppers. Logically it would seem that generic shoppers would find the quality satisfactory to the other brands they buy. If they were not pleased, they would not continue buying these items. This relationship was found to be significant. Progressive Grocer was interested in finding out what made generic items different according to the beliefs of the average consumer. They found that generics' lower price was asso- ciated with cheaper packaging and labeling and the absence 89 of advertising in both the published 1978 and 1979 reports. In both 1978 and l979,cheaper labels and no advertising were seen as the reason generics sell for less. In making the comparison between generic and non-generic shoppers, it was speculated that generic shoppers would see the externals, such as packaging, labeling and advertising as what makes generics different while non-generic shoppers would see lower quality the main difference. As indicated in Table 9, the results only partially confirm the expected relationships. More non-generic shoppers, than generic shoppers believed lower quality the main difference bebween generics and national brands, whereas 74% of the generic shoppers did not. This dif- ference was statistically significant. However, the externals mentioned, advertising, price, packaging, production & dis- tribution costs, and labeling, were cited just about equally by both groups, and thus, were not significant. Other internals such as product consistency and nutrition were also reported with equal frequency between both groups. The one finding that closely related to the significant relationship of perceiving generics as having the same quality as all brands, is that more generic shoppers said that they saw no difference between generics and national brands. This difference was statistically significant. That seems to indi— cate that generics are purchased by people who do not see them much differently from the other brands on the market. 90 TABLE 9 Differences Between National and Generic Products As Perceived By Generic And Non-Generic Shoppers Non-Generic Generic Chi-Square Differences Shopper Shopper Significance Yes No Yes No Lower price 43.5% 56.5% 43.3% 56.7% .9765 Lower quality 42.4 57.6 26.0, 74.0 .0151 No advertising 13.0 87.0 18.3 81.7 .3170 Cheaper packaging 17.4 82.6 24.0 76.0 .2535 Low production/ distribution costs 3.3 96.7 1.9 98.9 .5523 No difference 2.2 97.8 21.2 78.8 .0001 Don't know 15.2 84.8 9.6 90.4 .2325 Less consistency 4.3 95.7 4.8 95.2 .8780 Less fancy label 4.3 95.7 5.8 94.2 .6517 Less nutritious 1.1 98.9 1.0 99.0 .9305 Based on one of the original premises of this study, generic shoppers were assumed to be more price conscious than non-generic shoppers. This was thought to be one of the factors that influenc- ed their decision to buy generic goods. Because they are care- ful of what they buy, it would seem generic shoppers would also make a concerted effort to look more carefully for news of sales or other money saving items such as coupons. As seen by the results in Table 10, about the same per- centage of generic as non-generic shoppers actively search for money saving information. The results showed that the differ- ence between the two groups were not significant. 91 TABLE 10 Active Search For News Of Sales Or Coupons By Generic And Non-Generic Shoppers Non-generic Generic Search Shopper Shopper Yes 73.9% 26.I% No 76.7 23.3 Chi-square significance with l dif .6521 Along the lines of information search activity, infor- ‘mation sources consulted for national brands wereci interest. Since generic shoppers may not be so impressed by national manufacturers' advertising, it was postulated that for the national brands purchased they would rely on other sources. This same situation would hold true for labels. Cunningham's (1967) study shows that those higher in risk turn to personal sources more often for information on brands. Thus,assuming non-generic shoppers are high risk perceivers (See Table 15), friends and relatives was the category in which they were ex- pected to have the highest frequency. The results, as presented in Table 11, display some of the supposed relationships. Generic shoppers rely most heavily on their own shopping experience to provide the information they need on goods. Surprisingly,advertising was their second most popular source of information. The other sources were mentioned much less frequently, between 11 and 1 percent with 92 TABLE 11 Information Sources For Brands Purchased Of Generic And Non-Generic Shoppers Non-Generic Generic Chi-Square Info Source Shopper- Shopper Significance Yes No Yes No Friends/relatives 14.1% 85.9% 11.5% 88.5% .5872 Salespeople 2.2 97.8 0 100 .1307 Advertising 45.7 54.3 54.8 44.2 .2573 Experience 42.4 57.6 60.6 39.4 .6731 Government/ company reports 4.3 95.7 1.9 98.1 .3254 Samples 4.3 95.7 1.9 98.1 .3254 Don't know 4.3 95.7 1.9 98.1 .3254 Labels 8.7 91.3 4.8 95.2 .2750 Coupons 1.1 98.9 0 100.0 .2865 salespeople and coupons not consulted at all. On the other hand, non-generic shoppers relied on advertising most fre- quently as providing them with information and experience was a close second. Based on the similiarity of the two groups in using these sources for information, no significant dif- ferences were found. The last set of comparisons made between generic and non- generic shoppers was based on demographic data. Although fre- quency counts and chi square analysis, as presented in previous studies, showed no differences between generic and non-generic shoppers, it was thought that a t-test: of interval-level data may elicit the finer distinctions. Table 12 illustrates this data. 93 TABLE 12 t-Tests: Demographic Differences Between Generic And Non-Generic Shoppers Variable Mean t-value Probability Occupation Non—generic 41.0978 -85 .398 Generic 37.7019 Number of people at home Non—generic 2.6739 ‘2-20 .029 Generic 3.1748 Age of children Non-generic 9.8043 “1-91 .058 Generic 14.8558 Years of education Non-generic 14.0217 -.97 .333 Generic 14.4327 94 Using Reiss' intervally-scaled occupation index based on socio-economic measurement, there was no significant differ- ence between the two groups. The mean for the two groups did show a slight difference. The occupations associated with a 41 rating for non-generic shoppers are: construction worker (40), policeman (40), decorator (40), retail food business (43). The occupations closely associated with the generic Shopper's 37 mean are: fireman (37), agriculture (36). As illustrated by the closeness of the scores,there is quite a bit of overlmxfing. The number of years of education showed no discrimina- ting difference. A matter of fact, the mean level of educa- tion was practically the same for both groups. The average age of children living at home produced no significant differ- ence. Even though not significant, it seems non-generic shop- pers have younger children at home, elementary age, school children. On the other hand, the mean age of the children of generic shoppers is 14 or high school age. This finding is interesting because, intuitively it would seem that older children would be more set in their tastes for nationally branded products and less accepting of generics. However, it works the opposite way. The one demographic finding that showed significance was the number of people at the home address. Raw numbers were recorded and a mean calculated. The non-generic shoppers 95 have on the average 2.6 mouths to feed, while generic shop- pers have 3.1. As well as being consistent with the past generic studies, this result supports Farley (1964) who hypothesized about family size affecting brand behavior. He assumed that larger families may have to use a given dollar income more efficiently than smaller families. Thus large families appear less brand loyal because they consider sub- stituting some brands for others. By inference, this can mean that larger families are willing to substitute generics for national brands due to their possible economic limitations. Since income is often a sensitive subject, respondents were given categorical income groups in which to indicate their total family income. This was analyzed using chi square. Table 13 reports the results. There was no significant difference« between groups. However,it is interesting to note that in this study the highest percent of generic shoppers had incomes under $15,000. The largest number of generic buyers in Murphy and Laczniak's (1979) study had incomes between $19,000-$25,000, while the Progressive Grocer (1979) study had the largest group of generic buyers having incomes between $20,000 to $24,999. Due to all the conflicting findings income cannot be re- garded as a determinant of generic shopping. Some of the difference between study frequencies can most likely be attri- buted to sampling error. However, this study‘s income findings fell in the direction that was expected, more low income shop- pers buy generics. Since there were no significant differences 96 TABLE 13 Differences in Income of Generic And Non-Generic Shoppersa Non-Generic Generic Income Level Shopper Shopper Under $15,000 36.7% 47.5% $15,000-$30,000 35.6 31.3 Over $30,000 27.8 21.2 Chi-Square significance with 2 d.f. .3026 between groups, the percentage of generic and non-generic shoppers in all income categories must be treated as being the same. The question then can be raised, if income and price perceptions are not distinguishing between who buys generic items and who does not, what variable will explain this phen- omenon? High And Low Risk In Regards To Generic Shopping The idea of perceived risk, as originated by Bauer in 1960, was introduced. This meant that consumers were risk aversive and tried to minimize or reduce to some extent ”any expected negative utility” associated with a particular pur- chase. (Peter and Tarpey, 1975, p. 29). By applying the six loss components used by Peter and Tarpey in 1975, some addi- tional explanation of generic shopping was discovered. The six risk components were appraised on the basis of the degree of risk associated with generics. Certain com- ponents showed greater risk than others. Table 14 reports 97 frequencies for each risk component measured. Although this scale achieved an adequate reliability coefficient, the major- ity of the sample perceived slight or no risk at all in these areas. The risk component that received the most responses as ”great” was performance risk. Therefore, more people worry about the food not cooking properly or tasting up to standards. In addition it could also mean that people are concerned about the cleaning products and paper goods not working well. Physical risk also had quite a few "great" and ”moderate" risk ratings. Thus, some people are concerned with the product being safe. Almost three quarters of the respondents saw no social or psychological risk involved with the purchase of generics. Therefore, there is no risk to ego or social stigma attached to buying a cheaper product without a fancy label. Unlike the feeling of the focus group, most respondents did not see it as an inconvenience that generics are in a special section, grouped together rather than integrated into the grocery shelves. Only 32 respondents felt some significant loss of tinmadue to this. Lastly the financial risk of possible disappointment in the quality produces mostly a slight or moderate risk. Some saw it as no risk, probably due to generic's inexpensiveness. While very few consumers saw it as a great risk, they may feel any product that does not perform is a financial loss. 98 TABLE 14 Frequencies Of Degrees Of Risk Associated With Generics Risk Component No Risk Slight Moderate Great Financial 55 85 42 16 Social 178 18 3 1 Physical 94 42 24 40 Convenience 120 47 17 15 Performance 57 52 42 46 Psychological 152 20 13 12 Overall, how well does this theory fit generic shopping? Based on the results as pictured in Table 15, there is a sig- nificant difference between generic and non-generic shoppers according to their perceived risk levels. The highest percent of low risk perceivers were generic shoppers. Conversely, the greatest percent of high risk perceivers were non-generic shoppers. This relationship was expected due to the absence of reinforcing cues associated with national goods, such as fancy labels & packaging, advertising, and high unit price plus the others Roselius (1973) suggested. Thus, it was thought that those with high perceived risk needed those cues to re- lieve risk and would not purchase a generic because it does not offer a risk reliever. Low risk perceivers would not need 99 them. The results in Table 15 lead to the conclusion that the hypothesized relationship between risk levels and generic shopping hold true. TABLE 15 The Difference Between Generic And Non-Generic Shoppers Based On High And Low Risk Perception Non-Generic Generic Risk Shopper Shopper High 57.3% 42.7% Low . 38.3% 61.7% Chi square significance .008 with 1 d.f. p < .05 Turning back to the original study Progressive Grocer (1978) published which reported how generics move, perceived risk was applied to explain this situation. Table 16 shows the results. Aggregate analyses of high and low risk by product category and type of brand bought partially confirm the risk theory. Of all the products listed only canned vegetables and canned fruit showed a significant difference in type of brand by high and low risk perceivers. In explanation of this event, a quote from the focus group describing the low-risk perceivers seems extremely applicable. "I think the vegetables are all pretty good...What can yOu do with a vegetable. You don't make, just grow it and pick." The interpretation of this in relation to the results is that fruits and 100 TABLE 16 Perceived Risk By Product Category And Brand Purchase Behavior Don't Buy National or Generic Chi-Square Product Risk Product Store Brand Brand Significance Toilet Low .9% 74.3% 24.8% Paper High 2.2 80.2 17.6_ .3755 Napkins Low 13.8 56.0 30.3 High 8.8 68.1 23.1 .2008 Paper Low 10.1 58.7 31.2 Towels High 8.8 65.9 25.3 .5728 Canned Low 20.2 61.5 18.3 Vegetables High 14.3 82.4 3.3 .001 Canned Low 16.5 67.9 15.6 Fruit High 9.9 84.6 5.5 .0179 Bath Low 4.6 89.9 5.5 Soap High 6.6 89.0 4.4 .7829 Laundry Low 4.6 82.6 12.8 Products High 3.3 91.2 5.5 .1778 Peanut Low 8.3 76.1 15.6 Butter High 6.6 83.5 9.9 .4140 Mayo- Low 6.4 83.5 10.1 nnaise High 12.1 81.3 6.6 .2849 Spaghetti Low 9.2 74.3 16.5 High 14.3 79.1 6.6 .0691 Tuna Low 13.8 81.7 4.6 High 12.1 87.9 0 .1048 Soups Low 14.7 81.7 3.7 High 12.1 83.5 4.4 .8467 Pet Low 36.7 57.8 5.5 Food High 42.2 57.8 0 .0693 Cake Low 34.3 62.0 3.7 Mixes High 22.0 69.2 8.8 .0776 101 vegetables were the only products described by a package, (i.e., ”canned”). The significant difference with these products may not be totally related to high and low perceived risk. The category of people who do not buy canned fruits or vegetables may buy fresh or frozen. Thus, the chi-square may be picking up significance just by the dif- ference of the high and low risk perceivers in the ”don't buy" category. This finding does not explain why paper products sell so well. The study this information was taken from was con- ducted soon after generics were introduced so Table 3 showed which products were tried first. If paper products were, in fact, less risky items to try, then they should have shown up also in this study as the first products tried. In the trial of a new product, risk definitely plays a role. Arndt (1967) found that with high risk perceivers it influences their communication and purchasing behavior. High risk perceivers tend to be more brand loyal, thus are the late adopters or laggards in new product trial. Therefore, when asked which were the first generics tried, it was predicted that the low risk perceivers would have tried a greater number of products and a larger variety. These basic relationships were confirmed as presented in Table 17. The top four product categories which were tried most frequently were the identical ones to the Progressive Grocer report. In both groups, paper products were tried most, canned vegetables and fruit had the second highest number of triers- Cleaning products had the third largest trial rate, 102 TABLE 17 Generic Products Tried and_Perceived Risk Product Low Risk High Risk Spaghetti l3 5 Crackers 5 4 Jelly/preserves 2 0 Peanut butter 11 8 Mustard/ketchup 2 l Peanuts 6 4 Canned vegetables/ fruits 18 10 Paper products 42 25 Cleaning products 11 8 Pancake/cake mixes l 7 Tea 4 0 Honey 2 0 Spaghetti sauce 4 0 Plastic bags 5 4 Tomato products 4 2 Shortening 2 0 Chocolate syrup 1 1 Apple sauce 2 0 Coffee creamer l 1 Orange juice powder 4 l Tunafish/meats 0 0 Milk 0 l Mushrooms 1 2 Air freshener 0 1 Dog food 1 O Grated topping O 1 Total 142 86 n=107 n=89 103 with peanut butter as a close fourth. The rest of the products display a mixed frequency, all less than the top four. Secondly, although both groups tried the same first four products, the trial rate of most products was lower by the high risk perceivers. Also, of the twenty-six product categories mentioned by the sample, low risk perceivers tried every one but four, versus no trial from eight categories of high risk perceivers. This lends support to the original research conducted by Arndt and others. Product satisfaction was also tapped for this research. After trial, it was thought that high risk perceivers would be less satisfied with generics and return to their original brands. Low risk perceivers would not be subject to this effect. As Table 18 shows, this occurrence does not show statistical significance. Satisfaction seems to be equally felt among both groups. Looking back at the idea of certain products having an inherent risk to them, it was assumed that there might be some generic products that consumers would never consider buying because the risk was too great. First this question was asked of all respondents. The results are presented in Table 19. About the same amount of high and low risk perceivers would not buy some of the generic products. Therefore,there was no significant difference. 104 TABLE 18 Difference Between High And Low Risk Perceivers By Repeat Buying Of Generic Products Some Generics You High Risk Low Risk Would Not Buy Again Perceivers Perceivers Yes 47.7% . 45.4% No 52.3 54.6 Chi-square significance .7421 with l d.f. TABLE 19 Difference Between High and Low Risk Perceivers With The Decision Of Never Considering Buying Some Generic Products Some Generics You High Risk Low Risk Never Consider Buying Perceivers Perceivers Yes 43.2% 31.5% No 56.8 68.5 Chi—square significance .0910 with l d.f. 105 The next step, was to ask of those respondents who would avoid some products, which ones they would ignore. It was assumed that this relationship would show similar results as the previous variable. Certain products would be considered more risky and the low risk perceivers would see less risk in all of the items. Table 20 reveals that this relationship does not hold true. The items were viewed similarly by both groups. As was suggested earlier by Arndt, (1967) high risk perceivers have different communication behavior than low risk perceivers. Cunningham (1967) also confirmed this in his study of informal communication patterns in regards to a new brand on the market. Arndt said the high risk group was more willing to seek information. The relationship asserted for searching behavior was that since it is significantly related to risk, and risk to generic shopping, risk may provide infor- mation about the frequency of search. The results in Table 21 merely support past research. The results are consistent with other studies. Therefore it implies the reliability of the perceived risk measure to predict searching behavior. The chi-square was significant showing that this sample followed previously tested samples. Of those who do search actively, the frequency of this search was requested. According to the results in Table 22, high and low risk perceivers consult information sources with the same frequency. 106 TABLE 20 Perceived Risk By Generic Products Never Consider Trying Product Low Risk High Risk Spaghetti/rice Crackers Jelly/preserves Peanut butter Mustard/ketchup Peanuts Canned Vegetables/ fruits Paper products Cleaning products Pancake/cake mixes Tea Honey Spaghetti sauce Plastic bags Tomato products Shortening Chocolate Apple sauce Coffee creamer Orange juice powder Tunafish/meats Milk Mushrooms Air freshener Dog food Grated topping Total Ol—‘OOOU‘H—‘OI—‘NNOOOOO-DU'IU‘IG ON-Pl—‘OUO r> [\J n=107 OOOOHUOOOHNUOl-‘r-‘NI—‘OWO‘DO HNUJNNN 107 TABLE 21 Differences In Active Search For Information By High And Low Risk Perceiversf High Risk Low Risk7 Active Search Perceivers Perceivers Yes 82.4% 68.5% No 17.6 31.5 Chi-square significance .0243 with l d.f. When analyzing this measure it is important to keep in mind that more high risk perceivers were included than low risk perceivers based on classification of question #13 in the questionnaire. Even with a greater amount of high risk perceivers no significant difference emerged in frequency of search. Therefore there are no special patterns of searching activity high risk perceivers exhibit so marketers would know how to plan promotions for helping to relieve some of this risk. 108 TABLE 22 How Often High And Low Risk Perceivers‘ Look For Information Frequency High Risk Low Risk Of Search Perceivers Perceivers Everyday 25.3% 23.4% Two or three times a week 21.3 22,1 Food day in the newspaper 14.7 22.1 Day do shopping 22.7 22.1 Weekends 4.0 1.3 Three times a month 2.7 1.3 One day a ‘ week 9.3 6.5 Chi-square significance .7816 with 7 d.f. CONCLUSION The emphasis of this study was to provide information conducive to marketing planning for generic products. Due to this orientation the variables examined lend themselves to segmentation strategies. The objectives of the research were to investigate differences between generic and non-generic shoppers, either on a perceptual basis or by their shopping behavior. Table 23 is a summary of the significant results. The approaching conclusion section discusses the results in light of their impact on marketing decisions. Even though the main concern here is for the signifi- cant relationships, some mention needs to be made about a few of the variables that showed no significant differences. To begin with, budgeting or budgeting techniques can be ruled out as a discriminating variable. There are similar amounts of generic and non-generic shoppers who buy on some sort of budget. Therefore generics do not cater mainly to those watching their grocery money. Another variable that did not illustrate any significant differences between non-generic and generic shoppers is what qualities are important in nationally branded goods. The re- sponse which received the most mention was high quality by 109 110 TABLE 23 Summary of Significant Findings Of Present Generic Study Variable Groups Frequencies Table Benefits of generics Same quality as #8 national brands Non-Generic 9.8% Generic 24.3 Differences between national brands and generics Lower quality #9 Non-Generic 42.4% Generic 26.0 No difference Non-Generic 2.2% Generic 21.2 Number of people #12 at home _ Non-Generic x = 2.6 Generic x = 3.1 Risk Non-Generic Generic #15 High 57.3% 47.7% Low 38.3 61.7 Type of brand boughta Canned vegetables National Generic #16 Low risk 61.5% 18.3% High risk 82.4 3.3 Canned fruit Low risk 67.9% 15.6% High risk 84.6 5.5 Generics first tried #17 High risk n = 86 Low risk n = 142 Active search for in- #21 formation Yes No High risk 82.4% 17.6% Low risk 68.5 31.5 All were significant p < .05 aDoes not total 100% because ”don't buy” category not included in table. 111 both groups. Although advertising was cited least often, it has been hypothesized in research that the advertised national brand has the illusion of high quality. Rotfeld and Rotzell (1976) used objective ratings of Consumer Reports information to see if heavily advertised products are of higher quality. Their findings show that only with some products does advertising infer quality. A broad generalization can- not be accepted. The study previously mentioned posed an interesting way of looking at the generic data. Since both generic and non-generic shoppers viewed national brands as high quality goods,can this be attributed to advertising? Looking at the data reported for the differences between national and generic products, it reveals that more non-generic shoppers think the generics have lower quality. Could this be due to the lack of advertising? As Rotfeld and Rotzoll pointed out, this added-value phenomenon cannot be generalized. Thus, the lack of advertising is not what is affecting the sale of generics, since it was not significantly perceived by one group over another. However, what does seem to differentiate between generic and non-generic shoppers is the view of their quality. This is where potential marketing work can help to increase generic sales. By the percentages shown, the non-generic shoppers view generic goods as having lower quality more frequently 112 than generic shoppers do. Thus, they perceive the goods to be inferior. This same attitude shows up again when more generic shoppers perceive generics having the same quality as national brands while only two non-generic shoppers saw no difference. The problem this presents for marketers is quite com- plex. In order to move more non-generic shoppers into the generic group to increase market share, the non-generic shoppers are going to have to be assured of comparable quality to their other brands. Possibly an awareness campaign can be promoted by local retail advertising stressing the quality issue. How- ever, with some products the quality is not always equal to national brands. In the cases where it is not, such as liquid dish detergent, a possible halo effect can be hoped for to avoid false claims. More research for a marketing strategy for generics needs to be conducted to work out this problem. Another problem arises out of additional research. Where are the funds going to come from? Murphy in 1979 reported that generics mainly cut into sales of national brands (Murphy, 1979, p. 14), Progressive Grocer confirmed that national brands are affected greatest (Heller, 1978, p. 80), Assuming this is the situation then, national manufacturers may not be willing to contribute money to increase generic sales even though they may be supply- ing part of their overflow as generic goods. It would only cannabalize their money makers. Thus, the entire marketing effort will probably come from the retailer. 113 This situation is very delicate for the retailer. By making an effort to promote his generics, the retailer does not want to threaten the market position of his national brands or private labels. However, due to the need of the retailer to keep a competitive edge, these low priced goods must not be ignored. Within the marketing framework, another strategy that could be developed to improve generic sales is creating risk relievers. Since one set of results support the idea of generic shopping related to risk levels, this is an important undertaking. More non-generic shoppers are high risk per- ceivers than generic shoppers. This type of behavior was evidenced by the number of generic products tried by each group. The low risk perceivers tried 142 items versus only 86 items tried by the high risk perceivers. The last way this relationship was significantly demonstrated was in the type of brand bought. In the case of canned vegetables and fruits, apparent low risk items, there is a significant dif- ference between the purchase of national and generic brands based on risk. The high risk groups seem to be aversive to buying the vegetables and fruit, which nature made and can- not be so processed that they lose quality. If the generic marketer can offer the high risk perceiver adequate uncertainty relief,he may be able to increase his generic market share. Returning to Roselius' study (1971) on eleven risk relievers, some may be appropriate for the retail 114 outlet to employ. His study showed that high-risk perceivers used brand loyalty most frequently to relieve risk. Since generics do not have any brand identification and are still considered new this will not help. However government testing or approval along with free samples were also rated slightly or very favorably along with word-of—mouth. Thus,generic marketers might consider putting a government stamp on the product to show they have been FDA approved, for example. Providing free samples to shoppers as they stop by the generic section might encourage trial, and possibly stimulate word of mouth conversation about the generics. Further test- ing of relievers appropriate to generics is needed. Since the results also show that the high risk perceivers are more active searchers of information, any exposure may help to reduce risk. Providing some sort of leaflet on product information to shoppers may alleviate some of their doubts concerning quality. The demographic findings marketers have to work with are number of persons in household and educational level as suggested by Murphy and Laczniak (1979). This present study showed that the large families are more inclined to buy generics. This may be due to the larger packages as well as to the low cost. In order to reach the non-generic shopper with less people at home, introducing smaller packages may include the smaller families. 115 Of all the variables tested, there were few relation- ships found statistically significant. However, there is great consistency among the results in discriminating between generic and non-generic shoppers. As was identified in the introduction, generics are facing a leveling out and minimal decline in their life cycle. This problem has been attributed to the neglect of the marketing behind the product once it is on the shelves. In order to reverse or modify the sales trends, the results presented here provide information for developing marketing strategies for generics. Because generics are a unique concept in the grocery business,further investigation into the marketing strategies is warranted. Without advertising, brand labeling, and pro- motion behind them as product supports, generics defy tradi- tional marketing management. Therefore,there is wide opportunity for innovative individuals to provide plans for increasing the generic market. APPENDICES 116 APPENDIX A Unbranded Items Currently Available Grocery Chunk light tuna in water Medium grain rice Long grain rice Elbow macaroni Long spaghetti Spaghetti sauce meat flavored Macaroni & cheese dinner Orange breakfast drink Saltine crackers Graham crackers All purpose crackers Fruit & nut granola cereal Grape jelly Strawberry preserves lbney Creamy peanut butter Crunchy peanut putter Salad style mustard BBQ sauce Salad dressing Imitation mayonnaise Dry roast peanuts Yellow popcorn Chocolate flavored syrup Tonato sauce Catsup Cut green beans Cut beets Diced beets Diced carrots Golden cream style corn Whole kernel corn Sweet peas Tomatoes Whole white potatoes Mushroom stems & pieces Fruit cocktail Pink applesauce Fruit mix Irregular yellow cling peaches Irregular pears Purple plums Geletin (3 varities) Cake mixes (3 varities) Frosting mixes (2 varities) Brownie mix pancake mix White plates White napkins Plastic sandwich bags Trash can liners Lawn & leaf bags Facial tissue Toilet tissue Paper towels Standard aluminum foil Oil Shortening Black pepper Syrup Coffee creamer Black tea bags Instant chocolate mix Evaporated filled milk Dry dog food 25 lb. Laundry detergent phosphate free Laundry detergent-liquid Liquid dish detergent Automatic dishwasher deter- gent Fabric softener concentrated Fabric softener sheets Solid air fresheners (2) 117 Dairy Imitation cheese spread Imitation sliced american 60% spread quarters 60% spread solids Sour dressing Grated Italian topping Monteray Jack cheese HBA Baby shampoo Bath oil Mouthwash (2) Herbal shampoo Strawberry shampoo Aspirin Cotton swabs Tooth paste Plastic strips Source: Darrell Schmuker, Vice-President, Foods Merchandising, Meijer Supermarket, Inc. 118 APPENDIX B Questions Asked Of The Focus Group ShoppingiHabits: Are you on a budget? How do you save money? Do you consider yourself a bargain shopper? Would you describe your shopping time as routine or do you take your time when choosing products? At how many different stores do you do your shopping? Information Sources: Where do you get information about the brands you buy? Do you read labels or go by brand names for quality and ingredients? Consumer Perceptions: What differences, if any, exist between store brands and national brands? What differences exist between generic (un-branded) and national and private labels? Where did you first hear of generic goods? Have you/would you consider buying them? (again) What specific items do you buy? Which do you buy with the greatest frequency? How large a price differential would there be to entice you to try a generic product? 119 Risk: Do you think that your family could notice a dif- ference? Are there some products that you would never consider buying generically? Which ones? And why? Demographics: Number of people in household Occupation of wife Occupation of husband Level of education of both spouses Number of organizations wives belong to How family spends its leisure time. 120 APPENDIX C Questionnaire Code Sheet ? No. Variable Column Code 1 On a food budget 4 l = yes 0 = no 2 How budget works 5 1 = lump sum per week 2 = coupons 3 = stick to shopping list 4 = percent of paycheck 5 = experience teaches good shopping 6 = specials 3 Type of brand 6-19 0 = don't buy the product purchased 1 = national or store brand 2 = generic brand 3a/ Generics first 20-25 spaghetti/rice/noodles tried crackers 7 Generics never 28—33 jelly/preserves considered buy- ing NF