CONSUMER LIFE STYLES AND THEIR RELATIONSHIP TO MARKET BEHAVIOR REGARDING HOUSEHOLD FURNITURE Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY WALTER STEPHEN GOOD 1969 IIIIIIIIIIIIIIHIJIIIIIlIlIlIlIIlIIIIIIIIIIIIHI 1293 10818 5863_ This is to certify that the thesis entitled CONSUMER LIFE STYLES AND THEIR RELATIONSHIP TO MARKET BEHAVIOR REGARDING HOUSEHOLD FURNITURE presented by Walter Stephen Good has been accepted tonrards fulfillment of the requirements for Ph.D. degree in Forestry Major professor Date July 91 1969 0-169 F BINDING m.IIIIAE &_ §_III$' Illl --_.-__ "' """"-"‘O II - ABSTRACT CONSUMER LIFE STYLES AND THEIR RELATIONSHIP TO MARKET BEHAVIOR REGARDING HOUSEHOLD FURNITURE by Walter Stephen Good The household furniture industry in this country has historically been production oriented. Recent failures of the industry to increase its share of consumer disposable income has caused some concern among manufacturers and many realize that to improve their position they must change their orientation and payconsiderably more attention to the peOple who buy their products. This adoption of a more consumer oriented approach means that they must start communicating with consumers and gather information about their nature so that products and promotions can be directed more efficiently towards satis- fying appr0priate market segments. Such information can only be collected by means of sample surveys at the consumer level. To date most of these market research studies have been of a traditional nature relating peOple's behavior to Walter Stephen Good certain demographic characteristics such as age, income, education, etc. It is essential that these basic facts be known, but they do not provide the whole answer. It is felt by the author that these variables may not be very good pre— dictors of market behavior regarding household furniture. Perhaps more relevant information would be variables reflect- ing the individual's taste, attitudes and style of living to a greater degree. To this end, the two major hypotheses investigated in this study are: a) People live according to established-behavior and attitude patterns which can be identified and measured. b) These life-style patterns can be related to consumer behavior regarding household furniture. The data for the study has been collected from a mail questionnaire which was submitted to 2000 women randomly selected from the Lansing, Michigan area telephone directory. By the final cut-off date, 555 questionnaires had been returned. Of this number, 520 or 26 percent of the total were usable and the analysis based on these respondents. The questionnaire, itself, consists of three different sections: a) Six questions relating to standard demographic characteristics of the respondent and her family. b) C) Walter Stephen Good Five questions relating to the reSpondent's market behavior regarding household furniture. Eighty questions relating to how the respondents live, Spend their leisure time and attitudes and Opinions on various subjects related to everyday living. The life-style factors or variables are develOped by subjecting the eighty variables in part C of the question- naire to factor analysis. This procedure has identified a set of fifteen well-defined factors which cover a range of areas related to everyday family life. To test the ability of these variables to predict consumer market behavior,an N-way multiple discriminant analysis has demonstrated a) 1)) been used to separate groups of individuals who certain behavior in regards to: The type of retail outlet at which their last major furniture purchase was made. The styling characteristics of this last major purchase. In each case, three runs have been made using dif- ferent types gives the best separation among groups. of: a) and combinations of variables to see which set These sets consist The series of six demographic variables. Walter Stephen Good b) The set of fifteen life—style variables result- ing from the factor analysis. c) The combination of twenty-one variables. For the first question relating to market behavior, the set of fifteen life-style variables gives the best separation between the groups. These factors are able to correctly classify over 63 percent of the reSpondents who made their last purchase at either a department store or a furniture store. This is Opposed to 56 percent and 60 per- cent for the other two sets of variables. For the question relating to the style of their par- ticular purchase, respondents indicated whether the item can be classified as Colonial/Early American, Provincial, Con— temporary or Spanish/Mediterranean. In this case, the combined set of twenty-one variables gives the best dis- crimination, correctly classifying almost 42 percent of the re5pondents. This is opposed to 33 percent for the demo- graphic variables alone and 40 percent for the life-style variables alone. From these results it can be concluded that: a) Consumers do live according to certain patterns of behavior that can be measured and identified. b) These behavior patterns or life-style factors have greater significance than the demographic variables in being able to predict market be- havior from a practical standpoint. CONSUMER LIFE STYLES AND THEIR RELATIONSHIP To MARKET BEHAVIOR REGARDING HOUSEHOLD FURNITURE by Walter Stephen Good A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1969 p was? /‘ _- .J «' \K) ,\ ,/ I,» l )0 ACKNOWLEDGEMENTS The author would like to eXpress his appreciation to Dr. Otto Suchsland, Chairman of the Guidance Committee, for his tolerance and great patience during the conduct of this study and throughout the entire graduate program. His encouragement and refreshing sense of humor enabled the author to overcome many difficult moments during his aca- demic career. The author is also indebted to the other members of the Guidance Committee, Drs. John L. Hazard and Alan Sliker and Professor William B. Lloyd for their kind c00peration. Special appreciation is eXpressed to the hundreds of Lansingites who took the time and effort to respond to the author's questionnaire and to the many peOple who assisted in compilation of the data and preparation of the final manuscript. LIST OF TABLES. LIST OF FIGURES . LIST OF APPENDICES. Chapter I. II. III. IV. VI. TABLE INTRODUCTION A. Purpose . . . B. SCOpe C. Methodology . REVIEW OF LITERATURE. . . . . THE SURVEY. DEVELOPMENT OF LIFE STYLE A. B. Procedure . . Results OF CONTENTS CONSUMER MARKET BEHAVIOR IN HOUSEHOLD FURNITURE A. B. l. 2. SUMMARY AND CONCLUSIONS . . . . LITERATURE CITED. APPENDIX. Procedure . . Results Type of retail outlet . . Style of furniture. . . . O O O O O O O 0 ii Page iii vi vii award 19 32 40 59 67 68 75 9O 96 100 LIST OF TABLES Table Page 1. Comparison of reSpondents' profile on demographic variables with general statistical data. . . . . . . . . . . . . . . 21 2. Major furniture pieces or sets purchased by reSpondentS. o o o o o o o 0' o o o o o o o 25 3. Type of store or outlet at which respondents' purchases were made. . . . . . . 26 4. Style of respondents' last major furniture purchase. . . . . . . . . . . . . . 27 5. Length of time since the last major furniture purchase by respondents . . . . . . 29 6. Style that respondents would select if. purchasing the same furniture item 3931“. . . 30 7. Percentage of respondents who would purchase the same item in the same style or a different style by length of time since the initial purchase was made. . . . . . . . . . . . . . . . . . . 31 8. Partial correlation matrix used for determination of life—style factors . . . . . 34 9. Description of life-style factors derived by factor analysis of eighty variables in part C of consumer questionnaire . . . . . . . . . . . . . . . . 50 10. Factor 1 - fashion conscious. . . . . . . . . 52 11. Factor 2 - poor housekeeper . . . . . . . . 54 12. Factor 3 - careful shapper. . . . . . . . 55 55 13. Factor 5 — appreciation of the arts . . iii Table 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. nts between Correlation coefficie style variables. . . demographic and life- d actual population member- Predicted an own in Figure 6. . . ship for example sh Normalized confusion matrix for pepulations shown in Figure 6 . . or outlet at which Type of store purchases were made. . respondents' Raw data and normalized confusion matrix for six demographic factors related to type of retail outlet where furniture was purchased . . . . Raw data and normalized confusion matrix for twenty—one life—style and demographic factors related to type of outlet where furniture was purchased Raw data and normalized confusion matrix for fifteen life—style factors related to type of retail outlet where furniture was purchased . rative means for two 1 outlets and fifteen Table of compa types of retai life-style variables. . . last major O O C 0 Style of respondents' furniture purchase. . . . Raw data and normalized confusion matrix for six demographic factors related to style of furniture purchased . life-style factors f furniture purchased . iv Page 57 66 66 7O 70 71 72 73 76 77 79 Table 25. 26. Page Raw data and normalized confusion matrix for twenty-one life-style and demographic factors related to style 80 of furniture purchased . . . . . . Table of comparative means for four furniture style categories and twenty- one demographic and life-style variables 86 Figure 1. LIST OF FIGURES A life style hierarchy. . . . . Graphical representation Of the relationship between two tests, A and B, assumed to involve only two common factors. . . . . . . . Graphical representation of the principal components and rotated solution between two tests, A and B, assumed to involve only two common factors. . . . . . . . . . Discrimination of two populations on two variables. . . . . . . . . Probability that an observation will fall in pOpulation A or B with a given f value. . . . . . . Acceptance regions for three populations on two variables. . . Association diagram for six demo- graphic factors related to style of furniture purchased. . . . . Association diagram for fifteen life—style factors related to style of furniture purchased. . . Association diagram for twenty-one life-style and demographic factors related to style of furniture purchased . . . . . . . . . . . . vi Page 15 35 38 61 63 64 82 83 85 Appendix I II III IV VI VII LIST OF APPENDICES The questionnaire used to obtain information on consumer life-styles and market behavior. . . . . . . . Frequency distribution of respondents' answers to parts A and B of the questionnaire. . . . . . . . . . . Correlation coefficients between the eighty variables in part C of the questionnaire. . . . . . . . . . . Factor loadings of the eighty variables in part C of the questionnaire with varimax rotation for fifteen factors . . . . Description of the fifteen life- style factors derived from the eighty variables of part C of the questionnaire. . . . . . Output of multiple discriminate analysis for fifteen life-style factors related to two types of retail outlets . . . . . . . . . Output of multiple discriminate analysis for twenty—one life-style and demographic factors related to style of furniture purchased . . . . vii Page 100 108 111 131 138 153 160 CHAPTER I Introduction A. Purpose: The household furniture industry in this country has historically been production oriented. Furniture lines are produced by the manufacturer mainly to Suit his factory and production line or on a hunch as to what dealers will buy (13). Recently the industry has been faced with the fact that its share of consumer disposable income has remained static over the past decade and this has caused some concern among manufacturers (29). They wish to provide the consumer with furniture that will make them put a larger share of growing d13posable income into furniture purchases (13). The most obvious approach to this problem is increased recognition of the importance of the consumer by the manu- facturer or a consumer oriented approach to the marketing of furniture. This means communicating with consumers and gathering information about their nature so that products and promotional programs can be directed toward satisfying their needs and desires as closely as possible. This type of approach has generally been neglected to the present time but the necessity of moving in this direction has been recognized and stressed by many prominent members of the industry. Shaughnessy states in a recent article (36): ....our industry must find better ways of doing consumer research. We must find ways and means of learning much more from the users of our product than we know today. Only after we have a clearer picture in this connection can we hOpe to come up with the answers that will enable us to develop better methods of achieving the kind of growth pattern that is expected. Rothberg also feels that the industry must move rapidly in this direction. He states (34): We have fallen badly out of touch with the consumer....The future of retailers and factories alike rest with a com- prehensive, scientific understanding of our customer and how to satisfy her. From these remarks it would appear that the furniture industry recognizes the importance of the consumer to them and that obtaining and using information about their living and behavior patterns can be an important step towards attaining the industry's growth objectives. Such information can only be collected by means of sample surveys at the consumer level. To date most of these market research studies have been of a traditional nature in which a prOperly representative sample of the population is selected by an appropriate statistical method, information collected from the members of that sample about their purchase and use Of particular items or products and the results presented relating peOple's behavior to certain demographic characteristics such as age, income, education, etc. It is essential that these basic facts be known, but they do not provide the whole answer. Often, more weight is given tO this type Of data than is really deserved as the variables are only superficial characteristics of the con- sumer and may not be very good predictors of market behavior for certain types of products. What must be known, in addition, is what are the motives, attitudes, interests and Opinions which lead people to purchase one product rather than another, or to frequent one retail outlet rather than another. This would appear to be particularly true of hOusehold furniture where the style selected for the home is felt to be a reflection Of the individual's taste, atti- tudes and style Of living (2). The Objective Of this study is to develOp a more meaningful description of the consumer than the set of demographic characteristics presently used in most market research and show that this description can be practically related to market behavior regarding household furniture. B. Scope: The two major hypotheses to be investigated in this study are: a) PeOple live according to established behavior and attitude patterns which can be identified and measured. b) These life-style patterns can be related to con- sumer behavior regarding household furniture. As a result, the research is directed toward Obtain- ing information from consumers regarding their activities, attitudes and Opinions on numerous aspects of everyday living and to sort out the underlying order in this large number Of empirical variables in terms of another, smaller set Of variables termed "factors" or "principal components." These factors are then viewed in relation to their value as predictors Of market behavior with respect to househOld furniture purchases. C. Methodology: The primary technique used to gather data for this In study is a mail questionnaire developed by the author. the course Of conducting this study, the questionnaire has been submittedto two sample pOpulations as follows: 1) To pretest the questionnaire wording and question format, the form.was submitted to wives Of staff members and faculty within the Natural Resources Building at Michigan State University. These individuals were asked to try the form and to com- ment on how the content, structure and compre- hensiveness of the questions could be improved. 2) The final improved version of the questionnaire (Appendix I) was submitted to two thousand residents of the Lansing, Michigan area for reply. These individuals were selected randomly from the Lansing area telephone directory and the letters addressed to the female member of the family. Upon receiving the reSponses, the information was coded and subjected to computer analysis on the CDC 6500 available at the MiChigan State University Computer Center. The library programs utilized were BMDOSM and FACTOR A, Technical reports No. 31 and 34 Of the Michigan State Uni- versity Computer Institute for Social Science Research and the BASTAT routine, STAT Series Description NO. 5 Of the Michigan State University Agricultural Experiment Station. From the results of this analysis, conclusions have been drawn regarding the hypotheses previously mentioned and the value Of these techniques on research in this area. CHAPTER II Review Of Literature Due to the nature of this study, the literature in a number of areas which relate to the problem has been reviewed. These areas are as follows: 1) Consumer market studies relating to household furniture purchases and preferences. 2) Life style and its use in market research. 3) Application of factor analysis and multiple discriminant analysis in market research. Since research on each of these topics has been con- ducted relatively independent from each of the others, the areas are discussed individually but an attempt is made to show how each is important to_this present study. 1) Consumer research relating to household furniture. The consumer and research in the area of con— sumer behavior are tOpics that have come to be important to the furniture industry only recently. The first important study was commissioned by the Kroehler Company in 1958 (38) and concerned consumers' needs in furniture and attitudes towards furniture by socio—economic class and stage in the life cycle. The study was redone in 1963 to see what changes in consumer preferences and buying patterns had taken place in the interim. Although this information is presently con- sidered out Of date by the company, presented indicate that life style has considerable bearing 1 . a number of conclusrons on the type and quantity of furniture selected to meet a family's needs. The most important Of these are as follOws: -a family's first concern in furnishing its house is centered upon providing itself with all the items necessary to its chosen way of life. ~furniture needs differ according to social status and they differ according to age. -women experience vague feelings that their whole way of life, and the expression of their personali- ties, is somehow bound up in and defined by the furniture they own and use. More recent studies have tended to follow up on some of these general ideas but have dealt with quite Specific a3pects Of market behavior. Bourne (5) dealt with the influence Of the consumer's reference-group on either a) the purchase of a product, or b) the choice of a particular brand or type, or c) both. He concluded that furniture, like clothing, magazines, and toilet soap, is found in all homes, M Personal correSpondence with Kroehler Manufacturing Company, Public Relations Department, December 17, 1968. causing their purchase to fall outside the area of referencc~ group influence. The visibility of these items, however, together with the wide variety of styles and types available, makes the selection Of particular kinds highly determinant on reference-group influences. Schulte (35) conducted recent investigations into the role played by style in the purchase decision for household furniture. He considered a number Of areas including the importance of style to the purchase of a wide range of furniture items, the stage in the fashion cycle Of today's most popular furniture styles, the difference in acceptance Of furniture styles with varying geographic regions Of the country, the degree Of acceptance of various styles from one room Of the home to another and the style preference in relation to the price Of the furniture item. His major conclusions are as follows: —style was rated as extremely important in the purchase Of occasional tables and sofas and relatively unimportant in the purchase Of recliners and beds. -Mediterranean/Spanish, Country French and Italian Provincial styles are currently in the rapid growth stage of the fashion cycle while Traditional, Modern and Contemporary are well into the declining stage Of their popularity. ~some styles have wider acceptance in some geographic regions than in others but, in total, the differences are not dramatic. -the informal styles tend to be more pOpular in the family room or den and the more formal styles popu- lar in the living room and master bedroom. —the two styles which show the greatest sensitivity to price are modern and Mediterranean/Spanish with modern decreasing in popularity rapidly as price increases and Mediterranean/Spanish increasing in pOpularity with increasing price. A study conducted by the Chicago Tribune (8) in 1959 is somewhat dated but provides an interesting in-depth look at the average consumer of household furniture. Its findings indicated that social class is an important factor in furni- ture consumption and also an important variable in explaining the style of furniture owned by a family. In 1959, modern was the style most frequently found in homes but the pro- portion increased at lower levels Of the social scale. Contemporary, period, traditional and early American were termed "prestige" furniture, and their owners were found most frequently in the middle social class. Ownership of provincial and Danish modern styles, however, appeared to bear very little relationship to social class. They also suggested that each 10 Of these styles seemed to imply very definite images to Examples of these feelings are: their owners. owners Of this ste stress the Contemporary: functional aspects of their furniture but show greater interest in its esthetic value and its versa- tility for blending with other styles. Early American: for these owners, their furniture exudes warmth, comfort—-a homey feel- ing. Provincial: they are lovers Of fine woods. The most recent industry efforts in the area of con- sumer research are a set of three studies commissioned by the Home Furnishings Marketing and Research Council, a cor- poration made up of 18 industry organizations. Each of these reports deals with a different segment Of consumer activity and together they hoped to provide a basis for guiding the industry to better satisfy the needs and wants of its customers. The Arthur D. Little report (3) approaches the problem from the industry level. Its purpose is to examine the role of industry actions in enhancing the sale Of home furnishings, the trade's perception of the impact Of these actions on consumers, and the problems Of securing cooperation among various trade levels. The approach is directed at 11 suggesting how the industry can improve its methods to have a greater effect on consumer purchases of durable goods than they do at the present time. The Social Research and National Family Opinion reports (37, 28) approach the problem from the consumer level and attempt tO evaluate a number of aspects of consumer behavior. National Family Opinion (28) attempts to characterize con- sumers who fit into each Of the following four categories: 1) Those who have purchased furniture within the past year. 2) Those who have not purchased furniture within the past year. 3) Those who intend to purchase furniture within the next 12 months. 4) Those who do not intend to purchase furniture within the next 12 months. The description Of individuals in each category has been based on a profile of demographic characteristics such as age, geographic location, annual family income, family size and composition, and type Of dwelling. Results are presented” for each Of these categories and for a number of product categories such as furniture, rugs/carpeting, and mattresses or boxSprings. For example: 12 Recent Purchasers: Furniture a) Heavier than average furniture purchases were made in the Middle Atlantic, East North Central, and Pacific regions. b) Fifty-four percent Of all furniture purchases were made by homemakers 24-44 years Old. c) The highest percentage Of furniture was bought by those in the $7000-$9999 income levels. d) Eight out of ten furniture purchases were made by families who owned their own homes. While this information is interesting, there is a great deal of overlapping in the variables from category to category so that no clear differentiation exists between consumers who are recent purchasers or are not recent pur- chasers, or who intend to buy or do not intend to buy. The Social Research study (37) eXplores the rationales and motivations which govern the housewife's purchase of home furnishings. It deals with the questions Of why home fur- nishings are important to the housewife and what brings her to market and to a purchasing decision. The approach is Primarily attitudinal or motivational in nature with vari- ations explained on the basis Of life cycle, social class and personality factors. 13 results. Although the area of life style is touched on or implied in a number of the papers, particularly by the Kroehler Report (38), Bourne (5) and Social Research (37) it is not gone into in any great depth as a basis for explaining vari- ability in behavior. 2) Life styles in market research Life styles are becoming recognized more and more as important indicators of the way consumers act in the marketplace. Dr. E. Demby of Fairleigh Dickinson University is quoted by Marketing Insights as stating (25): Attitude and style of living, or psycho- graphics, play a far bigger role in determining which peOple buy what products than do traditional demographic factors of age, income, education, occupation, and size Of family. Their effect on furniture purchases has also been recognized. Edward Frank, when discussing the peOple who purchase the 14 more modernistic styles of furniture stated (2): PeOple who buy these home furnishings don't really fall into an age or financial group as much as they do into an educational or taste group. Lazer (22) defines a life style. He feels it refers to the distinctive or characteristic mode of living, in its aggregative and broadest sense, of a whole society or seg- ment thereof. It is, he states, a major behavioral concept for understanding, explaining, and predicting consumer and business behavior. It is the result of culture, values, resources and other environmental factors. The place of life styles in marketing is illustrated in Figure 1. Wilson (45) has developed a set Of twenty living pattern scales which he relates to the respondents reported market behavior regarding usage of a number Of products, number of hours spent watching television and exposure to a number Of magazines. This information is then cross-referenced to Obtain a regression estimate ranking Of the value of each of these media as an advertising medium for each of the pro- ducts considered. He concludes that such non-demographic variables significantly increase our ability to account for variation within these activities. Pessemier gt a; (31) identify fourteen activity, interest, and Opinion factors and eight personality factors of housewives and discuss the relationship between these 15 [Culture and Society] I.— Group and Individual Expectations and Values Is— FLife Style Pattern }’ and Values "'<"'-‘1 _ - I I I [ Purchase Decisions '- — 4 - - 1 F.— . Parket Reaction of Consumers }"'--J Figure l. A life style hierarchy Source: (22) 16 factors and market related behavior. Specifically, they use these twenty-two factors as independent variables to determine their value as predictors Of a) advertising slogan awareness, b) brand recognition, c) purchase concentration by brand for several product classifications, and d) purchase of a local service commodity. They conclude that non—demographic characteristics markedly improve the description of market segments and they provide predictive power that goes well beyond the capacity of standard demographic measures used in the past. Their results also indicate that the measures on activities, interests and Opinions seem to be better pre- dictors than the standard personality factors. 3) Factor Analysis and Multiple Discriminant Analysis in market research. The statistical techniques used in this study were not specifically designed for marketing or market re- search applications but have found considerable use in this The origin Of factor analysis is generally attributed area. to Charles Spearman (17) in 1904 as a method for determining the common elements prevalent in a series Of psychological tests. The technique is basically a way Of describing the inter-relationships between a large number Of variables in terms Of a few mathematically derived factors. Two of the earliest marketing studies worth noting using this method were conducted by Stoetzel (41) and Twedt (42). Stoetzel theorizes that complex consumer behavior is 17 caused by a few simple motives that can be found in the patterns Of that behavior. He illustrates this premise with a factor analysis which tentatively eXplains consumer pre- ference for nine liquors in terms of their sweetness, price and regional pOpularity. Twedt, in his study of advertising readership, carried the procedure one step further by test- ing the independent variables that the factor analysis suggested were most influential. This was done by calculating the multiple correlation coefficient between these suggested variables and a readership criterion. His results suggest that the predictive value of the variables is quite high. Both Spector (39) and Mukherjee (27) have conducted studies similar to Stoetzel in that they end at the point of reporting the factor loadings and identification of the indicated variables. Spector identifies a set of six general factors which explain peOple's perception of a cor— poration's personality characteristics or its "corporate image." Mukherjee had his subjects rate a cup Of coffee in terms of fourteen attributes to determine which of these attributes is most closely associated with overall preference. His results indicate that individual differences on coffee ratings can be best described in terms Of the variation on comforting taste, heartiness Of flavor, genuineness of pro- duct, and freshness. He considers these four factors the important motivating principles governing consumers' coffee preferences. 18 Pessemier 23.2l (31) and Wilson (45) both use a fac- tor analytic technique in the studies described previously. Their approach, however, approximates that of Twedt where the derived factors are tested by relating them to individual market behavior with the use of multiple regression analysis or similar statistical method. Discriminant analysis has had relatively limited application to date in marketing studies. Evans (12) makes use of the technique to discover which of a predetermined set of variables best predict brand ownership in automobiles. Pessemier e£_gl (31) performed discriminant analyses in an attempt to classify subjects into brand buying categories. Specifically they were searching for activity, interest and Opinion or personality variables which demonstrate high pre- dictive value to determine users of a particular brand of toothpaste, two different brands of cake mixes or subscribers to a TV cable service. Massey (26) feels that the technique has excellent potential for providing a set of aggregate similarity indices for a number of audiences of various advertising media. He illustrates this by evaluating the similarities among the audiences of a number of FM radio stations. CHAPTER III The Survey A five-page questionnaire (Appendix I) was sent to 2000 women in March, 1969. These women were selected at random from the Lansing area telephone directory, which covers most of Ingham County. Ingham.County is situated approximately 80 miles northwest of Detroit in the center of the state of Michigan. Lansing and East Lansing are situated in the northwest corner of the county and are the 'major'metrOpolitan areas within the county. It has a total population of 240,700 people (40) with 125,100 residing within the corporate city lbmits of Lansing and another 35,500 in East Lansing. This population is distributed into 70,000 households with an average annual income of $10,270 per household in 1968 (40). VThis is a good deal above the United States average of $9012 per household for the same year. Most of this income is derived from the payrolls of Oldsmobile- Fisher‘Body Divisions of General Motors Corporation, Michigan State University, and the government of the State of Michigan which has its capital in Lansing. By the final cut-off date, or five weeks after mailing, 555 questionnaires'had been returned, a return rate of 27.8 percent. Of this number 520 or 26 percent of the total were 19 20 fully usable and the analysis has been based on these questionnaires. Unusable returns were caused primarily by failure of the reapondents to answer all the questions and misinterpretation of the questions resulting in answers which could not be included in the analysis. The questionnaire itself consisted Of three different sections: a) Six questions relating to standard demographic characteristics of the respondent and her family. b) Five questions relating to the respondent's market behavior regarding household furniture purchases, specifically their last major (over $50) purchase. c) Eighty questions relating to how the reapondents live, spend their leisure time and attitudes and Opinions on various subjects related to everyday living. The respondents' profile on each of the demographic variables compared with general statistical data for Lansing and Ingham County is shown in Table l. The sample does not appear to be entirely representa- tive of the pOpulation in Ingham County as a whole but does include sOme members of each of the important classifications. Respondents tend to be married, 35 to 49 years of age, be better educated and have higher family incomes than is the case for the general population. The author feels that this 21 Table 1. Comparison Of re8pondents' profile on demographic variables with general statistical data. 1960 State Demographic Number of Percent of Census Journal Characteristic Respondents Respondents Percent Percent Marital Status: Single 8 1.5 21.73 5.4 Married 488 93.8 64.7 79.5 Widowed 14 2.7 10.2 11.0 Divorced 9 1.7 3.3 3.2 Separated l .2 1.2 .9 Family Size: 1 or 2 members 146 28.1 32.4 32.3 3 or 4 members 226 43.5 40.9 40.8 5 or more members 148 _ 28.5 26.6 26.9 Age of Household Head: 24 and younger 27 5.2 15.34 ~6.7S 25 to 34 117 22.5 24.0 23.2 50 to 64 136 26.2 20.3 23.7 65 and Older 28 5.4 11.2 15.9 Education of Household Head: GrahaSohoOl or less 23 4.4 27.14 n.a. Some High school 69 13.3 19.7 n.a. Graduate High School ' 159 30.6 29.2 n.a. Some college 127 24.4 11.3 n.a. Graduate college 142 27.3 12.7 n.a. Table 1 continued Demographic Characteristic 22 Own/Rent‘Home: Own home Rent home Total Family Income-1968: Under $5000 $5000-7999 $8000-9999 $10,000-14,999 15,000-19,999 Over $20,000 1960 Number of Percent of Census Respondents Respondents Percent 473 91.0 72.3 47 --9.0 27.7 23 4.4 30.6 59 11.4 37.0 87 16.7 14.2 226 43.5 13.0 ) 77 14.8 3.8 ): 48 9.2 1.3 ) 1figures for Ingham County. 2figures for Metropolitan Lansing indicated. State Journalz Percent n.a. n.a. 26.3 27.8 16.7 29.2 3percentage of female pOpulation over 18 years old. 4percentage of male population 5figures for 1ngham.County. n.a. Source: (40) not available 23 is to be expected with a questionnaire of this type. It is quite long and requires a reasonable degree of comprehension to complete. This would tend to eliminate the poorer edu- cated and less interested individuals. The emphasis on household furniture appeals to women that have a high degree of interest in their homes and its decoration. The majority of these are felt to be reasonably well educated, middle to high income peOple who own their own homes. A sample with this distribution is not necessarily bad and may be more effective in achieving the objectives of the study than a sample more representative Of the general population. A study recently completed by National Family Opinion, Inc. (28) indicates that a) families with homemakers 25 to 54 years old account for 75 percent of all furniture purchases. b) families with incomes from $7000 to $20,000 account for 64 percent of all furniture purchases. c) eight out of ten furniture purchases are made by families who own their own home. From this it would appear that the sample distribution is very representative of the furniture buying public so the conclusions will be more relevant than a sample based on the general population. The market behavior of the reapondents in regards to their last major (over $50) furniture purchase is shown in 24 Tables 2 to 6. In some cases this information is compared to data from the National Family Opinion, Inc. survey (28) to give an indication as to how representative this sample is of the national furniture buying public. From Table 2, 53 percent of the respondents' last furniture purchases were for the living room or family room, 19 percent for the bedroom and 9 percent for the kitchen or dining room. This compares favorably with the national breakdown of 53 percent for the living room or family room, 22 percent for the bedroom and 10 percent for the kitchen or dining room (28). The most pOpular pieces purchased are living room.sets, sofas, divans and occasional chairs and bedroom sets. These four items account for over 63 percent of total furniture purchases. Furniture stores and furniture departments of depart- ment stores are responsible for over 90 percent of the sales to respondents (Table 3). The independent or chain furniture store is by far the most important retail outlet for the distribution of household furniture. Other outlets such as discount stores and mail-order houses are relatively insig- nificant when considering the total volume of goods sold. Contemporary is the style most favored by respondents. Over 45 percent of thesample purchased items that are in the contemporary or modern category (Table 4). Colonial and Early American pieces are second in pOpularity while Provincial, l) t I.) ) 25 Table 2. Major furniture pieces or sets purchased by respondents Number of Percent of N.F.O. Item Respondents Respondents Survey Living room set 75 14.4 9.01 Sofa or Divan 79 15.2 14.0 Lounge or occasional chair 94 18.1 18.0 Dinette-kitchen set 46 8.8 11.0 End or coffee tables 28 5.4 9.0 Bedroom set 80 15.5 11.0 Chest, dresser, etc. 21 4.0 3.0 Other 95 18.6 25.0 Total -52O 99.9 100.0 1Figures represent purchases in the 12 months prior to July, 1967. Source: (28), Appendix II 26 Table 3. Type of store or outlet at which respondents' purchases were made. Number of Percent of N.F.O. Store or Outlet Respondents Respondents Survey Department store 154 29.6 24.01 Discount store 3 .6 -- 2 Furniture store 323 62.1 64.0 Mail-order house 3 .6 -- 2 Interior design shop 15 2.9 2.0 Wholesale outlet 2 .4 -- 2 Other 20 3.8 10.0 Total 520 100.0 100.0 lFigures represent purchases in the 12 months prior to July, 1967. 2Included in "Other." Source: (28), Appendix II 27 Table 4. Style of respondents' last major furniture purchase. ‘ Number of Percent of N.F.O. Style‘Category' Respondents .Respondents Survey Colonial/Early 1 American 172 33.1 34.1 Provincial 75 14.4 9.8 Contemporary 236 45.4 50.62 Spanish/ Mediterranean 37 y 7.1 5.5 Total 520 100.0 100.0 lFigures represent purchases in the 12 months prior to July, 1967. 2Includes figures for both Modern and Contemporary. Source: (28), Appendix II "J 28 Spanish and Mediterranean pieces claim smaller segments of the market. Respondents appear to be relatively satisfied with the styling characteristics of their original purchase. A number of them would now prefer to have the same item or set in a different style but they are a relatively small percentage of the total. Contemporary appears to be suffering the most from changing tastes, declining 7 percent in popularity (Table 6). Spanish/Mediterranean and Provincial registered the biggest gain in preference, increasing 4.2 percent and 1.6 percent reSpectively. The desire to change to a different style appears to be directly related to the time since the item was purchased. From Tables 5 and 7 we can see that over 90 percent of those respondents who made their purchase within the last 12 months are satisfied with their original style selection while only 77 percent of those who purchased the piece over 5 years ago would choose the same style category if purchasing the same Piece again. This may be due to either changing tastes on the part of consumers or merely a reflection of the fact that they tend to tire of a particular style after a time and want to get something different. 29 Table 5. Length of time since the last major furniture purchase by respondents Number of Percent of Eléifi Respondents Respondents Within last 12 months 236 45.4 1 to 2 years 127 24.4 2 to 5 years 122 23.5 Over 5 years 35 6.7 Total 520 100.0 Source: Appendix II Table 6. Style Categogy Colonial/Early American Provincial Contemporary Spanish/ Mediterranean Total Style that r if purchasin item again ’ Number of Respondenps 178 83 200 59 520 1Figures represent int following July, 1967. Percent of Respondenps 34.2 16.0 38.5 11.3 “ 100.0 eSpondents would select 9 the same furniture ended purchases in the 12 months 2Includes figures for both Modern and Contemporary. Source: (28), Appendix II f—‘ C2) 31 Table 7. Percentage of respondents who would purchase the same item in the same style or a different style by length of time since the initial purchase was made Length of time since Would purchase Would purchase last majprypurchase same style group different stylepgroup Within last 12 months 91 9 1 to 2 years 91 ' 9 2 to 5 years 83 16 Over 5 years 77 23 Overall 88 12 CHAPTER IV DEVELOPMENT OF LIFE STYLE FACTORS A. Procedure: The life style factors are develOped by subjecting the eighty questions or variables in section C of the question- naire (Appendix I) to a factor analysis. The computer program used is Factor A: Principal Components and Orthogonal Rotations develOped by the Michigan State University Computer Institute for Social Science Research (CISSR). I Factor analysis is basically a technique for represent- ing a large number of tests or measurements, each made on many. objects or persons, in terms of some smaller number of variables or factors. It describes the inter-relationships between this large number of variables in terms Of a few mathematically de- rived factors. A number of numerical procedures are available for performing factor analysis but discussion here will be limited to principal components analysis since it is the basis for the CISSR program. Principal components analysis is a technique which sYStematically extracts factors sequentially that have very little correlation with one another and such that the first Will explain as much as possible of the variation in the Original measurements, the second will explain as much as Possible of that left unexplained, and so on- 32 33 The general starting point for conducting a factor analysis is the correlation matrix resulting from calculation of the correlation coefficients between each pair of measure- ments. A portion of the matrix obtained for this study is illustrated in Table 8. The complete matrix encompasses all 80 questions or variables from part C of the questionnaire. (See Appendix III). From this matrix, factor analysis extracts the under- lying factors which are independent Of one another and which account for most of the variability in the original set of data from which the original intercorrelations were obtained. The computations involved in this procedure are quite complex mathematically but basically involve the solving of many equations simultaneously, one for each correlation in the matrix. An example of the type of equations involved is as follows (32): r = (A's loadin on factor I) X(B'S loading on AB 9 factor I) +(A's loading on factor II) x (B's loading on factor II) + . . . . . + (A's loading on factor N) x(B's loading on factor N) rAB = correlation coefficient between variable A and variable B The solution of a case between two tests assumed to have only two common factors can be shown graphically as in Figure 2. The points A and B have been joined to the center 34 oo.a 0H ma.o oo.H NH.o oa.o no.0 es.o ma.o ~o.o- eH.o ao.o so.o oH.o oo.a mo.o| Ho.o mo.o mo.o oo.H oa.o ~o.o- mate oo.H Ho.OI oo.o oo.H ma.o oo.a m n o m o HHH xwocomm< ma.o no.0 mo.OI ma.o NH.o no.0 oH.o oo.H ”oouoom mH.o ma.o on mo.o No.o m eo.ol oo.ol- m oo.o mo.o e oo.ol mo.on o ¢H.o oo.o m mo.o Ho.ol v Ha.o NH.o m oo.H om.o N oo.a H N H manowuo> muouomw adhumuomwa Mo coaumcaeuouoo now new: xfluume cowuoaouuou Hmwuwmm .m OHQOB 35 8\ Graphical representation of the relation? ship between two tests, A and B, assumed to involve only two common factors. Figure 2. Source: (18) 36 or origin and the square of the radiating lines or "test vectors" is equal to the communality of the variable or the correlation of the variable with itself. In this example ha = hb = l but this is not necessarily the case. The cir- cumscribed circle of unit radius is intended to represent the fact that any point on the circumference indicates that the whole of the variance of that set Of measures is completely accounted for by the two common factors. The correlation coefficient between A and B is given by the equation (18) rAB = ha hb cos gAB and the factor loadings Of variable B on factors I and II are represented by B1 and 32 respectively. From Figure 2 it is evident that the relationship between tests A and B would be preserved even if the reference axes of factors I and II were allowed to rotate to new positions. Rotation of these axes is often used as an extension of prin- cipal components analysis in order to obtain more meaningful factors. Orthogonal rotations are one of the more popular approaches to this problem, generally using either the varimax or Quartimax methods. Generally Speaking, the Quartimax method is a method of rotating the axes so that each measurement is described in terms of as few factors as possible. The varimax method, on the other hand, obtains a rotation Of the factor axes so as to minimize the number of measurements in which any one factor occurs (9)- 37 TO illustrate the solution shown in Figure 2 we can assume a hypothetical case where we are dealing with only two of the eighty questions. Again we will assume that no errors of measurement exist and that all of the variance can be accounted for by the two hypothetical common factors so that ha = hb = 1. The correlation coefficient between tests A and B has previously been defined as: l) rAB = ha hb cos ”AB a) r (A's loading on factor I) x (3'5 loading on AB factor I) + (A's loading on factor II) x (3'3 loading on factor II) The coefficient, r , can be determined by both formulae from AB information available in Figure 2. Figure 3 shows a specific case where the angle between the positions of variables A and B is 45“. Therefore, r AB 18 equal to .71. To determine r from the factor loadings of each AB variable, we see from Figure 3 that the loading of variable A on factor I is .52 and on factor II is .86. Similarly, the loading of variable B on factor I is .96 and on factor II is .25. Therefore, r is equal to (.52) (.96) + (.86) (.25) AB or .71. This is the principal components solution to the problem. Figure 3 also serves to demonstrate that the relation- ship between A and B is preserved even if the reference axes 38 Graphical representation of the principal components and rotated solution between two tests, A and B, assumed to involve only two common factors. Figure 3. 39 of factors I and II are rotated to new positions. If I and 11' represent the new position of these reference axes, the loading of variable A on these new factors becomes .22 and .98 and that of variable B becomes .84 and .54. Deter- mining TAB from these new loadings_gives a result of .71 or the same as the principal components solution. This information is generally presented in tabular form as follows: Principal Components Solution Rotated Solution Factor I II I' II' A .52 .86 .22 .98 B .96 .25 .84 .54 Each Of these solutions is a valid solution to the problem and it is up to the researcher to select the one which gives the more meaningful and identifiable factors. This explanation is a great simplification in reference to this study since, rather than two variables or tests, we are dealing with eighty. A graphical explanation of this case is impossible since each factor must be represented by a As a result, anything beyond the two factor separate dimension. In our case, with case is extremely difficult to illustrate. fifteen to twenty well-defined factors, it becomes impossible. 40 B. Results: The results of the factor analysis on the eighty questions in part C of the questionnaire are shown in Table 9. The generation of fifteen factors by the Varimax Rotation method accounts for 49.84 percent Of the eXplainable variation These initial fifteen factors in the entire set of variables. Although are quite clearly defined and readily identifiable. additional variation could be accounted for bygenerating more factors, it was felt that the difficulty in identifying the dimensions and the small pr0portion of total variation accounted for by each additional factor does not warrant including any more in the analysis. Table 9 shows the primary dimension for each of these factors and the percentage of total variation accounted for by each. Table 10 to Table 13 illustrate some of the basic con- cepts of factor analysis and show how these dimensions are determined. Each table contains three columns of information: "Question Number" and “Mean Score” are simply reportings of the items in part C of the questionnaire and the average score of all reapondents on that particular item based on the 1-5 scale used with the questions. The final column, “Factor Loading” presents the results of the factor analysis. The remainder Only four of the strongest factors are presented here. are presented in simdlar form in Appendix V of this study. ’ "/1'5 0 Table 9. Description Of life-style factors derived by factor analysis of eighty variables in part C of consumer questionnaire PrOportion of Total variance Factor Factor Name Percent Number 1 Fashion Conscious 5.07 2 Poor Housekeeper 4.31 3 Careful Shopper 3.52 4 Disinterest in community affairs 4.51 5 Appreciation of the arts 4.06 6 Sports Spectator 3.24 7 Do-it-yourself homemaker 2.58 8 Conservative shOpper 3.05 9 Child oriented 3.53 10 Modern Thinker 3.62 11 Energetic 2.77 12 Weight conscious 2.42 13 Sports participant 2.51 14 Socialite 2.83 1.82 15 Self-centered 51 Since factor analysis is a methodology by which a series of items or questions are identified which tend to form a com- mon pattern and not all questions are identical, it follows that all questions do not measure the basic underlying dimension to the same degree (45). The analysis provides a "loading" to indicate the degree which a question measures the particular dimension. This is a numerical quantity ranging from -1.00 to +1.00 and the higher the positive loading, the more the par- Negative ticular question defines the underlying dimension. loadings are interpreted as just the reverse. In practical work with factor analysis, loadings as high as i 1.00 never appear. Generally, a loading of 1.40 or up is considered quite good and one over 1.50 to be strong (45). In this study the author has adOpted a loading of 1.40 as being the minimum acceptable. An exception to this is factor 15 where only one variable had a loading greater than $.40 so the acceptance level was lowered to 1.30 to obtain a better concept of the dimension. As shown in Table 10, factor 1 has been labelled ”Fashion Conscious." The process of naming a dimension de- rived by factor analysis is a rather subjective procedure. Basically, the researcher puts a label on the results that, to him, best typifies the common element that seems to underlie the various questions in the Scale. From Table 10 we can see that each of the first five questions have high positive 52 Table 10. Factor 1 - fashion conscious Question Mean Factor Number Score Loading 41 Fashion in clothes is more important 1 than comfort to me. 4.10 .70 59 Dressing fashionably is an important part of my life. 3.24 .69 19 I enjoy trying the latest style in hairdo's. 3.18 .66 35 I use eye shadow or eye liner three times a week or more. 3.63 .66 63 I have COpied the way peOple dress on television or in magazines. 4.00 .60 66 I dress for comfort, not for fashion.2.39 -.56 38 I generally have at least one outfit that is the very latest Style. 2.67 .54 62 I presently own a wig, fall or other hairpiece. 3.70 -51 55 I have several different shades of lipstick to go with different dresses. 2.79 .44 PrOportion of total variance - 5.07% 1The smaller the mean score, the greater the general agreement with the question. 53 loadings and refer to placing great emphasis on dressing fashionably and the use of techniques and items which tend to make a woman appear to be "in fashion." Item six "I dress for comfort, not for fashion" appears to deviate from this trend but since it has a negative loading, the Opposite mean- ing must be assumed so that it does follow in the same line. The remaining items also support the earlier supposition so "Fashion Conscious" would appear to be a reasonable label for this particular factor. The same procedure was used to label each of the fac- tors. Table 11 shows that factor 2 has high negative load- ings on questions relating to cleaning and maintaining a well- run home and a positive loading on the question indicating carelessness or lack of concern for these matters. For this reason factor 2 has been labelled "Poor Housekeeper." Similarly from Tables 12 and 13, factors 3 and 5 have been labelled "Careful Shopper" and "Appreciation of the Arts" reapectively. These fifteen life style factors have been used in the analysis of consumer purchase behavior regarding household furniture. To facilitate this procedure, it is necessary to determine a factor score for each respondent on each of the fifteen factors. This has been done by means of the follow- ing equation: 54 Table 11. Factor 2 - Poor Housekeeper Question Mean Number Score 2~ I really enjoy most forms of housework. 2.32 40 I really enjoy cleaning my house. 2.51 58 My idea of house cleaning is "once over lightly." 4.04 73 Keeping my home nice satisfies my creative needs. 2.43 18 My husband compliments me on the way I run the house. 2.43 l I Often redecorate my house or apartment. 3.06 PrOportion Of total variance - 4.31% Factor Loading -.81 -.ao .65 -.51 -.43 -.42 55 Table 12. Factor 3 - careful shOpper Question Number 37 17 4 56 Mean Score I study the food ads each week so I can make the best buy. 2.67 I shop for specials. 2.33 When I find a coupon in the paper I clip it and redeem it the next time I go shOpping. 3.05 I watch the advertisements for announcements of sales. 2.25 Proportion of total variance - 3.52% Table 13. Factor 5 - appreciation of the Question Number 52 45 68 15 Mean Score I enjoy listening to classical records 2.84 I generally prefer classical to the more popular forms of mUSlC. 3.16 I enjoy going to concerts. 2.97 II enjoy going through an art gallery. 2.82 I enjoy spending leisure time in museums. 3.22 Proportion of total variance - 4.06% Factor Loading .77 .72 .71 .70 arts Factor Loading .82 .76 .75 .65 .55 56 n . . . Factor Score i j- E51 (factor loadingk) (questionnaire valueik)/n where i = respondent i = l - 520 Factor number j = 1 - 15 k = variable number k = 1 - n n = number of variables with factor loadings greater than : .40 in factor j. With this formula, each respondent is assigned a.score on each factor ranging from +5.00 to -5.00. Due to the coding of the original questions from 1 - Strongly agree, to 5 - Strongly disagree (see Appendix I) and the influence of negative factor loadings, the lower the absolute value of the respondent's factor score, the higher her rating on the factor. The question may arise as to whether we are really measuring anything different with this type Of life-style variable than with the regular demographic variables used in previous market studies. Table 14 answers this question. It shows the correlation coefficients between the fifteen life-style variables and the six demographic variables in Part A Of the questionnaire. From the table it can be seen that many of the factors are relatively independent of the demPgraphics and even though a number of the coefficients are Statistically significant, the values of the coefficients are 80 small that only a weak linear relationship is indicated. 57 Correlation coefficients between Table 14. demographic and life-style variables Correlation with Demographic Variables Variable 5:1 5;; A23 5:1 A-5 A-l Marital status - - - - - A-2 Family size -.08 - - — - A-3 Age of household head .04 -.34 - - - A-4 Education of house— hold head -.03 -.00 -.20 - - A-S Own/rent present home .03 -.09 -.20 .05 - A-6 Total family income -.22~ .06 .04 .37 -.15 C-1 Fashion conscious .04 -.Ol -.23 .17 .09 C-2 Poor housekeeper .06 .06 —.09 .21 .10 C-3 Careful shopper -.04 .12 .09 -.07 -.05 C-4 Disinterest in com- munity affairs -.05 -.08 —.l4 -.14 .12 C-5 Appreciation of the arts .00 -.04 .06 .28 .03 C-6 Sports Spectator .00 -.09 .09 .03 .07 C-7 Do—it—yourself home- maker -.01 .10 -.13 .07 -.02 C-8 Conservative shopper -.05 -.02 .09 -.09 -.08 C-9 Child oriented .04 -.02 .04 -.12 -.02 C-10 Modern thinker -.02 -.01 -.06 .25 .02 C-11 Energetic —.04 .06 -.12 .07 .05 C-12 Weight conscious .05 -.15 .22 -.01 -.03 C-13 Sports participant .05 .00 -.15 .02 .01 C-14 Socialite .01 -.23 -.01 .13 .02 C-15 Self-centered .01 -.08 -.03 .11 .06 of i .11 or larger is a/ a correlation coefficient significant at the .01 level. .18 .11 -.19 -016 .12 .05 58 The largest coefficient, .28 between variables A-4 and C-5, accounts for less than 8 percent of the total variation between the two variables. Some explanation may help to clarify the implications of the data in this table. Several life-style patterns such as interest in watching sports events (factor 6), conservative shOpping behavior (factor 8), and self-centeredness (factor 15) appear to be well distributed throughout all socio- economic groups. Other patterns, such as fashion conscious- ness (factor 1) bear some relationship to the socio-economic classifications. Highly fashion conscious women appear to be younger, married to husbands who have a high level of formal education and have higher than average total family incomes. Careful shOppers (factor 3) tend to have larger families and lower family incomes as would logically be expected. CHAPTER V CONSUMER MARKET BEHAVIOR IN HOUSEHOLD FURNITURE A. Procedure The primary technique for relating the life-style factors develOped in the previous chapter and the demographic variables from part A of the questionnaire to certain aspects of consumer market behavior is the method of N-way multiple discriminant analysis. It is a statistical technique for making forecasts or estimating structural parameters in problems where the dependent variable appears in dichotomous form, i.e. did or did not purchase Provincial styled furniture. Its use and interpretation are much the same as in multiple regression analysis, i.e. a linear combination Of numerical values for two or more independent variables is used to pre- dict the behavior Of a dependent variable. The computer program used for this analysis is BMDOSM: Maximum Likelihood Classification also developed by CISSR. Basically, the procedure in this case attempts to pre- dict to which group an individual belongs, based on the sets of group means on each variable considered, together with the set of sample variances and co—variances of the variables. That is, the individual is assigned to the group whose characteristics are most like his own. Since it is known 59 6O beforehand which group the person actually belongs to, a table of correct and incorrect classifications can be pre- pared. This table is commonly known as a ”confusion matrix" and the fewer the misclassifications of individuals to groups within the matrix, the more distinct or dissimilar are the groups. To illustrate this technique we shall briefly discuss the two-way situation with two pOpulations and only two variables. The initial step in this form of analysis is to estimate the coefficients in a linear discriminant function. An example of this type of function in terms of two variables, x and Y, is as follows (26): fi = cxxi + cin The subscript i is for each individual considered in the analysis. A critical value of f is determined such that if the individual's f value is above the break-point he is classified in one group and if it is below it he is assigned to the other. The function f is defined so that it discriminates between members of the two groups in the most efficient manner. For example, assume there are measurements on two variables for a sample drawn equally from two pOpulations, A and B. Figure 4 represents the scatter diagram for this sample. Now we have an additional measurement and it must be assigned to either A or B in a way that minimizes the probability of misclassification. 61 Y break-point a b a b -".mrsclassrfrcatrons a be! 0 C a s Papulation A Population 8 region region Figure 4. Discrimination of two populations on two variables. Source: (26) (the c's) Of the discriminant function on the information pro- vided by the original Observations. Having values for the c's allows us to assign a value of f to any possible combination Of X and Y, whether from the original sample or a new sample. We can then use mathematical methods to estimate the prob- ability that, given a particular value of f, the observation would fall in A. The probability distributions for popu- lations A and B may look like those illustrated in Figure 5. Figure 5 also contains a vertical line which represents the discriminant or break-point value of f. The break-point is set half way between the means of f for A and B, so, at this point, an Observation has about an equal probability of falling in A or B. The shaded areas on either side of the break-pOint give the total probability of misclassifying a particular observation. The same concepts apply to the general or N-way case as well as the two-way example. (Expansion of the case to consider three pOpulations of individuals (A, B and C) describable in terms of two variables (X and Y) shows more clearly how the procedure works with a number of groups and a number of variables. Figure 6 has a scatter diagram which shows the variable values Of individuals in each of three pOpulations. The problem in this example is to define three mutually eXClusive regions (a, b and c) which exhaust the x-r space. Population A / l7 N2 L vvvvv \/ a in population A or Source: (26 ) 64 2h Region I JJLL E x Population A o Populatilu II * PopuIaliuu II [J Pnpularinn means Region c as Figure 6. Acceptance regions for three populations on two variables. Source: (26) 65 The region boundaries should be set up so that when the X and vaalues put an individual into a given region, it is more probable that he actually is a member of that population than of any other pOpulation. In Figure 6 the lines separating the three regions represent loci of equal probability for their reSpective pairs of regions. When a given observation lies to one side of the threshold line, the probability that the Observation belongs in this region is greater than for any other region. This maximum probability criterion does not eliminate mis- takes in classification but if the classification process were repeated many times with similar groups of individuals, this procedure would result in the lowest possible prOportion of errors. Each of these regions is determined by using the sample data to estimate the parameters of linear discriminant functions for the pOpulations. These are represented by the lines 23. Zb and 2c in Figure 6. Once the parameters of these functions have been estimated, the boundaries are set so that each discriminant line (2) bisects the angle between its respective boundary line, i.e. Za bisects the angle between ab and ac. While the means of the variables fall within the acceptance regions for their reSpective populations. they do not have to lie on the discriminant lines. 66 Figure 6 also provides the information for using con- fusion matrices to evaluate the similarity Of populations. Table 15 shows the correct and incorrect classifications for the example while Table 16 shows the same information in normalized form. Table 15. Predicted and actual pOpulation membership for example shown in Figure 6. Predicted Actual A B C TotaII A 12 1 2 15 B 1 l3 1 15 C 4 1 10 15 Total 17 15 13 45 Table 16. Normalized confusion matrix for populations of example shown in Figure 6. Predicted Actual A B C Total A .80 .07 .13 1.00 B .07 .86 .07 1.00 C .27 .07 .67 1.00 67 The figures along the diagonal indicate that the number of correct predictions for B is greater than either A or C. The Off-diagonal values indicate that a member of A is most likely to be misclassified as a C, and a member of C to be wrongly classified as an A. From this one can conclude that populations A and C are more nearly alike than are A and B, or B and C. This procedure has been used in this study to evaluate the similarities between groups who have demonstrated par- ticular market behavior relating to two important aSpects of household furniture purchases; 1) The type of store or outlet at which their last major furniture purchase was made. 2) The style category to which this last major purchase belongs. B. Results An N-way multiple discriminant analysis has been per- formed on the questionnaire data using questions B-2, "At what type of store or outlet was this purchase made?” and B-3, "Of the attached diagrams, which page has a drawing which best represents the styling of this item or set?" as the dependent variable. (See Appendix I). In each case, discrimination has been attempted using three different combinations of independent variables: 68 l) the Six demographic variables in section A of the questionnaire. 2) the fifteen life-style variables derived from the eighty questions in section C of the questionnaire. 3) the combination Of twenty-one demographic and life-style variables. This enables comparison of the ability of each type of vari- able to provide maximum.separation among each group in the dependent variable and also to measure the complementary effect of combining both types Of independent variable. 1) Type Of retail outlet In the original questionnaire design, respondents had seven different alternatives or choices from which to identify the type of store or retail outlet at which their last major furniture purchase was made. However, it can be seen from Table 17 that the two Options, department store and furniture store, account for over ninety percent of the total replies. Because of this highly skewed distribution and the tendency of the computer program, BMDOSM, to equalize the frequency of occurrence among. groups, the author has decided to restrict the analysis to a two-way discrimination between these two groups rather than the seven-way analysis that was originally intended. These two types of outlets are by far the major retail distribution channels for house- hold furniture so discrimination between them should be much more meaningful and easier to comprehend. 69 Table 18 Shows the confusion matrix for the classi- fication of 477 reSpondents on the basis of the six demo- graphic variables in both raw and normalized form. Entries on the diagonal of the raw data matrix signify correct classifications or hits, while the off-diagonal elements represent misses. For this case the percentage of hits is 56.6%. The nonmalized matrix is obtained by dividing each of the raw data entries by their row total. These new entries represent the probabilities that an individual who is actually in a given group will be so classified. This result of correctly classifying 270 of the 477 reSpondents is not very indicative of good predictive ability in a two-way discrimination. A completely random basis of classification, such as flipping a coin, would have correctly classified approximately fifty percent of the sample. Increasing this percentage to 57% by using the demographic variables can not be considered a significant increase. Somewhat better results are Obtained if the demographic variables are combined with the fifteen life-style variables and discrimination attempted on the basis of all twenty-one factors. Table 19 shows that this procedure correctly classifies 289 of the 477 individuals or 60.6%. This is an improvement over the previous case but the proportion of hits is still not high enough to state that these variables have (great predictive value for discriminating between the two groups. 70 Table 17. Type of store or outlet at which reSpondents' purchases were made. Number of Percent of Store or Outlet Respondents Respondents Department store 154 29.6 Discount store 3 .6 Furniture store 353 62.1 Mail-order house 3 .6 Interior-design shOp 15 2.9 Wholesale outlet 2 .4 Other 20 3.8 Total 520 100.0 Table 18. Raw data and normalized confusion matrix for six demographic factors related to type of retail outlet where furniture was purchased. a) Raw data matrix Predicted __‘ Actual Department store Furniture Store IITOtal Department store 86 68 154 Furniture store 139 184 323 Total 225 252 477 Total hits = 270, Percent hits = 56.6% b) Normalized matrix Predicted Actual Department store Furniture Store Total Department store .56 .44 1.00 Furniture store .43 .57 1.00 71 Table 19. Raw data and normalized confusion matrix for twenty-one life-style and demographic factors related to type of outlet where furniture was purchased. a) Raw data matrix Predicted Actual Department store Furniture store Total Department store 100 54 154 Furniture store 134 189 323 Total 234 234 477 Total hits - 289, Percent hits = 60.6% b) Normalized matrix Predicted Actual Department store Furniture store Total Department store .65 .35 ' 1.00 Furniture store .41 .59 1.00 Discrbmination based solely on the fifteen life-style factors gives the best results for separating individuals who made their last major furniture purchase at a department store or a furniture store. From Table 20 we can see that these factors correctly classified 302 individuals or 63.3% of the sample. This is an improvement of almost seven percent over the case using only the demographic variables. A percentage Of hits in this range does not mean that this type of factor can be accepted without question for predicting consumers' buying patterns but it does indicate that their attitudes and opinions regarding the way they live do have considerable influence on where they shop. 72 Table 20. Raw data and normalized confusion matrix for fifteen life-style factors related to type of retail outlet where furniture was purchased. a) Raw data matrix _: Pregicted _7 Actual Department store Ffifhiture store Total Department store 100 54 154 Furniture store 121 202 323 Total 221 256 477 Total hits = 302, Percent hits = 63.3% a) Normalized matrix Predicted Actual Department store Furniture store Total Department store .65 .35 1.00 Furniture store .37 .63 1.00 Table 21 illustrates some of the differences between the two groups in terms of their life style. It presents the mean scores for individuals in each of the two groups on each of the life-style factors. Because of the scaling procedure used in the questionnaire and the method for determining the individual factor scores, small absolute values for the mean indicate a high degree of association with the factor. For example, in variable 1 the score Of 1.65 for furniture store shOppers indicates that they are more fashion conscious than dePart'ment store shOppers who have a mean score of 1.78. Also, they tend to be more conservative in their shopping behavior as indicated by a mean score of -.04 Opposed to a Score of -.08 for department store shoppers. 73 Table 21. Table of comparative means for two types of retail outlets and fifteen life-Style variables. Department Furniture Variable Store ‘ Store Fashion conscious* 1.78 1.65 Poor housekeeper 7 - .81 - .79 Careful shopper* 1.80 1.93 Disinterest in community affairs*-1.84 -1.98 Appreciation Of the arts 2.12 2.11 Sports Spectator 2.87 2.77 Do-it-yourself homemaker 1.96 1.94 Conservative shOpper - .08 - .04 Child oriented 1.08 1.06 Modern thinker -l.55 -l.60 Energetic .71 .65 Weight conscious 1.93 1.97 Sports participant 2.23 2.23 Socialite 1.83 1.77 Self-centered - .24 - .27 *Means are significantly different at .10 level. 74 By looking at the means for those variables which tend to show some degree of significant difference we can Obtain considerable insight into the characteristics of individuals in each group. However, it should be mentioned that difference between means is not the only factor that contributes to discrimination between the two pOpulations. The degree of correlation of the variables between groups and the complimentary effect of some variables acting to— gether are also important factors but looking at the means should give us an adequate profile of each pOpulation. From Table 21 we can see that individuals who pur- chased their last major furniture item in a furniture store tend to be more fashion conscious, interested in social activities and relatively carefree in their shopping bee havior. ReSpondents who shopped in department stores appear to be more careful in spending their money, have less interest in community affairs and activities, and do not have as high a degree Of fashion consciousness. From this we can picture the typical furniture store shopper as the contemporary idea of the modern housewife, oriented towards entertaining and participating in community activities. She is more likely to buy things currently in vogue or that strike her fancy, with price primarily a secondary consideration. She is, however, conservative in 75 her shopping behavior preferring brand-name merchandise with which she is familiar or that has been recommended by her friends. The average department store shOpper appears more as the "plain" housewife. She is not very social oriented and takes very little active part in community activities. Her overall attitude is more self-oriented. She is the type who is apt to do a great deal of shOpping around before making a purchase. Price is more important to her and she is likely to pass up something she really likes for something that is just acceptable or to purchase unbranded merchandise and new products, if the price is right. 2) Style of furniture In question B-3 of the questionnaire, reSpondents were asked to select a diagram that best illustrates the styling features of their last major furniture purchase frOm several pages of drawings. This serves to classify the pur- chase into one of four different furniture style groupings. These groupings include almost all styles currently available, yet have quite distinctive features from one to the other. Table 22 lists these categories and the number of reSpondents assigned to each group. Contemporary is the most popular style category accounting for 45.4% of total respond- ents' purchases. Colonial/Early American accounts for the 76 next largest proportion of purchases or 33.1%. Provincial and Spanish/Mediterranean have the smallest frequency of purchase with 14.4 and 7.1% respectively. Table 22. Style of respondents' last major furniture purchase. Number of Percent of Style Categosy, Respondents Respondents Colonial/Early American 172 33.1 Provincial 75 14.4 Contemporary 236 45.4 Spanish/Mediterranean 37 7.1 Total 520 100.0 A four-way discriminant analysis has been run using these four categories as the dependent variable and the three combinations mentioned earlier in the chapter as independent variables. Table 23 shows the confusion matrix for the classification of 520 reSpondents by the six demographic variables in both raw and normalized form. For this case, the total number of hits is 170 or 32.7%. The primary reason for this low percentage of hits is that these variables are unable to differentiate individuals who purchased Colonial/ Early American from those in the other categories. We can see from the normalized matrix that the probability of an 77 oo.H me. mm. mm. mo. COOGMHHOUHUOZ . \amwcmom oo.H mm. me. am. «a. mumuooEoucoo oo.H -. en. es. Ha. Hososs>osa 00 H VN. hm. 5N. NH. CMOflMQEfl XHHMN \Howcoaoo Hmuoa coosouuouwooz muchomEoucoo Hoflocflooum coowuofi¢.>aumm Hoouo< II, \nmflcomm \Hmwcoaoo oooosoono Rename OONHHoEHoz An we.mm u when onoouom .oea u nuns Hobos omm mad mma «ma em Houoe hm ma. m m N :ooconnouwooz \nmflcoom omm mm Hoe om mm auouooeoucoo mm ha on on m Howocw>oum mud No no me an GMOHHOEE mHHmm \Hoflcoaou Hobos coocouuouflooz wuonoofioucoo Hmwocw>oum ascended manna finance \nmasomm \Hoflcoaou .rr oouosoono xfluumfi sumo 3mm Am .oomocounm OHSDHSHSH mo oa>un on ooumaou nuouomm oasomumOEoo Rem How Rename conomnoo ocuwaoauoc Odo sumo 3mm .mm manna 78 individual who purchased Colonial/Early American being classified as such is only .12 or less than the probability of her being assigned to any of the other groups. These variables are most clearly able to differentiate Spanish/ Mediterranean as almost half of these individuals are classified correctly. The fifteen life-style variables do a much better job Of discriminating among the four groups than the demographic factors. From Table 24 we can see that the number of hits is increased to 206 or 39.6% of the total. This is primarily due to the ability of these variables to properly classify purchasers of the Colonial/Early American style. The prob- ability of correctly assigning these individuals has been increased from .12 in the previous case to .40. Some de- crease is encountered in the assignment of the Contemporary group butthe Provincial and Spanish/Mediterranean groupings improved slightly. The best discrimination is obtained if the two sets of variables are combined and separation is based on all twenty-one factors. The total number of hits is increased to 217 out of 520 or 41.7% (Table 25). The probability of the individual being assigned to the proper grouP ranges from a low of .37 for Colonial/Early American to a high of .46 for Spanish/Mediterranean. This range is much narrower than for either of the other two cases so the probability Of 79 csocsunouwooz \aonaoam musuooeousou aswocw>oum csoeuse< mausm \stcoaoo ensues xwuuse oonwaseuoz an oo.n am. as. es. «s. .H .mm. en. es” em” ww.s mm. am. as ma. oo.H om. mm. . as. oe asuoe :socsuuouwooz ausuomEoucoo Asaocfi>oum escapees mausm \nnsosom \Hoesoaoo oooonooso no.mm n one: osoouoo .oom u omm and one see and em Ir as s o m omm am mm mm on me as es an on «an on an on so Hobos csocsuuonflooz ausuOQEODQOU Hsfiocfl>oum csowusem Mausm \nnsssom \aseaoaoo nose Hobos Hence csocsuuouwooz \nossoam musuooesucou assocw>oum nsowuoa< hausm \stcoHou Henson oononeoso .oomsaouoo onsuwcuom Ho Tampa 0» oousaou muouosm samum rowed cooumwm How xfiuusfi scamsmcoo ooueasEHo: ocs suso ssm anuse suso 3mm as .¢N OHQMB 80 we. on. «N. we. csocsuusufiosz so A \oossoom oo.H ma. me. ma. NN. hasHOQEoucoo oo.a ma. mm. mo. ma. assocw>oum oo.H ma. mm. mm. mm. escapees mansm \Hossoaoo asuoe csscsuuoueooz SHsHooEoucou stocfl>oum csOwHOE< mausm assuos \Smflcsmm \HsflcoHou I oonosooso LI xwuusfi osNfiHsEHoz An wh.Hv u was: ucoonom .ham A one: asuoa omm mod mod maa and asuoe hm SH o m o csocsuuouflooz \nnsseon emu es mos mm mm ssosooswusoo me «A SH um oH asfloefi>onm «SA on ow mm «o csOfluofid hausm \Heesoaoo asuoa csocsuusuflooz husuooeoucoo Hsflocfi>oum :soflnofi< wausm asnuos \sneooon \Hnesoeoo oooosoeso anuse suso ssm As .oomszousm ououflcunm mo Tampa 0» oobsaou mHObosm Oecosnmoeoo ocs OHSHMIOMfiH scormucssu How xwuusfi conSMSOO osansEHoc ocs suso 3sm .mm mansa 81 an individual being prOperly classified in each of the four groups is approximately equal. To give us some idea of the groups that are most dis- tinct from each other we can use association diagrams for each of the variable sets. Figures 7, 8 and 9 illustrate the direction of most likely misclassification for each set. Each figure has two diagrams, the upper showing the direction of misclassification for the largest off-diagonal entry in each row of Tables 23, 24 and 25 and the lower showing the direction for the two largest Off-diagonal entries in each table. From Figure 7 we can see that the groups most clearly differentiated by the demographic factors are Colonial/Early American and Spanish/Mediterranean. This means that indi- viduals who purchased Colonial/Early American furniture are least likely to be misclassified or confused with those who purchased Spanish/Mediterranean and inversely. Those most likely to be confused are purchasers of Contemporary and Spanish/Mediterranean as indicated by the reciprocal arrows in the largest misclassification diagram. Purchasers of Colonial/Early American and Provincial are also most likely to be confused with Contemporary but the reverse is not true to the same degree. Figure 8 illustrates the situation when discriminating by means Of the life-style factors. Here the clearest 82 Colonial/Early American Provincial II - ‘! SpanlSh/Mediterranean Contemporary a) Largest Misclassification Colonial/Ear 1y American Proyincial I » ) Cbntemporary . .‘r_ SpanlSh/Mediterranean b) Two Largest Misclassifications Figure 7. Association diagram for six demographic factors related to style of furniture purchased. 83 Colonial/Early American Provincial Se.- SpaniSh/Meditgrranean Contemporary a) Largest Misclassification Colonial/Early American Provincial I -v Spanish/ I Mediterranean Contemporary b) Two Largest Misclassifications Figure 8. Association diagram for fifteen life-style factors related to style of furniture purchased. 84 separation is between Colonial/Early American and Provincial but Spanish/Mediterranean and Contemporary are still the groups most easily confused with each other. All other relationships are reciprocal except that Colonial/Early American purchasers tend to be disprOportionately associated with Spanish/Mediterranean and Provincial with Contemporary. The associatiOn diagram for the combined set of vari- ables is shown in Figure 9. As was the case for the demo- graphic factors alone, the clearest separation is between the Colonial/Early American and Spanish/Mediterranean style groupings. The same reciprocal relationships exist as with the discrimination based upon life-style factors but purchasers of Provincial styled furniture tend to be associated with Con- temporary even though Colonial/Early American purchasers are more likely to be misclassified as Provincial buyers than those preferring Contemporary. Table 26 enables us to Obtain some insight into the characteristics of individuals in each group by analyzing differences between their mean scores. For variables one to six or the demographic variables, high mean scores indicate a_9reater degree of association with the variable. For example, the score of 3.80 on education of household head for respondents purchasing Provincial styled furniture indicates that they or their spouses tend to have received more formal 85 Colonial/Early American Provincial i SpaniSh/Mediterranean Contemporary a) Largest Misclassification Colonial/Early American Provincial 4 I I I o I I . .._..._...___..._...._ I h ‘ ' Contemporary S ' . panis [Mediterranean b) Two Largest Misclassifications Figure 9. Association diagram for twenty-one life- style and demographic factors related to style of furniture purchased. *The pair of dotted lines denotes a tie. 86 .Ho>oa as. on ososoemso maucsoflmwcowm mus wcsozs em. I mm. a um. I mm. 1 so.s ss.a sm.a oe.H ooswwmwwmmwon ”Hm mo.~ mm.~ sH.~ ss.~ o . . on em ucs HOHuHsQ muHOQw . em.a mo.~ hm.H hm.a no on o o o o aflomgoo unmflmz omH mm. vo. mo. mu Owuomuscm .ha mm H: mm at om Hr om.HI Hexagon one o . .H no.H mo.a H . . o s we NH. o o OoH UOHCOHHO UHHSU .mH mo I oo. I mo. I oo I summons o>flus>uomqoo .ea mo.~ vm H AH N oh.H stsEsEon «HomHSOMIudIOQ .mH @50N vmoN FmoN QMoN *Houmuowam muzwomm oNH MO.N mdefl hmoH mo.N mHHM 05“ NO COHHMHUOHQQ‘ .HH mmoHl h$ofil mmofll omoHI WHHMHHM huHsnEEoo ca umououcflmflo .oH mmoH omefl NmoH QFoH Hmmmogm HfimeMU om mm. I no. I we. I Ah. I Hommoxomdon Room .m @¢.H Ghofl MQoH mwofl CWQOHUWCOO GOHSNMW oh NMofi m®.m mdov when {OEOOGfl MHHEMN HMHOB cw «H.H mo.H mo.H mo.H oeoc boomono boou\cso .m mv.m om.m om.m wm.m osoa oaonomooa mo cofiusooom .v vo.~ oo.m om.m mo.m osoz oHozomoon Mo was .m oH.N oo.~ em.H mo.m swam wawesm .N mo.~ ~o.~ NH.~ mo.~ nsuoun Honest: .H csmcsunouwooz awsnomflsucou aswocw>oum csofinoam mansflus> \nmfiasmm mausm \Hossonoo .moansans> oaaumIOMflH oss OenmsanEoo scorhucssu ocs moauomobso samba onsuflcuom soon How mesoE o>flusnsoeoo mo oabse .om oabse 87 education than reSpondents who prefer either of the other style categories. variables seven to twenty-one or the life style variables, on the other hand, are interpreted so that lower absolute values for the mean scores indicate a high degree Of association with the variable, as previously explained when discussing Table 21. An analysis of the differences between means in Table 26 produces the following group profiles: Colonial/Early American Respondents who purchase furniture in this category tend to have a relatively low total family income averaging from $8000 to $10,000. For the most part, they are not overly fashion conscious, tend to be poor housekeepers and exhibit very price conscious shopping behavior. Their main interests are their children and Sports with very little enthusiasm shown for the arts, community affairs or social events and activities. Provincial These reSpondents are, on the average, older, better educated and tend to have fewer children presently living at home than individuals preferring either of the other style categories. They appear to exhibit a definite interest in the arts but very little for do-it-yourself projects around the house. Entertaining and other social activities are an 88 important part Of their way of life so they Show some concern about their appearance, particularly their weight. Contemporary Respondents who purchased Contemporary styled furni- ture are very closely associated with those who prefer Colonial/Early American in that they tend to have a low average total family income and Show very little interest in being considered fashionable. They also seem to indicate very little interest in community affairs, the arts, sports or social activities. In fact, this group exhibits no clear distinction on any Of these factors. In general, they appear to be very average, expressing no strong desires and content to go along with the crowd while purchasing furniture primarily for its functional rather than its aesthetic value. Spanish/Mediterranean Purchasers of Spanish/Mediterranean furniture,in the main, are younger, generally under 35, more energetic and more affluent than individuals in any of the other categories. Fashion appears to be extremely important to them and they engage in diet programs and active sports to maintain their physical appearance. Their interests include the arts, social activities and the community in which they live. They appear to be quite outgoing and, in general, portray an image of what is commonly termed "young moderns." 89 This analysis Of group differences can be made as extensive as desired by the researcher. In this case with only four groups it is relatively easy to identify the major group characteristics for each category. If more groups are involved the association diagrams of Figure 9 can be used to pick out the pairs of groups connected by fewest arrows. These groups are the most widely separated and pair-wise comparison between them gives a good profile of group characteristics while reducing the total number of compari— sons to a manageable size. To illustrate this procedure, our data in Figure 9 indicate that the most widely separated groups are as follows: a) Colonial/Early American compared with Spanish/ Mediterranean. b) Colonial/Early American compared with Provincial. 0) Provincial compared with Contemporary. The means for each of these combinations can be com- pared to Obtain group profiles similar to those we have already presented for each of the four furniture style categories. CHAPTER VI SUMMARY AND CONCLUSIONS This study has attempted to provide some insight into the characteristics of consumers of household furniture and to see how well these characteristics can be used to predict their market behavior. Rather than restricting the analysis to standard demOgraphic variables such as age, income, etc., another dimension has been added, that of life-style factors which define the consumers' interests, Opinions and attitudes concerning the way they live. These factors are derived by subjecting a series of eighty questions from a mail questionnaire to factor analysis. This procedure has identified a set of fifteen well-defined factors which cover a range of areas related to everyday family life. To test the ability of these variables to predict con- sumer market behavior, we have attempted to discriminate amonggroups of individuals who have demonstrated certain behavior in regards to: l) The type Of retail outlet at which their last major furniture purchase was made. 2) The styling characteristics of this last major purchase. 90 91 In each case, three runs have been made using different types and combinations of variables to see which set gives the best separation among groups. These sets consist of: 1) A series of six demographic variables. 2) The set of fifteen life-style variables resulting from the factor analysis. 3) The combination of twenty-one variables. For the first question relating to market behavior, the set Of fifteen life-style variables gives the best separation between the groups. These factors are able to correctly classify over 63 percent of the reSpondents who made their last purchase at either a department store or a furniture store. This is opposed to 56 percent and 60 per- cent for the other two sets of variables. For the question relating to the style of their par- ticular purchase, respondents indicated whether the item can be classified as Colonial/Early American, Provincial, Con- temporary Or Spanish/Mediterranean. In this case the Combined set of twenty-one variables gives the best discrimination, correctly classifying almost 42 percent Of the reapondents. This is opposed to 33 percent for the demographic variables alone and 40 percent for the life-style variables alone. From these results we can conclude that 1) Consumers do live according to certain patterns of behavior that can be measured and identified. 92 2) These behavior patterns or life-style factors have _greater significance than the demographic variables in being able to predict market behavior from a practical standpoint. The factors identified and used in this study are by no means inclusive of all aspects of consumer life-styles. Many other components of living behavior can be developed with other sets of questions and many of these may exhibit _greater predictive value than those used here. A long term trial and error procedure would be required for develOpment of the Optimum set but the results of this study indicate that this may be a worthwhile direction for industrial market research to take to get a more complete picture of the people who purchase and use their products. Although the results of the discriminatory analysis indicate that the life-style factors have greater predictive ability than the demographic variables, they still lack a great deal from being able to perfectly discriminate among the various market segments. Developing other, stronger variables is one way of improving the overall perfonmance but there are indications that stratifying the population by age, income level or other means will also give a better dis- crimination. In the pretest conducted with wives of staff and faculty members within the Natural Resources Building of Michigan State University, the life-style factors alone 93 correctly classified over 80 percent of the respondents as to the style category of their last major purchase. Admittedly this sample size was much smaller but an improvement of over 40 percent in predictive ability can be considered quite significant. The explanatory capabilities of the variables certainly appear to be much greater when the pOpulation is relatively homogeneous in income and occupation but differing in their attitudes, Opinions and preferences for particular styles Of furniture. This study represents only one appraoch to the market- ing problems of the furniture industry. With this type of information about their customers, they can promote and direct their products towards particular market segments. This will enable them to push their products through the distribution channels more efficiently and effectively. Another, and perhaps more important, aspect of the furniture industry's marketing problems is to overcome the traditionist attitude of many of the larger manufacturers insofar as styling or design is concerned. They produce what they think the public should have or variations on styles that have been pOpular for decades and even centuries. O'Hanlon (29) quotes the chairman of a large company as stating: There is really nothing new in furniture design. It's like a stack of sheet music. You play your way through and then start all over again. 94 This statement has been very true for many years. The extent that this philosoPhy has permeated the industry is illustrated by a quote from the same article by an execu- tive of the Kroehler Company: "If someone came to me with an original design, no matter how beautiful, I'd turn him down." To support their point, these peOple point to the recent success of "instant antiques“ or furniture that is artificially distressed in the factory to give it the appear- ance of grandmother's favorite table (29). To date this philosophy has worked as the public has purchased their product and many companies have prospered. Perhaps this has been because furniture is a necessity to every family and many consumers have been forced to make a selection from the limited lines available. Now with con- sumer disposable income expanding beyond the necessity level, the failure of the industry to keep pace with other consumer goods industries indicates its failure to satisfy consumer tastes and needs. The fact that today's consumers are ready to accept completely new designs not linked to tradition is exemplified by the success of Herman Miller Inc. of Zeeland, Michigan. This is a company that has expanded by ignoring traditional tastes and designing their products to meet a human need. Another indication of this is the rapid expansion of plastics 95 into furniture construction, where tremendous growth is pre- dicted not only in applications to simulate traditional wood panels but in the use of plastics as plastics. Research in this area to provide designs more accept- able to today's modern consumer will create a greater desire on her part to buy more furniture. The combination of the market pull developed by pro- viding products more desirable to the consumer and the push created by manufacturers directing their products and appeals to appropriate market segments should stimulate the demand for household furniture considerably and enable the industry to maintain or increase its present share of consumer dis- posable income. LITERATURE CITED 1) 2) 3) 4) 5) 6) 7) 3) 9) 10) ll) LITERATURE CITED Anderson, T. W., 1958, Introduction to Multivariate Statistical Analysis, John Wiley and Sons, Inc., New York, N. Y. Anon., 1969, Plastics make themselves at home, Business Week, Feb. 8, pp. 50-52. Arthur D. Little, Inc., 1967, The Im act of Industr Activities on Consumer Purchasingof Home Furnishings, Cambridge, Mass. Baker, L., 1968, Looking backward and forward, Furniture World, May, pp. 25, 38-39. Product, brand and reference-group influence, pp. 351-353 in S. H. Britt (ed.), Con- sumer Behavior and the Behavioral Sciences, Jo n WiIey and Sons, Inc., New York, N. Y. Boyd, H. W., Jr., and Westfall, R., 1964, Marketin Research: Text and Cases, Richard D. Trwin, Inc., HomewoodiéfII. Bourne, F. S., 1966, R. E., 1968, Past tense and future progress, Burow, 32 -33. The Twin Citprurniture Digest, Jan., pp. Chicago Tribune, 1959, The Study of Furniture, Chicago, Ill. Collins, 6., 1963, Factor analysis: how it's done, pp. 242-249 in A. Shuchman (ed.), Scientific Decision Making in Business, Holt, Rinehart and Winston, Inc., New York, N. Y. Cooley, W. w. and Lohnes, P. R., 1962, Multivariate Pro— cedures for the Behavioral Sciences, John Wiley and ons, Inc., ew Yor , N. Y. Cox, W. E., Jr., 1966, Response patterns to mail surveys, Journal of Marketing ReSearch, Nov., pp. 392-397. 96 12) 13) 14) 15) 16) 17) l8) 19) 20) 21) 22) 23) 24) 97 Evans, F. B., 1959, Psychological and Objective Factors in the Prediction of Brand Choice — Ford vs. Chevrolet, Journal of Business, vol. 32, no. 4, pp. 340-369. Finch, T. A., Jr., 1967, The marketing concept; a maturing industry, Furniture WOrld, Aug., pp. 21-45. Ford, N. M., 1967, The advance letter in mail surveys, Journal of Marketing Research, May, pp. 202-204. The Furniture Industgyand Its Forman, J. B., 1950, U. S. Department of Commerce, Potential Market, Wishington, D. E. Gordon, C. 0., 1968, Furniture's future, The Twin City Furniture Digest, May, pp. 12-13. Harman, H. H., 1967, Modern Factor Analysis, University of Chicago Press, Chicago, T112 Harper, R., 1962, Factor analysis as a technique for examining complex data on foodstuffs, pp. 410-416 in R. E. Frank, A. A. Kuehn and W. F. Massey (eds.), Quantitative Techniques in Marketing Analysis, Richard'D. IrWin, Inc., Homewoodi’Ill. Henrysson, S., 1957, Applicability of Factor Analysis in the Behavioral Sciences, Almquist and WikselI, Stockho1m. Kendall, M. G., 1957, A Course in Multivariate Analysis, Griffin, London. Kroeger, A. and Bing, L. S., 1963, Western Home Furnish- ings: A Market Study in Two Parts, Sunset Special Report. Lazer, W., 1964, Life style concepts and marketing, pp. 139-139 in S. A. Greyser (ed.), Toward Scientific Marketing, American Marketing Association, icago, III. Leary, T. D., 1967, New forces seen changing shape of furniture industry, New England Furniture News, Nov., P. 0 Lucas, D. B. and Britt, S. R., 1950, Advertisin Ps — chol ‘ 'and'Research, McGraw-Hill Book Company, Inc., New York, N. Y. 25) 26) 27) 28) 29) 30) 31) 32) 33) 34) 35) 36) 37) 98 Marketing Insights, 1969, Demographic study reveals differences in buying patterns, vol. 3, no. 9, p. 6. Massey, W. F., 1965, Discriminant analysis of audience characteristics,'Journa1 of Advertising Research, vol. 5, no. 1, pp. 39-48. Mukherjee, D. N., 1965, A factor analysis of some qualitative attributes of coffee, Journal of Adver- tising Research vol. 5, no. 1, pp.235- 38. National Family Opinion, Inc., 1967, A Home Furnishings Consumer Study: Buyers and Intended Buyers, To e o, Uhio. O'Hanlon, T., 1967, 5,350 companies = a mixed up furni- ture industry, Fertune, vol. LXXV, Feb., pp. 145-149, 178-182. Oppenheim, A. N., 1966, Questionnaire Design and Atti- tude Measurement, Bas1c BoOks, New York? N. Y. Pessemier, E. A., Teach R., and Tigert, D. J., 1965, The Consumer Behavior Research Project, Krannert Gfaduate School chIndustrial Administration, Purdue University, Lafayette, Ind. Ramond, C. R., 1963, Factor analysis, when to use it, pp. 235-242 in A. Shuchman (ed.), Scientific Decision Makingin Business, Holt, Rinehart and Winston, Inc., New York, N. Y. Rao, C. R., 1952, Advanced Statistical Methods in Bio- metric Research, John Wiley and Sons, Inc., New York, N. Y. Rothberg, M. F., 1968, We must become consumer oriented, Furniture World, April, p. 27. Schulte, H. J., 1968, St 1e Trends and the Purchase Decision, National Retail Furniture Association ‘ Reports, April. Shaughnessy, C. S., 1968, Looking ahead, Furniture world, June, pp. 25, 78- 79. Social Research, Inc. , 1967, inc Homemakers and Home Furnishi_gs, Chicago, Ill. 38) 39) 40) 41) 42) 43) 44) 45) 99 Social Research, Inc., 1958, The Kroehler Report, A Motivation Study, The Kroehler Manufacturing Company, Napervillef I . Spector, A. J., 1961, Basic dimensions of the corporate image, Journal of Marketing, Oct., pp. 47-51. State Journal, 1968, Facts and Figures on the Greater Lansing Trading Area, Lansing, Mich. Stoetzel, J., 1960, A factor analysis of the liquor preferences of French consumers, Journal of Advertising Research, Dec., pp. 7-11. Twedt, D. W., A multiple factor analysis of advertising readership, Journal of Applied Psychology, vol. 36, no. 3, pp. 2 - . Udell, J. G., 1965, Can attitude measurement predict consumer behavior?, Journal of Marketing, Oct., pp. 46-500 Weilbacher, W. M., 1967, Standard classification of con- sumer characteristics, Journal of Marketing, Jan., pp. 27-31. Wilson, C. L., 1966, Homemaker LivingPatterns and Marketplace Behavior, Harvard BuSiness School, Division of Research, Boston, Mass. APPENDICES I.I ilulllllll All-I'll“ I APPENDIX I The questionnaire used to obtain information on consumer life-styles and market behavior. MICHIGAN STATE UNIVERSITY m ammo - mallow 48823 DDAITKBN'I' OF INDUSTRY February 24, 1969 Dear Homemaker: I hope you will help us with a study we are doing at Michigan State Univer- sity. We are asking a representative sample of residents of Michigan to send us in- formation regarding themselves and their purchases and ownership of household In this way, we hope to relate behavior regarding household fur- furniture. All information niture to the_living patterns of various groups of people. _ . . that you send to us will be kept strictly confidential and no part1c1pat1ng individuals will ever be identified. This study is being made by Michigan State as part of its research program. It is not sponsored by any business firm or any political organization. You will never be asked to buy anything as a result of this study, nor will you be asked to make any contribution of any kind. I hope that you will fill out the questionnaire and send it back to us as soon as conveniently possible in the self-addressed envelope which is en- closed. You are very important to us in this study. Something else which may be of interest to you is the work being done by the Cooperative Extension Service of Michigan State University. They publish a number of informative publications, most of which are available at no cost. The enclosed booklet has a full listing of these reports, and you may use the order blank at the back to obtain those of interest to you. I particularly recommend "The Selection of Upholstered Furniture" and "Wood Furniture" listed under the Home Furnishings Section. If you have any questions regarding the survey, please call me at the Univer- sity, 353—cane, and I shall be glad to answer them. Thanks very much. Walter S. Good, Graduate Research Assistant Department of Forestry - Wood Science 100 Dear Homemaker: . 4‘ o The following is a series of questions relating to living patterns and household furnituZZ§P::;:gi ences. Please mark the answer to each question clearly with either a (V’) or an (X) in uestion priate box at the right hand side of the page. Also be certain that only one answer per qthe are is checked and that all questions are answered. Ignore the numbers beside each square as y used only for coding the final information. A. 1. Present marital Status: Single 'E31 (6) Married C32 Widowed [J3 Divorced [2])4 Other [35 2. Family size: (presently living at home) 1 or 2 members [31 (7) 3 or u members [32 5 or more members 03 3. Age of household head: 21+ and younger 01 (8) 25 to 3k E]2 35 to #9 E33 50 to 61+ [1“ 65 and older [35 (9) Some high school Graduate high school Some college #. Educatibn of household head: Grade school or less E3 Graduate college [3 \J'IF'UJI'UP" 5. Own present home E] (l0) 1 Rent present home [12 6. Total family income — 1968: Under $5,000 Ill (ll) $5,000 - 7,999 [32 $8,000 ' 93999 DB $102000 ' 114,999 on $15,000 - 19,999 C15 Over $20,000 E36 B. The questions in this section are concerned with the last major (over $50.00) furniture item or set that you purchased for your home. If more than one item or set was purchased at the same time Please relate your answers to only one of them. ’ 1. Which of the following items or sets was this particular purchase: Living room‘set E]; (12) Sofa or divan [3 Lounge or occasional chair C13 Dinette-kitchen set h End or coffee tables [3g Bedroom set 0 Chest, dresser, etc. E37 Other [38 2. At what type of store or outlet was this purchase made: Furniture department of department storefljl- (13) Discount store C32 Furniture store EEB Mail order house h Interior design shop :35 1(31 3. Of the attached diagrams, which page has a drawing which best represents t 1+. 5. C. The following questions may appear to be very unrelated to ho blishing living patterns. Please try to answer certain that all questions are answered. with the statement orlumwwellyou feel it applies to you as indicated by tial for eSta l. 2. Wholesale outleth6 Other [37 of this item or set: Page I, Colonial-Early Americanfjl- (1h) Page II, Provincial 2 Page III, Contemporary [33 Page IV, Spanish-Mediterranean [:124 Other [3 Please explain: Approximately how long ago was this purchase made: Within last 12 monthsfijl (15) l - 2 years [32 2 - 5 years [33 Over 5 years CD“ If you were considering the purchase of this item or set again, styling features that you would select now: Page I, Colonial-Early American 1 Page II, Provincial 2 Page III, Contemporary 2 Page IV, Spanish-Mediterranean Other [3 Please explain: (16) Agree or applies Uncertain \J'IJTUJNH I often redecorate my house or apartment. I really enjoy most forms of housework. I go to the women's club, church ladies' group, or some other women's group which meets regularly. When I find a coupon in the paper, I clip it and redeem it the next time I go shopping. People should spend a lot of time with their children talking about their activities, friends and problems. I enjoy going to concerts. I definitely watch what I eat to keep my weight down. I often wonder where others get all the energy they seem to have. Indicate by your respons Disagree or does not apply I Strongly disagree or definitely does not app y Agree [31 E31 [31 E31 [31 C31 C31 C31. C32 C32 [32 C32 [32 E32 [32 he s tyling features which page best represents '3, usehold furniture, but them as honestly as pos '1 e the extent of your agreeme» the scale below: . Strongly agree or definitely applies Uncertain D3 [33 C33 [33 ' aree ,4 On D1883 1* '~ [3h C35 LW [3“ [35 “J they areas?” Sible am} 113:: I .. oooo Agree 9. I try to buy things that represent good value I for my money. [3 10. I have a good deal of respect for tradition. [31- 11. Looking after children really demands too much 1 of me. C] 12. One of my favorite community activities is work- ing with boys and girls in Scouting or other 1 group activities. D 13. I was active in sports when I was in school. E31 1h. I nearly always wear nail polish. [31' 15. I enjoy spending leisure time in museums. [31' 16. As a rule, I don't buy new products until I hear something about them from people who have tried them. [31 1.7. I shop for specials. [:11 18. by husband compliments me on the way I run the 1 house. [3 19. I enjoy trying the latest style in hair-dOYS. C31- 20. When children are ill in bed, parents should drop most everything else in order to see to their comfort. [31' ?l. I have gone on a strict diet to control my 1 weight one or more times. C] 22. I have helped to collect money for the Red Cross, 1 United Fund or March of Dimes. [j 23. I dislike any changes or interference with established ways of doing things. [31 2h. Parents should regularly visit their children's school and talk wit. their teacher. [31‘ 25. I enjoy making some of my own clothes. [31' 26. I am able to work for long periods of time 1 without feeling tired. E] 27. I go out to lunch with my friends quite often. E31 28. I take a keen interest in politics. ‘ E31" 29. I hate to throw things away even though they 1 are of little use anymore. E] 30. Parents should take a lot of time and effort to I teach their children good habits. C3 31. I read the sports section of the paper. C31 C31 32. I enjoy fixing up and repainting old things. 1()2 Uncertain D2 D3 [12 D3 C12 D3 [32 DB D2 D3 DZ D3 [12 [33 D2 D3 [32 E13 E32 E13 [32 C33 C12 [:13 C32 E33 C39 C33 C32 [13 C32 [33 E32 [33 C32 [33 E32 E13 [32 E13 C32 [33 E32 [33 E32 [33 E32 [33 [31: Du Db Disagree Db. DS (25) an (35 (26) Du D5 (27) [35 (28) [35 (29) [35 (30) E15 (31) [35 (32) E15 (33) US (31*) E35 (35) [35 (36) [35 (37) [35 (38) E35 (39) [35 (ho) [)5 (hi) [35 (#2) C35 (#3) [35 (ha) [15 (us) [35 (#6) C35 (#7) E35 (#8) -h- Agree Uncertain Disagree 33. I take advantage of low calorie foods to help 3 h [35 fl; me and/or my family keep our weight down. D1 D2 D D u [15 (33‘; 3h. I often buy products or brands just on impluse. E31 [32 [33 C] 35. I use eye shadow or eye liner three times a 1, 5 (517 week or more. D1 D2 D3 D C] ’ 36. I used to bowl, play tennis, golf, or engage h 5 (52‘) in other active sports quite often. [31 C32 [33 C3 C3 37. I study the food ads each week so I can u 5 (fl) make the best buy. [:11 [:12 D3 D D 38. I usually have at least one outfit that is the h 5 (it? very latest style. Bl [:12 D3 D D )4 [35 (55? 39. I like to work in community projects. [11 [32 D3 U (A h [35 So #0. I really enjoy cleaning my house. [31 C32 [33 [3 hi. Fashion in clothes is more important than h 5 GI comfort to me. DJ" D2 D3 D D h2. Once I have made a choice on brands, I am likely 2 E13 E3“ [35 (i) to use it regularly without trying any others. C31 C] “A h [35 W h3. I have a lot of energy.‘ [31' C32 C13 [3 1. ‘ S . (6C1 uh. 12:22:82? my elf very good at sewing and/or D1 D2 D3 all 05 AS. I generally prefer classical to the more pOpular h 5 (&l forms of music. D1 [32 DB D D M 1+6 Parents should try to arrange their home for their 1; Us (4; ’ 2 children's convenience. 81‘ D D3 D . . . . . . 5 (63' h7. g:aITSlt with friends in their homes a great [11’ [12 C33 [3“ [j h8. I often find myself being critical of the l 2 [33 [3h [35 (W) \ younger generation. [3 C3 6) M9. I feel embarrassed and uncomfortable when I l 2 [33 [Bu [35 (5 am asked to entertain strangers. C3 [3 . {a 0. Children brin a husband and wif closer to h [35 5 each other. 8 e [11' C32 E33 E3 ”) 51. :ewgzgerally 80 out for dinner at least once [31' [12 [33 Elk [35 (l o 5 (UV .3 u D 52. I enjoy listening to classical records. [31- C32 E3 E3 53. I get great satisfaction from experimenting l 2 [33 [3h [35 ( with new recipes. [3 E3 (. 5h. I like to watch or listen to baseball or 1 2 [33 [3“ C35 football games. U D 55. I have several different shades of lipstick to go with different dresses. 5 D1, D2 [33 U“ D / - r e a a n at ‘ C v (:1 t)alllzti Jion. . v r h [K n ‘ Se 1 C . : &LA 0 As n f g; 1,. I 'I 5 a ‘ . , yr 1. " ,,’ r l nu . I l U]. 1n . L H‘ pr / I r n C2} I . / \ Q 1 Se ‘ l v. {11‘ 1 . 7‘. + a . ‘ h - 'AAe A .. l ‘ .‘4 GI 3.; l, i >‘ l ‘ «S /7 O AnllLrLCan ‘ K 3 £1 a: V-L V \— E)ar U 1‘1 COAAUHUXLJ. JV 8 Q : area 1 \; I‘Il {la 1 s ..-L A O we _. "‘ }.lnk4n5 d O t I .~ 7‘ ‘ ’ v $- 5 1 — r '- ‘ I I e ‘1 (x J V 1 (I . é 8“ ‘r . 0 Ii 1“ ‘.I \, I ‘ G l y s I v “ A ,1 0 'fl; r I". .L 1‘9 .el v ery une'“ ' asy wnen othe or peo ples' m ’f, a I t‘ . nlnk ‘ .n ma' now r . my you r L. h ' n“ I o cring u ? Pdrents t 70 T ‘ p Children pr Oflay d0 “0+ k A. .‘la‘fe D C)perl U new ‘.- 4L ’3rh ‘ y. lor" sonalr a Cfilrhiirio'r J worked - ‘ v x)... art {so - ic . L con"; 3 Que camoair °$Ger 1w . 1 bn a”) Wbell‘ r 1" ‘. . h” I RD e_aclvely ceo a‘ ‘v set 4 ‘ wav f In m ’fi . .ron V ‘ ‘y wa fin ween. i grands I've ys. ”h Aln" r . nev . I. e 3 needs 5 Ly home n. r heard of , ice sat‘sf' . ” $ A. _ r ‘ (A. To r; ies my creative c..”“9 ha‘ ‘ ~ ) vain h ll tne fu Do. n of v- shoppin"r ' r?. ‘7 g 13 tr ‘ “‘C’” raw - ylng ”ls; fir‘ I ‘I CD new or hospil‘Id work “or . +- h L P “f vaL on a a scrvi ,9. I s 1 regular ba 'Ce organiVat' ‘ e do” 818 s 'lon . b ' . 77 I uy things on ' . lm ‘ en+ V . pulse certain freq“ . - dently ' in my h iome L Cl HO O y I e. I lik e to or ' ' ganize community . project S. . . I (r0 9 bowling often [31 [31 [31 [31 [31 D1 Uncertain [32 E32 [32 [3 2 [3:2 [3:2 [3 2 [32 1k \ u pa I a _, 1 c p r O 'n. l()3 [33 [33 [33 [33 [33 [33 [33 [33 [33 [33 [33 [33 [33 [33 '2 [33 [33 [33 [33 [33 [33 [33 [j 3 Disagree [3h D“ [3h []M []h [3h [3% 01+ D n [35 [35 [35 (7) (8) (a) [35 (10) [35 (11) [35 (12) [35 (13) [35 (1h) [35 (15) [35 (16) [35 (17) (18) (19) [35 [35 [35 (20) (31) (22) [35 [35 [3‘3 (23) [35 (2h) [35 (95) [35 (a6) C35 (27) ,[35 (an) [35 (Pa) [35 (30) [35 (31) COLONIAL - EARLY AMERICAN PROVINCIAL my, ,_ |_| CONTEMPORARY gk' E Q 45” N If ”~11 - MEDITERRANEAN IV SPANISH APPENDIX II Frequency distribution of respondents' answers to parts A and B of the questionnaire. 4"" ‘D‘Au )u“. H J—V4‘4‘ A~ N4.ulu uto.I i... .s...r__.. 1.- “I .._..‘ _ --_ r '”7 1 r—-. I . I““7”' .——-I--T—" 7"”‘7"! —"_' '“T‘"’" - I-I I --- - - --- I ,.,I;I- f . -:I-iI!I:-IITI-III-III:II-IIII I-IIfI-:I-!-I- I .w ON“ omflod ’O‘uofl f. . >).U.J_-......L¢ . mzo_»<>xwmmo .o.m ¢zmmmo .=.m 24m: rmx»s x444: Izos -- - - --. - . - mn.n ma.om an.ov cm.~m ofl.n _:iuqu zqz I - ..I:.:- ..-:----II ---IIII-.1II -I-- II:-I-I--I IIIIIIIII=NIIIIIIc9fi-.IIInHoIIIIIsfiHIIIII~mII:II:II>nauzayim - - I, .II III III-- m .-. m «- w. 3.5- III--4- m c/ - - I - - III: III. I- -I: I l-.. .. -- II- ...-.I- II -3-.III-I---III.III -II- .-.III-.II.-IIIIIII.I.IIIIIII.II I--. III-IIIIIIIIIIQFIIII-III III-3n..-- ..-I . a: . “Zn: 7 own Nmn. voa.~ . . c ; >J;u::mri mzo_»<>xmmmo .o.m za.u:uwls mzouh<>zmmmo .a.m zuzusxsmri. .III;.IIII I P n b h IEI:II:IM1!¢aLu I. _ n\ - 3.-.C-V«U a .u ”NJ-(I ---I III. 33:: I - III ;:.I-II:- I I I - - II-I - - II:IIIII!IIIIIEIIIIIIIIIIIIIIIIIIIiIIIIIpa:IIIIt-.:.H ..cz ”4nd" - a .mm...:..(.a- I - ,- mean-42%. -33 335....-- :33.-3...m<-.s.c.--mr..._m-_:sw..-I.m...-I-me_-:.w....s.3n._s Gil-LI -. €F(.U 5")" II- “ - Q.‘ ( .~R.dA- .4 ‘I I? [I _... T-__r T . I . own awn.” Hmo.m c .. .J. u::...r.._ .u mzo—»1>zwmmo .=.m 2zmmmo .a.m z:wmmo .o.m 24w: nun-E»; x :3... 1.3... n.a.o an...“ ev.nv .K.ov mn.an “v.0 TLunI. xx: -- -- -I-I 1- I mw mu mwm um .nm n». Serif... IIIIII II. C IF .- n. W. .I .5 J .I .04/.5005. I I I. ...I III III I. ..I I. I I II I . III II. ..I - I. - I -I . II. I. .II.II.|...I.IIII.. I..I IIIIIIIIIII. I II-..|. 0g-.. . I .I. . :I..- I} N “ha” own :5. coat" c ; 3.5.1.1. mzo:<>zmmmo .:.m 25?. «his :53; 3:3 0:.0 «~33.» , guru... 2Jgu..o.I.x._.I III.III.II IIIIIIIII \ H .. £35.) . III I 01.33.”; wwiuwx. f) I. I -. I I II-.[- III . II III II-IIIIIIIII . II .II III I II . n; III.IIn.!. .21.. ”Anew I...” in .1319. ...I...--.I.- I: .0.on def-moms ..~...m.5..w..; INN-ImmulmuflflmnMam-PumaImm.o-s.p...I....i.rHa_.-,.-.-MENU-1.3..“ -I- - - - I‘m” ‘P(Ad 9‘}? unsila‘ ..-- 'I' .. YIIII -1--- I ..--‘ll. 0 own amn.d HMO.N c ; .J.u:3urI . II... .IIIIIIIIIIIIIIIII.IIIIII.I-. .II. I I. I; .I- I- I... IV“ WZO~P1>ImMDO oaom 2z¢nxo .a.m zJru;I-rq wzo_»«>xwmmo .c.m 24w: au1»; x.<4: :3.u n~.a wb.¢fi ov.nv mn.0w mm.an mv.v _:uJIUL :1: I I---IIIII- -.--III- III. III III-IIIII II -SVIII:III-ommiI-I; I am .3 II. , ...-23L... . I. c HI r WI 1“ V III. ...) .mCmOug- I - I .I --II - I I II I I- II - - I- II - I.-- -I - I- I III-IIIIIII-I -- -II -I -I-IIIIIIII-II.-I II--- :I-:o.~91- .-.: .I- .'~- ;. .,.z. N “nu.“ ONn NSN. QOO.H c c >J.v;JLr¢ mZo—h<>zmwmo .:.w Z);u;anu.;- I II I“ W III..—IJ . .. II! N/prwi ....JaoNK I.) I- I I III-II IIII- .IIIIlI-III IIIIIIIII III-II I-II II III III. I: I-IIIIIIIIIIIIIIIIIIIII. I- IHQIIIIII-IIIIIOII. 37’” unrucw « _wm ¢H¢o I . -. - ; .I- ., oooa Juzv< ooow.rm»4(x . mm-.m.owIfi< m:am_mu:o 20“»:ahx»n_a >uaI4:mI¢ ..n <-(- 3 ..\ \. f I—Nfinl‘ III-III III III." I-IIIII‘“ --‘lIOlat. -IIOII.III III-IIII'o’IuOIIIII'-.IIIIIIIICIOI0 I’ll 'Inl'li‘lo‘l"- II ' . TI - I III- -I-I I III I I ::II I-.. I I I-I- III-II II II II..IIII-II.II.-III|.-IIIII.I II- IIIIIIIII-I. III-I - - III- II - II .II I. I II I I I I II - - We - I -III I. . I-I-III-III ........ -I.III.:I-IIII-IIIIIIIIzIIIIIII IIII-IIIIIII I-I:IIII.:-:I-::;-II . I - II .- -I---I - -. I I I - I -II I -. I - --I--I-I-I-III I-IIIII-I I-IIIIIIIIIIII-I-III- I. I - I- - I I H own ”mo.“ aem.~ a ; .- 3.3-r- . ,fi wzouhqzlmmo .o.w 24m: rut-kc. .../:3: 3.3. on.an cv.xm :m.nr mx.¢c _:;31u; ?:: r.- I II I- - I .III. III I .. .II .I I I- I I II I...-IIIII-.-I ..- I- -.I- I..I-II- III! .III- III-III III-I-II-IIIIII30-II-.III-.. CD“ I-III ~..I- II.._II.~\ r. - .I . IIvJ.IU...M..I-._r-.. V.: I. ......... - .I- .II-.IIIII 2I.-III:.I-I:. I I m w II-gba;---- nu _ - l -- - I . -II- I - I - I I I I -. - II I I --- I - .I I II .. .IIIII.-I.-I.II I .IIIIIIIIIII-III- -.--IIII.:-II..I II I. I III IIIII IIIIIDZ-IIIIII I. I. I .Ir. r .. ....C U:..4.... a y l own mna. nfio.a c g .I “2,.x. “ wzoZkZlmmo .2.m 24w: 9?»: .244? I3 3 m n\.c anmo .=.m .«w: .-;». ..44: .239 W xd.~ an.m- ac... Io.q< ...uzu. ‘4. II - ..- -I. I- -I I I- - III- -I I- -II---II--- I I - -.IIII I-I-IIII: III; III IRWIIIIIKMO-I I .mNI-ImhfiI IIII -..x-DJATJHEITII _ VII. ..--III III. -.I. II c m h L. --.uncuI-I2III-m _ "a. I- II.I.-II. - I -I - I III I. -- - III-I- I I - I I. ...III...- I. -II- I..--III-.II III .III.IIIII-IIII.-IIIII.I-.-IIII I-MIu II-II. --. .0 - .Fa IU.I..I.<— M w- - .: - - ..- -- ... - - _- . . - .II; I , I ..-, -t .1 -II-_ .5 a H _mm u4IJ:u.‘ _VCI II I I (II. ltd- I! III I- '3! I I II III I III I III! III III-I! I .. III ..III II-I.IIIII.I I II'IIIIII. I III! II IIIIOIIIIIIIIIII-III III I II I I I APPENDIX III Correlation coefficients between the eighty variables in part C of the questionnaire. wmoo. ¢o~«.n la n vm awmn.m sovm.m emmn.n n~:n.m ‘ 0 \3 Na am mnun.n as ac woou.~ «men.» o: mama.» me “moo.~ oomh.~ . - - I un~¢.~ we. «coo.v “c mm«m.~ a. mono.» on Hake.“ on anno.~ an IIIIuI.II- .III-IIIIIIIIeoono.~IIIII on I. ocwo.nIII-. mm «nmm.n I- c» I. «ado.~ I.. an II.em«o.~..----~n nonm.n an VI.--- - II-.-.I.~oon.a I an I-uoaa.n.-I. o~ memo.“ - om..I-onnm,nII .: «a -I «nun.~ om s»mo.~ mu _I--- - I II--. --I «smo.u- II «N I, «oon.n- - mu ~ooe.~ -.II swno.n «u . oom~.~ om oosa.n on IIII-I.-Is-III .IIIIIIIIlznun.J~:IIII-ouIIIIoowntuIIII-Iha.- .occo.~-;III-o« II-~o«~.nII;.. ma, mnao.n. ca mafia.» na . . VII sz,- -IIIIIvovm.nI.IIII~«II; ovoa.c-- .-aa cama.~II I- o« .oo-aa-.-...o onoo.~ a Nocs.u a “II II -I- I .I-:I-I-IIII«»«on~II:IIIo,:eIIoano«a -I---» - I «0.0.».-- -Ie- IIIns««.n -I n ommn.~ ~ ance.» a w II.IIII-IIII:I.I:, -II I I .- I-II .: III-I.-- ,I I .m2¢ W - I- I III----.I..IIIIII:I.III- I I--- I-- III - - I --- II I -I - -- .o.«uo~.xmxo.auum.xca. m k.zaou «» .III-I-. --II - ...... IIIIIIIII. ....... I-II IIIIIII I moo“ ..4.maa coco «wusaz _. - coo.«u mzoutwwao zo mnm>4¢7¢ achu«u I- I- - I-II:IIII-II-- :III:IIII. - .IIIIII IIIIII.I: - -I I III- III-- . «aurora wz_»=o¢ «cu wc¢x 44 IMI. - --IIII I I::IIII :IIII- I I I II- -II III-I I III - -I-4 momueuamzupaoan .. ma‘ma nm¢u Joapzou cum” omopoflu .2! I! I anumm .I-"l'l.|-l ” monumcuu -uozmaz...«>oxraz.. (mm .nwsopt .II IIunzz- . on. -ou< zmooxm.. 2<3wzoI.. HuaII»z:ooama.. momoyz; «ohmah.o .wm¢ IIIIII:.nou.Fz.Ito¢. oma(04 m>.»:omxm mm<»-><4¢w>o cumin- II-IIII|‘|..I'I.III c0 . 9|.IIIOI mm

mc om; onwn.v on smnv.~ as voou.n so oown.» no NQON.N mm I. - -II I II - - I TI I III -.I l.- I I I II I II _II- I .. w-I . I- II - -- - - .fi - m III. I I- II- - I I I- I - I I - I - --I- I - f-I I- I -I--. I :-I- I I I - - -I -I- I - I - - I _ 3 a .1. I I III - I - I I I I I .- - - - 1.;IA.II .III III- II-I II: III-I I -I I II I “III-II III I- III I II II-IIIIII I I- I- I I I I I I - - II II- IIIII I cocooweoqorxmwxzk w:n.q>7wc~w, - I-IIII II-I-IIIIIIIIIIIIIIIIII-IIIIIIIII;I .. I .-I I..-III-,-;I- I -.:-:-..; . .. I o Izo_»a¢ p2,“: “ a“ .>mJu_¢=IJm.x hoz u~ mp‘»o¢ o» mac»uw4e~az.4w~x m u .zo~»<»oa go ma>» mI-I I - - III-I--:II- III- I I- I I- - I -I - I - -- - I - ..oI..zo..‘+oa mo m_x‘ 4p~4«2:xzou .I-III-IIIII III-II .I- .I.I.- ...-II. I I. ...-II- .. .-.II-.--I--I--IIIII.-I I. :-:I : .. .I_ ..:I o. .2236 SE: _ TI - .I -- I I I - I- ampumgwm mac”. m flz: .III-.I .IIIIIIIIIIIIII-IIII-I...I-I- I... ..- - I - -I...,---I :;I--;III..- I-I-.I.I III .mw4n<~a.> on «on mzo.».>.mmmc cN — mwo“.n q~.fl.c nmzc.o- . mmda.o «n.a.; fim°°.°r . , nmuo.o. , nona.c. ~nmn.o. :- --- ;-:-: sf1,-;s-nnH..o :-.°~°.m .. mmeo.° 1..¢.n.; :e:;;:¢szzo;lz‘z:s«no°.o ..... e~o«.¢--u ammq.. Hz; - - --;;;:-: e:!.n luv: uwuwu. ‘::.ov.e.c.:n .«m«.°.: wa- --- .2-ws:!:s:1:.iila-nnsuro.i‘sauvo.t.§, :quuu;:;!:w- . onWotarruewneo.co-asnwna.o. w!-;.4w;-:=- - --woo..o;-:;nnvo.c:.- wzz;l:;-l-- ;:::en-:.::s::--o~ac.e.;s:¢«o«.o.-- mu.;1:;;- . :a1.aq::z;s-uz1s.“(no.°;.:;d°~a.:,ga-“we“.o.-z w- -Al ;-.. ,; ;-a;:esr:;esa~mo.ozyninocn.c.s-."moa.o‘ w; ,- - 5:-a;:-:xr:,nova.o-a;.mo=o.c[:;.~svu.o.- W--::- .:,-:i ..‘1:::::::::mnme.o;:!.¢~oo.casaoa~o.o.l (I-.; - l -.:-:;::-;saeudm..o;l:1oceo.c.- .o'vo.o :nug‘-., -:-;:-:a,-n::zj;~noo.aq:,n¢mo.c,- éflFlIE,-.:Iik‘ilifiiii::||3-6. :.-: e, : o~nu.° ocfio.e‘ m¢oo.or mr,;eii- . - .: 1,;. -:-;;i oom..or-l-.mco.@. oau~.o W:.- I .- .; :;;i: 3-.:-Titdocofiui-: mwnu.c-;;eoowa.o wuiri--:;.:,:‘9a;z:.:-4:-;:.:z;!;;;:eauco.w i:.eoao.on- . ....... .:,-::.. - g -- -- a- 2.....; r- :; coco.“ - T: - :--:;,:--:;:ssi:- -.; I aA -- w.- ------ :-: .1- -1 --i- -s-;-:;‘ ,1 - fl-- , - - :,--:,-a a - -: ,«-- r g e w-:- - I --:.::g---,- m .s:. .1---.:-: -4 :.- r‘ - - ‘ :I- ...... -----Pu. o - c eomfi.o nqvo.o cmno.o oeso.o {~«o°.. -nnh°.o .jlno.o.° omen.o t.womo.o : :oooo.or. ‘-oono.oy;; smua.a,_ csmo.o. onwo.o- vvoo.oc nond.o ,onoc.o .{oono.°.r_:mo¢o.o homo.o omflo.o. «oo°.o: -;von«.o .oo~°.or .ma..o; mao°no comonoo cvnono :~n.°.o q- coconut oooano : oo~°w°.,:.noflo.° enoonu. nhoouo.i.gomaauqe omao”°.; -nmno.o- ‘sowo.o .o~»o~o.- .'-lll \ m~m°»o coco“; cocauo ‘cooowo -,~n~o~o. coco.“ 1; emma.o coach”. I I ' c -.Boo»... .fi«°»o_ .lt nanouo "moo.= ¢mno.o ~noa.o. oaao.o. ooou.o vomc.o. o~n°.o. so«°.ol «on°.o,: «duo.e ocmo.e ¢~n~.911-:~noo.o;-. ocvo.a«un - .nao.o.‘ -uooa..- .«aco.q -mncc.9i-- «sm°.o ;gzso=~.e -.:;smno.o‘:,.os~c.o n««~.c;.:.omo°.. ‘ :omno.c i;-ov~a.oy;; os¢o.o.. : ; ..omc.os-.mem~°.c .oo«.°‘;;,o.on.o oaoo.c.. ,- oom°.° B~.omo.,:~.do.fiw. cmhnno ‘iscuwo,oy 1-«ono.o,-s-asvo.e : :..ono.o.r.:«u~c.o -‘.-¢o.o i -d-ho.o.o‘ _oamo.o.- :wmao.o :-nm~«.o. - anao~o.‘ ~°-.° --ocoo.o., moo«.° nomo.oc. A.~on«.o. oocu.o ..~o~«.g onso.ow. afloa.e ~0~o.o 9. omeo.o.. ,uoso.a. ouao.o«v ..aoaowou ounfl.o unwa.e - -ocdo;o. ..ooco.o .ooa«.c : -oo°¢.« 5:;ooau.o; -.umno.o,: - ‘ - no.9.n;. o~ca.a - -- e-‘i-;:. ....... -:-ooou.«:r- n w.v n asha.c¢ sv~o.° mcoa.o onuo.e .wvoa.o. nooano4 wnso.o mano.o. ..onna.o A.onoo.o _tmsan.c .o«n°.o. w~o°.o .caua.o. m«o«.c ovdo.o mow“.co onoa.g . ooo°.° wmco.o.A nooo.a ”moo.o., oon«.o gmflno.c .ovaa.a. mc.°wa .svono.c.. noco.os cnoo.o novn.o coco.oa snac.oo mmno.e chew.o nmvo.oJ ~0no.o o~c«.o noma.a 0059.9: wooa.o cued.ow canu.o omno.o sono.o wooo.o. oouu.o «mwo.u nmoa.oc vnva.o mooo.oo nomc.o onao.ou sun«.o oncn.o coca." a an no .xukb4<2< 7o~k no mIOH»30¢ m. w W M BOCOIO- CNBCI‘- 'h'alc ‘l‘ruf‘ll‘ “hl.\.: . ...—Ml”- -..-‘- . -— -‘5 . —v—..... .132 canc.o cco9.o. .o-°.o n-o.o «Noo.o. vvvo.o nuoo.ou «moo.oo °~u°.o .ose.o ~s«..o. «mma.oo voo..n n~H=.o m~«..c. oumc.e oonu.o. no-.o. ~a-.° ooa~.cu oono.o. onua.o «cno.o mafia.=. uueowo. nooo.eu noco.ou ad nvno.c cano.c. naw°.c omeo.o 3:... «nao.c. mnco.o. u~¢o.oo vmam.o. «daa.c» ooua.a muoH.c nmao.o; mmto.c “oco.c cmm«.e omsa.c nomu.cn «den.c oon«.e oo««.c mo~°.o o~n°.c ccam.o eavo.n‘ oc~°.°. o°a«.a o~go.r. mnxo.o omna.ou «uno.ou mooo.o sono.oc oovo.o mmcomo umo°.o. «vmauo oaoo.o. avno.u onco.o onmn.o woNOmo. onco.o «aco.o caho.o onoouo. «mao.o onoowo oocOyo «cco.ou .aouo.o., ovnomo nosa.o. ammOflo! nowoqo awfioqo noofi.o cana.o sund.o; anvo.oc nooo.o. «moose vsma.o mvao.o. onca.o «smo.os oaoo.or onso.ou mmou.ou nooo.o mwcu.o- smoo.o so~«.oa nn°«.on oono.o onmo.o ouv°.o mvno.o. souo.or onuo.on «moo.oo unmo.on mauo.oc 0009.9 c.1nu3 nox9uo mooono wokenoa account nose»: mafia": oo~o~o. emNQuoc .coOona. vane»: “maouo.. .osmona tmomonou- onoono avomno Ammouou Yuccauo. ongouo.., nomouo. nmwcno mono»: “magma Dadono. «moons. vmmnnon onNono vvsono vrq«.o ovco.o somo.o. cano.o memo.o «mmaoo. woao.os sooo.os coop.o camo.oo muvo.o ¢ow°oo cowoooW .somm.a coao.ou ”moa.o amco.o coco.ou wnwo.o «mma.o owmo.ou ocna.o vmvs.o voca.oc onco.o. ovao.° wnvo,ou covnoo‘ nmcm.cu used.»- coao.ou emmo.o. sono.o- mcdo.o. nomc.oc v-o.cc mono.oc unoc.ou woso.o mnmo.o * mnwo.ou odmn.o: wo~«.ot waoo.o mmsm.au "mm”.o con~.:c snao.o woao.eu Hv~=.oa mooa.oa Amso.cs uuoo.o. mavm.oc novo.on w~¢n.oo moce.cu nnm°.c mnno.o flooo.oc. oaoo.o “heo.o anca.o «mma.o Hm¢«.o wmnn.o .cmsa.c «moo.o «nfio.o- m~co.o mono.o “fina.o nvnfl;o Amnfi~.a “oco.o- veno.c mwuo.o Homfi.o memo.c mono.o. mv°n.on omao.o .. .o‘ \Hch c oofio.o mnna.c. mdmo.o moao.o. ofioo.a. emo«.c. shoo.o ammo.o. mnc«.o. flnmvjo. onho.o nano.o. whoa.o mona.o. cofic.o mmno.o oacflqo moamno. Mafia.c. “mmo.o boogie. coma.o cand.o. mnaa.o. 505°.q. mvflm.o. nvwv.o. th?~31)J newc.o omao.o owco.o wmoo.o nfloo.o «wo.o moao.o. .oono.o ceda.o mann.o nano.o wmmO.o «Hafi.o mama.at oooo.ow scoo.cu onwfi.o wmno.o 5505.0 Herc.ou mmmo.ou Hocc.ou nv««.o 75).),l . nn on ma vn ma _ 1 vn a. ‘ Hr ow “ND l nu“... ..____—--_ _- . —‘ '-"-~—.--—9-'9...._._..'._- --- H . £133 m\99.9. 9999.9 9999.9. 9999.9 9n~9.9 9999.9 9999.9 9999.9 9999.9. n9n¢.9. 9999.9. 9999.9 99n9.9. 9999.9. m9~9.9 99.9.9. 9999.9. 9999.9. 9999.9. 9999.9 9999.9. 9999.9. c999.9. ~9m9.9 9~9~.9. 0999.9. 9999.9 9999.9. omva.9. cann.o. 9999.9. cmma.9. mnn9.9 9~«9.9. 9999.9. 9999.9. cuoo.9. mmoo.9 999v.9 9999.9 n~¢9.c. «999.9 9999.9 ~999.9. c999.9 9999.9 9Q99.9 0099.9. Nn99.9 9mn~.9 v9o9.9 ~099.c wvn9.9 c~n9.9. 9nn9.9. m9m9.9 9999.9 9999.9. 9999.9 thooo 9999.9. 9099.9. 9999.9. 9~99.9. 9999.9 ~99~.9. 9~99.9. 9999.9 vmx9.9 9mn9.9 9999.9 999999. 9999.9 9999.9. 999999 v~99.9 9~mv.9 9999.9 999999 ~c~999.9 9999.9. om99.o 9999.9. 9999.9. 9999.9. 9999.9. 9999.9 999999. cu¢9.o. noncoot omvo.9. nV99.9. 9999.9. mono.9. 9999.9. 9999.9. 9999.9. v~m~.9 ammo.o 9999.9 9~m9.9 9999.9. 9999.9._ n9«9.9. mooo.9. 9999.9. mooa.o. 9999.9 a999.9. 9999.9. v999.9. 9999.9. vo~9.9. 9999.9 9999.9 9999.9. 9~vv.9- mmno.o «099.9. ov9~.9- masons. 999999. 999999 .99999 999999 .999999. 99.999 99.9.9 999999 999999. 999999. 999999 999999. womomot 999999 999999 999999. 999999 999999 999999 999999 999999 999~99 999999. 999999 999999 999999- mnoxuo haven: O'Cono «999.9 m~99.9 9099.9 n999.9 ¢9~9.9. 9¢9~.9 nmmm.o 9909.9 9999.9. 9v~9.9. mmn9.9 oomooo 9999.9 enon.9 9999.9 9on9.9. nku.o. oaw9.o «999.9. 9999.9 9999.9 9999.9. 9v99.9 9999.9 mo99.o. nvm9.9 ovmw.o ¢~n9.9 YOflOoOI OflNOoOi v9nc.9- 9099.9- ocao.:. ovom.o. 999a.9. mn99.9. mo~9.9. 9999.9. 9999.9 w~99.9 v999.9 9999.9- 99>9.:. 9nm9.9- 9999.9. 9m09.9. monc.o. 9999.9. 9vv9.9- 9999.9. vvv9.9. 9999.9. omn9.9. 9099.9 unno.9 9on9.9- wouo.o. mono.o- mNno.a quo.ol mnnfl.c «999.9 m099.9 c~n9.o nwoo.9. con~.9 once-c c909.9- 9999.9 ov99.o 9999.9 N999.o- .9999.9 omm9.9 no~9.9 mv¢9.9 ¢ow9.9. «999.9 o9n9.9 omn9.9 ~999.9 n999.9 nnvo.9. 9999.9. wv99.c. hmuu.o on99.9 Hmco.o NNcuoo movNoo “no.9 C‘s. 9999.9. mn~9.9. 0999.9. n999.9 M999.9 9999.9. vnm9.9 9999.9. ooo~.o- v~9«.9- 9999.9. 9999.9 9999.9. 9999.9. v099.9 9999.9. nwn9.9. n¢99.9. 9999.9. ~9w9.9. w~c9.9. nmm9.p mmmono. 999m.9. «999.9. 9999.9. vmmN.o m9m9.9 9¢vo.9 vm99.o om99.9. 9999.9 -¢9.9. «~99.9. 9999.9. Nuca.o smm9.n 9999.9. 9m99.9 00:9.9 cmm9.9 9v99.9 cnoo.o- 9999.9 9999.9 9~99.9. 9999.9 9999.9 n9n9.9 nn99.9 N099.9 .999.9- cowu.o ov0e.o- (V 5 'J‘ I!‘ V! lf‘. 0 m #‘n‘JOUO' Ovation ',o¥'u° 00' ‘900uo I‘ CC‘~U- (‘.\ 0.0 H! I1 p\‘~ IDII ...!C ut' - ...—-... - —.-_ 'onc— ..-, ...— q--.‘ .. .-.. ‘134 «sno.o ooou.c auan.o. anwu.o oxau.o cmvo.ou oaao.os mmwn.ou osc~.o. nodo.a. ~am«.o oomv.ou mooo.o mnem.ou oona.c. oome.o «coo.o «oma.o coca.o nude.o. caw..o. acmo.o. onwn.os e~««.os sdou.os vsao.o nano.oc oaou.o no«a.° onca.o caco.c soco.c owcn.cc unuc.o csao.o somo.o ovco.o nuan.b mono.o n~c9.cu mama.c oono.cu unvo.o ownd.a onn«.o.. 0900.9. moe~.o mooo.o. amm..o «sno.o¢ maco.ou vuoo.o cann.o omvomo oo-mo ooo..o. «geomo. «gnome. o~oo.o ooadqoc anaemo. namo.or annexe» annowou «oooqo. nono.o. mvoo.on swmo.o nnmo.o smo«.o vnno.or “vac.ou nvoo.om .meflo.o codo.o nn~«.o Ndoo.o onno.ou «ono.ou tona.o¢ omeo.o¢ coco.os svaa.o o~n«.o. omco.ou coco.o onco.ou coco.o msoauo omoono «cannat Nana»: hmaono mmcano vcnouo- ahaono nomonoo vmmonoo ovmauo «ono*o.. Radono cmoanoo ¢ovonoa cwomno ocnmuo anOno amNOuo. «vacuo- adOOno cmoono. Homa.ou ammo.o coho.o vamo.o cvno.o omso.o nmoo.o Noon.ou oncOJOo ovoo.ou manu.o .n~««.o snno.o mood.o. coco.o. macaqo coma.ou noo«.o omco.ou mowdoo «noooo occo.o vwmc.os waum.ou mnaa.o ooou.ou wmao.ou moan.os ammo.ao mano.a mowa.a snao.au oooc.o. “00”.: vaso.on odn°.o acco.os oanm.oc cmsa.ou nmuo.°u ««o«.on macawoa oosa.ou unoo.ou oaod.oc onvo.o hnna.c .ano.o mqmd.c nooo.o con“.o amho.o whoo.o wbso.ol ammo.oo mo¢«.o Afiooo.o mvmo.o cova.o ooac.o ‘omco.o vao.o vooo.o chm.o wowo.cu ammoool mano.as wnco.o. cnm~.o. moon.o. uooa.cu mnvo.ou nono.o. tocm.o. ooaa.on noao.on ccna.o memo.o. vcsa.o emad.os ~mva.oc fimoo.ou vmho.q. wwna.c whmo.o vonowcu amen.cu unmo.° mano.o ocfiw.o noma.o ovmu.o unho.ou cmmc.o wom~.o mono.o omvo.o hocd.ou mace.o saco.ou moan.o ncco.oo 00cm.ou voho.o Hvrc.o -c0.o cmmv.o .oflo~.o ceco.o onuu.o c: on as us on ms «5 .\v§..v INVI- ‘ - llo.fli.k ICII I w '9’- _ow—9..- - -. “a-.- ..- -- £135 5999.9. 9m99.9 9~c9.9 n9~9.9 9¢99.9u 9999.9. 9099.9 9999.9 9on~.9 mmv9q9 mooo.9 «999.9 9999.9 999~.9. n999.9: “moo.9u 9999.9. 9999.9. 9m9a.9o «non.9. 9999.9 namn.oc hv99.9n 9999.9 9mn~.9. mom9.9n mm~9.9o Odnfl.o mu wnx9.9u 9999.9- 9929.9. 9999.9 oom9.9 9999.9 9999.9 9999.9 ov¢9.9 9999.9 990~.9a ”999.9 9999.9. m~99.9 9999.9. 99~9.9u 9999.9 '99~.9c 9~n9.9o 9999.9. 9999.9 9999.9. uvn9.9; vo9a.9 9999.9. 9m9~.90 9999.9. 9999.9. 99 $99.9 r‘. m999.co <~n9.9a 9999.90 99m9.9| Nooa.9 oo¢9.9 nnv9.9n nmao.9|. ncm~.9¢ nvoo.o N009.9| .9999.9u hvno.oo ntho.al vmoo.9o. vooa.9u v9m9.9 9m99.9n nmna.9o 9999.9: om99.9 9~n9.9 m999.9a ~999.9 omN9.9 9999.9 unco.ot nu 9999.9: 9999.9. mnwn.9u 9~99.9n 9999.9. 9999.9. mmm9.9c 9999.9 9999.9. 1999.9: 9999.9 m9m9.9 «muo.9 oowa.9u 9099.9: m99«.9u 9999.9: 9999.9. nv~9.9 9999.9: m~09.9u 9N99.9 NONo.ol NH (mac 9 1 Tllrl:-.: ..:. :1mosnw. a. cosmic" ovnxu .m canno.l" monan1o s, .i 1111 m4a1. cnmhcool 1 Inulmohcm .jflnuoflfl Gamma-d mtoncsm K \ ,i‘lllll‘ . I‘\I I N Kali |V_ . .( m~aau.~llnl|nomnu.n. manna.” nannm.o. anon“ menar 1111;111:111-.. l. t m2 «Inn ll;11!-11:11:11111111111111511 ;llluillz;1lz-nrlt.1 . :.1:11|:1:a-,~n. "awe a. awemaem zeitsom. Willi. I; ,1 :11 . , scan.nn.fl.7s I ...ruua >u<2uc.. Itlllllllt 11.11.. _ . . C‘U r muomr>m cocoa 11-111 1 l A 111). '11.;an 111111111 1 no ocean a on mu =m__~t .Jw season1a1-nn;.zaficoc : : mace ~zu shxhh om .---sy..:1s.;1-1xl.l . a- H... ~1m.llaem>«myl~kn2N 1 on 11nnc°c 1 -11.-11 l mz:_n 113 «fiance mu Nncoe 2:1 ; '1‘ l!‘ r . 1 1: n ..isll:l.xnnlllzxoz~ .-kank n... «_mllllmmohmllnkmnn. :-.nficacals-lll. zaeuam J_.m naoaon ..s:-sl{: 3:1:11211111 1 :nlt-11-l-1f§.:1 a. w>zbzu 1 1:11.--...tllz.l ,1111111,.ocaoal:l1uzw44~u1 mpz~on sanzw . .---tlil- (llllilli-ll>ll1-1 11.1-1..c M. . -111. - 1- Wiliillllnlll1li-llllll 1! J I. mmuzq~u<> W1? ..-;.:;-.1-£§§§1:§ill-11111 -- .1-1llali!i1 --11_1-1:i- : m111- . ------ ; - . :1111 o o . :11t- Nvmo.o 2 . fl.22; 1. -11. 1111!. 11111 131 111é§il£é.:-i3-: -. -1 - - n+.-mmfloc.a. msenOna onavo.o. oo»no.« as.~n.~ sandasn mousse“. meownwm-1111mwnoswow1.- anmow«:: 1 w4m .11-1,.11-11111111111111.111-.ll.11111!-11|.ll..111-11111111.1..11-111.111-1111.:11: 1111 1.111 11113111111111.1311 ----11111-11|.-11.1..lll-1- .1111111-.1-...-.1-.111-1- 1- .1... .... .3. .1 - , 1.- ,11 en~n m iaozu _ .- -._.-- z -1 :1, _ ;-.-1. --:-21-1:- 11-1:-;- ..:; -; -. - -. , 1. 1..-: ”99-nsoc.a ~c=.nanm.flt ~o=.~o~n.n. moo-anns.n. ~93..na..~. moo.flflac.n ~s=.oonm.a noo.a~no.n. Noo.oc~o.n ~ao.«oa~.n ma r “co.annc.a ;~co.mmoc.v 1~oo-oann.n..moo-onma.a..”93.noxvqa ~oo.nvnm.n1.uso.dnmd.m- Noorvxao.d-1~eo.namn.o. moo.ntom.m -.H . 1;,1;;: . “octaanc.a oc=.cnno.o moononom.m- moo-nnaa.n «ostensm.n moocmsoa.a. mso-~mov.v. mootmncx.o. meotowmm.m. mootflmoc.m n“ . uoaI¢HVConl Neal3m0h.NI:NODliknflomt Noeldnnm.d-doaloocd.w dootchamad.1NDOdevo.Nt HoDIOd\aoH.1NOOINCOOoN1 Ncotuooc.w NH -;L- - - r «oo.onom.s meatvooc.a uoa.ooom.o. moc-ovoo.a aaarsmoo.d moosmnnc.a mastscvn.o. nocifiaa~.o meatowon.cu mootvmuo.m fie . "oo.n~.n.n ~ca.onv¢.hc Noo-oo~c.~. moo.ofimm.d.-o=.ov¢e.¢2 ~90.~nan.m -so.a~oo.¢.-noonn\~n.~. mco.mmfie.~ -oo.n¢on.n ,ofl1,-;-t--.1. “couaxmx.au meatncaa.o Nootomoa.fl moouawon.u eoatnooo.~ moo.o:ac.at vac-confi.fl. mooraao=.e mcotenvv.ot mootflnoc.a o . "octaoxc.~. Nco.om=e.H ga°319aan.a..~oo.oexm.m .moa.¢mmo.vn1moo.no\c.~.-ao.novo.m .moo.aoaa.~. wco-mvnw.m. mooucen«.vt m :11 ------ a “concanm.a Noatamlw.« Noouoxon.w Hoosmwwmsm Moaoflcxo.v «connwnmwfl Van-mama’s: dooum)o:.w Noorvmmn.nl Ncoudnmm.h h r «oatoawa.v ocatnooo.h -oo.¢m~c.v1-~oo-aeko.e.13:3.oanfi.a1-~:o.vnna.o swaauonam.or,moo.=~av.\ -~oo.nvmo.- Noo1xeo~.n o f. -.1: . ~ao.~nsm.m ~ootoooe.v. xoaaacke.m. doouonmm.a moa.cmnoio «oormoan.m naoaanao.o. mcotoxmn.n mcotmvac.o woornuon.v m r .oa.d\=c.v..~o=.oaod.«.-mco-nsac.~;-~oo-n.mona.-was.onaononwoowflonoun.:aao.~ncmnu .moo.aaan.n.1meo.senn.a.-~oc.aaon.vn v - 1 - 1-. outnsum.m. ~c=t~ane.n wooaanoe.mt «cosmoco.a 1~catomvvia-1~oo.mxxm.o11~=ooo=an.nw1nootwanm.n moa.HMcw.« 2°35?; n . «oatmnao.m.axes-navv.at1~oo-nanm.m11mootvnmn.nc was-ueno.m. ~oor~¢do.: usouocon.«- moo.an=m.m1laaotacon.n.1nootnmmo.fin «1.: ;:;1;1;.. ...-..:... ...-..:: eE......11e_--_-n.......,m.-.... ..--meet....w-eo-uémn .meme-w-...-.m..-.-..-....1..1-.-.-.,1,1_.fim.Hug-2.... . 1 ~.- .. . .. _ . -. . T -. fi-- .02 mccl-.:.:1f.11.tl:-:L.:1 ;-1:- 1-:1--..1- --:-1 .:.!.i. 1.11211111232111:111 é-f_!1-:;: . . - zcm14mm<4 zc~»0uw .v .l l-‘.1.vl‘l"l‘l‘-'lfl.1ltlllllll 3111. ..I’lnlOQI ‘I..)'lt!".l‘ lll" 1‘- . u 1 I f -1 121.:311 .1111 1 .1 ..112:.11i :3 x~zkco . :1- . u“ll\\ - III. .:.-II I ”OOleuJOon NODIN-nAYofi \u IPH’I II-K‘II I‘ll 1.! 11I-IIIILIII I 'll. .NmCOIIIf1NU ‘ NDDlmlfl-Oh .3 NOOI—Ucnv ' V3O'fl-JIU“VI'|| “CCI‘nc-H1 III "(l‘n'lll‘ . 111 1. .1 1 11 1 I 1 111 1111 11 11 1 1 ! 1 1 1.1111111 111111 111 11111 11111111111111 111-111-1111! 11.11-11.11- 111111 1 1 1 "1.11 1111 11111 111 11 1 1 1 111111 11111111 1 1 111 1 .1 11111111111111 11111111111111 111111111111111.1111-1..1111-111111111 - l .1 111- 1110-: 1 M .1111111111 M 11111- 1-::1111111- -1-111111 : -:-111 ... . 1 11.11 11.1111 11111-1.---. . 1111.111111- -1 1... - .. - - 1, 11 1 11 "1:11.111 -1 1 1 --11--11.11-:1---- .-11 11-11- 1 1- 1 1.1 .:.-1111.1 1-1111 -1. 1.1: - -1-1. --1 11-1 -11-111111111--1 1 -.111-111 -11-1-. _ 111-1-11-1z-111 - ,1.-1 - +111 N. 111111 111111 1111 111 - 1. 7.11.11 1! 11 1.11 1111 1111 1111-11.:111111-111 1 1111111 1111111111111 1111111111 m ,. 111111 1111-1-1.111111 hwy-111 -1.1--: 5 1 ...-11.1.. - - 1 -_ W1- 1 :11- ,111 111- , -- 1 - W. 1 111--- 1- 11 11111111 -1- 11.111.111-11111 T..- .111- -11-1- 1 -11 111.111.111.111- 1. - W .1. - 11 11 111 1 11111- -1- 1- 1.11-1i-...1-.111. 111-1 11 1:11 1-111 111-1111111111I1111.111.111.--11.1....11.1-11- 1111-11-1. 1- - 1- 11 1 11 - - 1 “co-aoco.m ~ca.mmwv.~. ~oo.uamm.a. moo-oamn.n. ~==.vv~o.a. noo.o~om.n nag-“mod.“ ~Do.monv.~. ~0010mmo.n ~ao.cmoo.v m“ . 2:72:13...- mconmcvm.fl1uooccmo«.~1woo-aim.c! ~331uooo.n-1~oo.m~oo.n. - naoxnazfiu1~oo§nnn..:.- ~couomon§9 wood-$0.6, . I .11 .-:. "no.a\=a.d Nan-91‘s.“ ~on-wc.c.~. na0.«cca.o. ~o=.oocm.~ moo-“Nao.~. ~=°1H-n.n. moo.anac.~- Nee-«men.m1 Nco.mwdo.e ma . .~OOISBn¢onO Noalmpflv0~fi 1VoslndcrofiI1N00lvvw0n.a..-. NDOINNDNuN1.NOOIOVHm-.ov {NaotmofinpnI-Nootn:\9o\ INootflmmmawt Nootv©oaoc NH 1- -, .1 ,.v woo-oans.n ec=.-cm.o wan-comn.fi1 woo-omna.a ~93.»mo~.a woo-Noon.o. ~39.9-¢.~. moo.aanfi.n. ~o015mom.e. ~oo.moom.m "fl . “coconnm.w ,ocatmumo.mulmcouov~o.wu1Noocovno.d13030tvmnfi—ai.mooumdnv.o,.NDOLNNaoer moorsxon.fit-Ncosmhnm.n Noolwxoo.n .o«.111:-. . ' wO°l\H)°.QC ficalflfidfiofl NODIhNHTod ”oatmcflmon NOD-.QMOmaofl WOOOVdVOoV VDOCMOdObMO NDDtND\Oo\ NDOOQCNNCV' NODINHficufi 0 v 3:13~o.m.-23.33513313351- moouainé1~313~§a 1 moo-msnfiu- 33.53;. - woo-$30.7. moogooamé. 29.33;. @1111: -- _. w°913¢n=.a n==.n=o~.n Non-$can.n Hoa1n¢xfi.h ~o=.¢aafi.cu ~oo.a~as.o ~=o.avmh.m. moo.ouae.~ ~oo.-qc.~. moo-Gnuo.n. s . "co1vxnn.o. mcauanon.az-wo=1»-n.a-~oe.v.ofi.o.-ao21050a1a ~001aono.ms1~3ounoom.au-oouocso.:. doo.mmcm.ac ~oarmoo~.m .o 1 - r “counnnq.o aoaavaco.v NoaINNOm.mu ~oo.«mmn.o waaaanno.sc «oouoaxv.n vaouo~o¢.ca noocnmac.x moocsmvc.m Nooamomv.c m . “EDINNdp—ufll VoaanHO-NI..0DDIOmn¢oN -.-NOOIQVMBoNliNDDIQGOmpnt NOOIONDQoOG1daolHMNNnN 1-NOODO\DN...\E1Nco.-Dmmv.n NDDIOmN¢.VI v 1 .v "oo.n\=n.m. ~c91~oso.s goo-«emc.~1 mco10noc.s ~o=.avuo.¢. moo-amoe.o uso.o~e~.n. Hoo:»~m¢.n wee-mmsu.o. moo-cwom.flc n . “calflxsm.n1moalnemho'1NOOUOOOK“V1Noalev0nNtldaaInQON.HOINQOINA".811N9DOOCNVbn1NOOINnNh.081-fiOOIQCDQoN ...nOOINHOmodl N 11.. ...s Ffi...w-fi...§3 20.2.: gin-£- énwfiw-mfifiduhfig.... gimmmwmggg. 23E.- ” . . . ..T- ~11, .oz-....c<..--.--1111-11.1111.-- ----111 111 11111111 11 1:: - 1111-1.---1.1.1-1-1-1,-1-1..-1 30x 4wm<4 zo~kuum 1111-111-1-x~z>co ,- z fllc‘n.‘ {.1 (cc-' ‘FI‘ I. I An ‘- CCC‘h-‘hk 1 V N ' i‘ ‘1‘ In ‘11 1~ \§‘ 0‘ \(F Ntu. uazlmmzh1zo1mz.».c accede.“ ~ca.»coo.~. ~o0130n¢.a1 ~091qono.a. ~o¢.~x~x.a. moo.o~ce.a ~=o1mmom.fi moo.nnvn.au ~on-¢ogc.n ~oouuomn.c ma n.acnv.m11wag-donn.~11~oouommn.mo1mcoumamn.cs1aoaunnwoJH11~ooaacoc.~ 1~=ounamm.om moo1novm.~u.~cc1nmco.nu moaanoo.n-:va ---1111111” o.¢an~.« moo1oo’q.fi uoo-\odn.~1 ”counsdc.n do=1afioa.~ woo.m\ox.ao uaosovn~.n. ~901~a~o.~1 mootoov¢.c1 Nao.m~mfi.e a“ . a.on\n.aa ~o=anc\n.H-1~oocncon.ms mooumcvc.v1-~oauoxov.m 1~oosmsac.~- Vaounnnv.v. Noormcnm.o11woo.w~om.mu mcouwwom.m .Nfl - 1. r. u-a\ao.v woa.amoc.v ~oa1omcw.fi1 moo1ecmc.~ xaaux~an.v noo.n~o~.n. Vaounfi-.n. moo.monn.o. Noa1~ume.wu ~oouados.m a“ . or~ac¢.a 1~c=.o:3«.uus~oauxcnc.wc.~oo.¢qam.a11~oauvmoo.~ -Noouooaa.o .~:ouocvm.-.mso.masc.ot.mcouvhcm.n Ncouomno.n o«.111111 1". a.o=)a.fi. Hoa.»a¢..a Noa-\amn.fl noo-¢wcn.h ~3313fio~.fl mao.~\flm.fl. u=01HmH=.n. woo.ooa\.n ~c010mnm.m. Noo.momn.fi a . ou~¢ac.mu.mcossann.n11aaonmnvq.a11~co1~mzo.m.1ooa.om¢m.dm.~oocnnmm.¢c1naosmvnn.n .noo.ooax.xu-mcosocmx.nu Nooumfioo.vs m, 1 1:1 111:.1 o.>v«m.« ceasefiom.s woou~mmc.m «ooumcwm.s ~331nnoa.~u woonwcam.o unocnvmv.v. wooucanc.n wcogacoc.nu ~oo.oq¢«.m a . aucnoc.w mcau,n=~.« 1ooo.om.«.a11wo91mnwc.ac-aaaoon~«.a 1woo.o¢om.~.1u:o.~com.m.1noo.-no.ou1~co.ofinm.oa-wootwmoo.m -o 1--- 1- 1r.; uuooaf.o ~c31-a~.«c wpaunmom.v1 ~o01ovnm.o ~o:.ocom.~. a001-ov.n ~=D.No~«.n. moo.m\on.o moo1v¢cs.o moosmmcc.c m . a-sn¢m.m. ~=31amdc.nn-woo1nvmn.fl;1maouovmw.¢.1~oa.~oo~.911wooumasv.hu1aaocomom.~ -moo.¢o=a.ou Nao1mnnc.« -mcaumnvm.c;,v_..1: - 1. ,' nuno\c.nu moo.oonn.m ooouao«¢.o1 moouvomo.n eoa.-nc.n. moo.o\an.n nnouoo°o.o. Hoo.=)e~.n «cacmqow.v. moo.aman.m n . u.c\on.n11~ooucnuapmnlmoouoomc.na1~oouuwoc.nu1woo.odmn.a.swoonccouaw11vaocmnnc.a11~oo.~¢om.c.1«co.mcno.~_1nca.-ao.fi;.~-g:i. 1111.1r_1 fluommo.n wen-memn.u ~oa-~«m¢.vs moouovoa.~ moouamoo.~ Nooovnvv.w waoomnvmueu mocca~n5.w neocmwao.uu aooc¢n«h.~ u . 1 . .11 nllllts If..i,111 I141: 1:! «11.. . .I:"\l1,'1.llvnl\ll.l1.1rlll. .l I . , W1v111 cz mc«u1é1 .. 111111111111:.11111.1 1- . a . - -.11. - . 1 . - 1cm Jumaq zouhowm 7-11 11111.1 xumh09 omqocu1-...1 11. l ;Uv.s Mluv‘lt (II!\. lll‘flul “IRW: \‘Cvt J.x!1o\ I. 2'; \ b 1;. , 11111;1111111-11:-1.-11111:-- fi-e . r 1 -1111-11111- .1.. : - 111111111- 1 1 -1 1:111;1-1- 1-1.- 5 1M 11111 .1111 1:11 - 11 11-11111 -- -11 fi1 . w unnnm.oc mflmno.mn p2¢hszu T11 1 1111- - nu~m . au~m ..azouo r1111111-1-1 1111 manozu 44. mon mkz.:«Itoo no m1.m:. 111- 1-1-. 7,. F 'IW‘H .I. fit 5« J1L‘!I\.II ' -..ll I‘lll'l -.‘lllllll'i's,’ 1.1"“! I'll. "llll II‘I'III‘V I -.:..li‘ ‘- I10 I! ..: 1 ' I! ' ‘l'l ‘-’- ..‘ ‘ ' .1 111111 I 1 I. 11 1 1".[111 1.11 1 11111]! 111 111111111111 11.1.11. 111111 1.1111II 1 1 1 l. ..1 | 1. 1 1 1 1 1.1l 1.1 11 i I. .\ 11111111 I 111111,! 1111.! 1111 ll. 1 I 11‘ 1 111.1 1.111 11 1111‘ I‘fill’ll 111.1 1 111 1111' 1 1 ..l1l11lll'111\'ll- 11111 I [1111' 1‘ 1 I 1 1 11.1 1 11 1 1.1 1 III-.1 c 11.11 .I 111.11 I. I.. '1 o 11 I. - I 111-1.11|1.1l|1111 1,111,! '11. .1111.‘11|.l.|.’1ll.1111111 1111 1! 11 1 1 .1 .- ..lllillllill 1 1.11111 1 YI‘I'IIIG-‘ ill-1111111.! 11.16111 1 - 1 . [I'll I’ll-11.1111I1Ilwlila . 1111’ 111 -1111 . I 1. c1 11 11 I 1.11 . 1 11 1.1 11 11 1111 1 l 11- I 1.1 111111-1111. I11111|l111 111111111 11‘. 1 1 11 1 [11111.11 1 Y 1.1.! [11111111111111.1111 1.. .1 111 11.111111111111111! 1.1-1 1 .11 11 11.1 1 111l111111111$111l111111111111ii1111111111111 .111 1.1.1 1'1'1‘111' 1 1 1 1 1 12.11 1111.11.11.11. 11111111. 1 1 . . 15.-1‘1!!! 1.11.1. 1|! 11.11 11.1.. lic . 111-.-111. I 1" 111.11.111.11! 1111111111111. .1 1 .1 .... 1 1 v... 1.11 1111.11 .1! I. 11:10.111111111111 111- 11,] 1 . . 1|] l‘...‘ ' l (1!" " I. \1I I’x‘ l‘fl. A, ‘II 11'. 1| 1 1 I I. 1.111111111111111... 1111 1 1 1111 1. 1.1 I11" I'll-1111 11- 1 a” -.l'li' .10-I.-! 11 1 - 1! 111111111111]! 11111 -11 f'11l1l111l11111l11‘1 1 , ..1.11I11.111 .21-1111.111- Ill 1 I 1. .l11111 .1111 It. I 1 1 11 1.111] 11111111111111“! 111.1111!- 111 1.11 l 1 11111111 I 1.1 1.1 1l|171|l 1.1111 1 111 1 .1- .- 1|11|111 1'11“- 11,111.11 1 I 1 1 1. [Du-111.1 I11 1 11 l- .- 1 1‘1"" 'Il- l." ,"h' 111111-l- 1 11,111 V111; III-111111111 11111111,. I I II, .1 1 ll 111‘1111.|11l11l011. 1 1 . . l 11-11111 I111! 11 1 11|1J11 11...|111.1I11.1.11.]!1111ltll‘l1cll1111-I"! 1 1.11111 $11!.11 111 111 I. 1 . 1 1 ’1 I l 1- . .. . 1| 1| 1| [1 (1’11! 1. .1. 1 111. ll. 1.11."!1111111'11‘- 1 1111.311“- 1’13.."“' 1.! In “I!!! 1‘! , 1 I -111- 1 '11. ~ 1 . l1 11.11111 ‘I-11lol111l1111 1 11.111111 1|.|11 11111.. 1 11 .. 1 {£111.11 111111 1 111 1 1.11 11- 111 111 11111 . P. 1 1 I 11" 1| 1.1.1111- .‘1111’1‘11‘11111‘111. l1 1. 1 ‘1 1111111111.I-llll.1l-1v1..111|1l.l11‘ 1.11 l 1 1 1| \ 1 l .I ' Q 1' I luv" 1 C.- .l '0. ‘31.]- ll'llll .11 1. ll‘lll . . I 1 1 all: 11 1 111‘- l1! 0 . 1 I. 1.1111111... I111. '11111 011.." ..lv 11! 11.111.113.111 11'11 1111111111.! 1111 11111 11 11 11.111 11 1.11 I .1 1 11 11.1 11’ 1 1 11 1 . 1 1 n . . 11?. 11 l . .11-1 1.111111111111-1111111111111111 £1111 .11-1.11.11-11.111.111111111111111 11 1.. ..I 11. ii... 114.1111 1 11 1 - . I 1 '1 ‘1 I | l . 11 . 11 - I I. _ NON v'lniotll'lllllv H fl .1 1...! 1113.131! 12.1.. i I- 1 . - ‘ . ' H \ U . - 1 ..... .1111 11. 1111 1.1111111"!!! 11111111. 11111111 I111! 1 I, 1 1 111111 .1 11111111'1111111 Ill-.11 1'1 1.1!l111l111I1111'1'l111111l-1. 1111111111... 1.. . . ._ . 11111.1 . l 1 1- 011. .oz uC(1111 : 1111111111- 1 111111 1 1 1 - 1111 11 i ,- 1 . : ..:..- ..: '11." -nmzawmm<-ym1a«m C «maauu 44~h 00\¢O\OO ocaon.n unmo¢.c mwnz«~a<> munmn.cu m~v0e.« mmmhc.m .- s m2 annoa.o emnmu.o cano~.« nofiom.« «no.9.n mnmmo.w no: nmnn 32m ... .! ocooc fiJm caacun ~zu uhnmh ~29 m>mpzm cocoa mudquu . mace: hzmou ...u Pfiguu WHY/(Mm m4a<~a .).qh m4mmhzm 1.32w4 aog~.cn.fl.7:a ...xuma >«¢z_m..e mgomz>m cocoa vfiooo nficoo wagon aflocn cacoa “coco cocoa mecca, cocoa nooco «been fleece conga Ifimontn Jduk ocOJ WON-h Vln~NKC ~\Mu.\l. '\ N ‘4 p\\o '\\\v1l\.’ :161 Y 0 up -- ' U-._~< _.—,- v t a F-A-‘flw— - .-r. nmfikcoo mu2<~t<> me<~m¢> ' O O O . . .___. m - —--_ v _..._., ___'--._. .7 ——.-—_—T.. ...—.— '- d00l36fim.nb NoelNkvc'r 'CClc‘A‘o DUI a 0‘ 'II“ Vt J \‘U‘J "oo.oono.u. accommnw.u ”ao-onmh.v.‘~c°-oo~o.fi “9°.mmac.~. ~c°.oovn.fl. "ocuofin¢.a. wcounaoo.u "occmomn.o. Neoaowflu.~. "ca.vonn.~. uco.ooon.m "cotanoc.«u manuamom.wn _ao.oomo.s. «counoov.v "oo‘oma~.m «co.¢oow.e. .aaaooo~.vu aco.n¢nn.m "oo.vomc.n Nco-noou.n. “co.ocom.n. moo.no~n.~. "oo.¢ovn.fl acc.ouoa.a ”ac.~nos.~. waacdcnm.n wao.mn~«.s Nao.m¢mo.o "ca.mo~«.a ~ao.oo-.o. "cogouv«.nn «connoau.nc ”co‘ono«.o. «conmnwn.w. J a .02 mo~fl.a wen-momm.nu mco-mnmo.n «oo.o~nm.a nooucnmm.« NODOOOHH-ht «oo.oon~.~u noouonoo.o «oo.o=¢~.o. noouoaov.u a°=.~nnm.a Nunavuoo.na ~°=.ofio¢.a mo=.~ndn.n accvnmom.oc maoucomc. n - - memo~.o. w soosn.u “nos".N osmoo." nvu¢o.c nodoo.fl. bmsoo.d «om~°.°. - odmcd.~ , oamoo.~, amcoo.fl W A cauoc.u. dunno.“ oamu».°. ~«v»o.u “00¢“.v nnnoa.w cocoo.n acoo~.n - coccc.H oaoma.~ M : mz p ‘ . Nana u cacao “cocuoou.u «conuosv.nc Noouumnn.n Noopacvv.n Noanvooo.mn "occaoco.o. mooavnun.nc «cocomvn.vu «co.oooo.o «oouonna.«c "co-oona.ns «pounuoo.s. «cc-«oos.m. Neouosnn.n wao.ad~m.n. fl fl .oz wo 24m: m4m<~a<, “Ht-4 In! meqq 7o~huwm X.&k(3 wU><~I<>OU ‘CIOOO‘ '1 o“lfl|‘n‘ ' 0“l"" '- -" I‘Iu‘ ‘ "I {“I. ‘ :" llll. | I I ‘ '1 I02 Munu‘lt l?.b\ l~\u.s-\J \Ir\~;.J..Uan ".¢.~¢...- 1.. ”a...a"~.s; ~o°.mvvn.u «ao.~n~o.d. ~oo.oo~«.o. ”9°.ooa..fl. mco.-h".c. _ooeummo.~- «co.nama.u "a°.mouc.~ nc°.o~oo.o «ao.oeao.m. moo.¢mo¢.o "cognoom.dn moo.wooc.ou "eo.ooon.n. ~co-»hhm.m “a°.maom.~ «90-0"om.~. “9°.oflom.~. aco.do~u.m .oo.om~n.~ moaufifiow.n. “co-oosu.¢. ~co.~nan.m "ocuacmo.' «09.5”...a co.»noo.a. ~c°-nman.« ao.oomn.a Nee..mvw.v “canonnn.v mcaumofio.ou ao.omoo.o Neo.mhv~.u. coguoov.n Noo.~o~o.s. n .02 maOU HfiflHHHHr(fifivlflHv{Q-1Hfifl flu“||l“‘ cz flunk“ O‘clc'i‘ o c I"- "‘l“ '1‘.‘ ““|Il"‘ ‘ lac!" ‘|' ‘F Y -... Q "T ._— -.—~_~ ...Mv~. ., n _ .. . . “eotuons.« noo.haon.h .aa;oonc.n. noaunoow.u , fl .fi M t cocoa.“ Hfi.~o.«. .oz mu nmwo~.co macaw.m mocno.~ c m2 noor¢fiom.n. noc.ocon.o Noa.onaa.a mootccme.~. moo.n~a¢.fl dooumndo.an Noocvodo.uu Nocunvmn.ou mooumomo.n. «acvunww.w Vac-ovuv.« ~09.ad~o.n. «oo.on«o.~. «ao.noou.n «co.~aoo... ~90.h»ao.. «cocnsna.~ «oouoauc.». noogsmos.~.‘«bo.oshfi.fl. a" am Neochm¢.nl cm mo7<~m«\ m4m<~m z«wr m4a«~w<> 7cm: w4c<_a<> v-Iv-QH 301 me¢J zoubuwm Xukk(7 m071u~u<>00 In"! “(CI-“Ku‘ok1h- ”CC-I'CKyt V- ACC|hlh0.NI NOO'HNDK 0." NOC"CNH only NQOIOAUI.N ...n ,, - '_ ,W.--.——n .-.... »‘ ~ 1— . n.' ...,‘a—T--.*.- . .. .- --—-. -.. --. “.1 “,0 1L _ “°°.mo~m.~. ~o0.noaw.m ”ea..voc.n..~=o-m¢aw.n. "oatmnmfl.n. ~co-~m~b.n. _oe‘homo.~- mco.noDu.m «no.0nnm.u noa-a-o.n ”co.uovn.cs «couamom.o "cocoacv.~. n¢9-omom.o ”no."ooo.m. Nco.mo~«.u ”aoefinvn.« wco.oa~m.n¢ «no.uomm.s. «o°-on¢o.m ”co.0¢e«.nu moa.vwn..~. ~°o.meoo.~. mca‘an~o.fi .oo.ufiov.c acouoono.u .oomohoc.ou nan.n«~h.o. moo.nm~¢.h ~co.no~o.fi .oaemoun.n «no.95on... .oc.oon~.s «oououov.w. oo.°ooo.s nc°.a~o«.n. .oz wo«m « . g n _ 7 . noo.o«oo.m Noo.mvnc.o. mooeoumm.x Noocooov.vn mac.nmno... moo-ancn.no moo.~mov.m. Hoo.oun~.« noo:ovvm.~o moocvwn¢.wo «ooamom«.¢ moo.onmo.uc «oo.ooo«.fl. ~e°.o~nu.4. «co-doom.«s ~ao.oomw.m ~ao.momm.~. neo.o“oo.d. ~a°.aooc.' ~ao-ficn~.h woo-ovnn.m woo.~mnc.~ ~9o-~o~a.m. Nco.¢no~.m Nco.oo~o.m Ncounnao.~ woo.mmoo.~. moo.enac.~ moo.onmo.uu «ao;on°m.~ «ca.~msfl.«s woouovnn.w fioa.¢aoo.fi. «ca.n«m«.u mooumuon.« noocssco... noouoov~.cu ~°°.oou~.m. ~°°.mvom.mu moouoono.m moacunmo.fi moo.-m“.~ Noo.o~a°.d woo.ovom.nu .Noornaoc.a doo.oonc.a "cocooo~.«. Ho°.um~«.fl. ocoo~one.fi woo-ocm~.mu Hectosdo.m ~oa.«mmo.mu «oo.o~mm.« ~oo.~moo.s. noo-mhos.a moo-vhnm.~. noorcmnn.~u noo.nov¢.n noovmcvn.n moouoamc.~u moo.vuaa.mu woo.onmm.wo nooumnmo.on noo.nfi-.ec Noouomms.fiu voouovku.m Noonovmn.mu nooamvmc.u mooumnoc.a Noonvhmo.ms wooucnoa.~a noocvoum.“n ncc-~onc.ou moo.vc~v.w- moonococ.nu Nealovflo.m Nootnvmm.v ~oo-~cmm.m. ~o°-¢oom.o. aeoummno.a. Ncounmmv.n waoufiamc.fl «ooufloox.flu woo-Vfioo.fi. “canouac.m moatmnoc.a oco.mnom.a «cococao.uc moonuoan.m maouuuso.~. doonnmaa.au Noounmoo.n mon-momn.a moouvxmd.n mooyfianv.c. moosonmh.~ noo.~aofi.o mooaonac.ao moosmafln.n woo.o~mm.¢. wooumame.o Hoo.nama.fi Noo:Hnm:.nu moouv~ma.nc «ocaocao.nt nooaflmon.o «oa.¢~ma.au «oouuwos.n mas-onom.« Noccmonm.mo ~oa.nuoc.«u Noauucmc.fin wossocwn.n ooauvn~«.m moasnoae.n moo-mocm.m meannen~.u mcouoflov.cu moormcmm.~o mooamnom.a “antenna.“ xoatvmoc.na moauaofin.m acouvhmo.as doc-Nomv.m Nooo¢o~a1wn noouovvm.c moo.figen.m mug-ovkm.h nccuvcmo.cc nee-an-.nc noasnfimm.m «go-wcnm.ma ncoumcon.nu nag-ocma.n noéufiqu.wu nos-muoa.nx ncaunnem.vn Nco-NcOc.uo ”co-«cKn.au moo.umnm.u. mcaunwou.fl moonvomu.ac ~macownn.o flflflHfiHHHfiHdfifiH-dflfifl In: mecJ 70~huwm xuthax wo><~k<>n0 I! API(V‘UI |,!('V§ubl¢!.l x “Mil-I.» ll lulu... KKJJ~H mksxuauu ..r‘hxék V«L.mu‘(M\J Y! (N\ NU?“ \x...\x ... M.,—...... ...-n.a.- 'v 0"" W“ .— .... .nr-N- .:.--.- .— . m—u—rT -.— -..... _o—vn. . - _____..,...,__..———o- ooo-.c ”cmnm.o mcfln~.c naomb.« maon«.~ nmooo.« enuge.«. -v¢>.« ns~«s.o. 1150 a .02 mu¢m a_m . ..4 as»: ..waJuu au44~u >mpzm aoooa . mace: .¢< m4u<_ s. mmnz<~a<> mummnool ? m4m«n. masoc.m 7, m4u<~m mmmN¢.~ >< mqm<~mu .-. : m2 ocooc canton unusn “mush «recon «shun ocean shuns shuns mnocan chums «nuns ocooo oosca oaacca omaoo mecca «recon onoco scoaon «coo: ococoa nfiaaca nflaocc Deccan among muqquu so. nmnn »< ombm«<2~muc. om cocoa m cm u mh cacao m on a om nficoo a mu m cm «goon g cm m cm Hanan D N" a so aficoa c a" . a ma “coco c cm m cm oocoo c «c a He mooao d .m« a ac cocoa c as - ; a co mecca «. ~— n mm mecca m mm a on “coco a so . c no cacao mhz.on >xkzm 1.32m; :«monzn JJ~u WON.“ _ . “.-..-qu " . —~- row-n... .. I u I ...- ... .— .....- .- "*‘--' "-'-CD ~ V ""‘ .7 t Y . A ' y . . » ; I u ‘ . | I w o 0 K ¢ :161 T . I Q ,~.— ‘,__‘___,_‘_~_ I i I ) I | I I . cho—fi - .-r— -V‘" . _ _ H.—‘-.—.~‘.._—“v..- _T-~-v~4-.__. : moanc.o muz<~m<> m4m<~m<> vaoou.c sovhn.o omnnm.o ovvon.o noomu.o ouoo«.o ouno«.o. oovNu.o nomnm.u -. mooun.o .--. mU2<~UQ> Noctonkw.ut do¢tflrw~.«| NOOINBVC.H ”COIO'Omunl NOE‘QDO'cH Net'ccco Vu' 'cclhflv WI T \I-LNLLU L «I.» "ao.ooac.hu Hoe.m«nu.n “no.0noh....moo.oo~o.« "oo.mmuc.~. ~o°.oovu.«. "oo.ofisc.«u wco;nfioo.a "catnomn.o. Ncotomam.m. "co.ecns.~. Nco-ooon.w "co.fi~ac.fl. woo-«mvm.~. _oo.oomo.h. Nco-noov.v ”cosomom.~ «co-ooom.¢. _a°.ooo~.¢. «ca.nvnn.m "co.vomo.n Neo.nooo.nc “ao.ac°m.n. Noo.ne~u.m. "ooavoen.fl aco.ofioo.a "coumnon.~. ~eo.«°nm.n woo.mn~«.s «ca.m¢~o.o "ao.moma.fl ~oo.ac-.o. ”oo.ooVfl.n. «ca.moau.n. ”oo‘onou.oa ~99.mn~n.~. _ fl .oz moo~5.5 ~==cmm~n.m Noecoocw.na uoaumcxw.mn NaosOD Nn “AGE “"V (_Ahzlf:"l_ ~9.136¢-0t -6.8462~0| 1-1662~ou 921-002 174-002 791-nn? OH" .0! um. ’3' «Na DOC 00C Ill #05 0mm ovn huh twp u NNC no: Doc not he: no‘ he! 0. “OI hoosn.fl ouuoo.“u ‘ .. ”cou~oo«.u acouuouv.nu moo-unnn.n ~oa9occv.a Noonvooo.ms "oocmocc.m. woo.¢n«u.n. moo-omvn.v. mac.oooo.o meo.aonfl.n. ~o¢-a~oo.s. ~oa.aoos.m. Neo.ounn.n woo.aflnm.n. N .02 wu X;&b(! wu><~3<>ou mmnmc.o “ounm.o afloo«.c no mmuz momom.on omooo.a .ocoNH.N naouommm.o| «N wu2<~w< m4m<_a< wu2<~w< wqc<~wf WU?<~G4_ m4m<.a? 2O HHfiHHfifir¢dHP‘HHv{c-Qc-4HN 3cm 4mm<~2<>nb CCIC'Q“0 “Cr-I‘vh.‘ ll OCCIn-Fcl‘ 1| (I..! I|1l I 'II IIan 1 III -- I1 0.52 “My“ )Fss J‘V~\.~1J 7.\~L)JKI. _ . 1-- - M M . . mu. m... “mac“.o “summ.o ononn.o mmwmm.o .wmmfl.o q fin.nm.o an'on.o mama».o onom~.o «~onc." _ T r ., —1 I m 909mm.“ "chdn.~ cmsdo.~ omoce.o ovvam.fl¢ _ h W ««.no.«. ensue.“ msnho.=. "av...“ .mmmo.n M W nun» weo.uonu.u noo.h«°n.h. .oo.oooc.w n°°.namn.v Noo.o“»v.~ .ao..on..s. nee.naom.d floo.ooov.+. «oo.v~na.u n~o«fi.c cncsc.o mnono.« fimomo.« n aaoco omnna.o nono~.a cocmo.o. oooom.n oncoo.o Hucko.o coavo.a cocoa.» magma.” Noecm.c c~nvo.~ csmoo.a . mzqmz m43 nocoo.c nomnm.c mnnOc.o .. mmuz<~z<> «mmcm.cv mnqma.m mocao.~ nootvamm.no noouovon.o Noonowao.n Noouoeme.mt «nonhuov.a «N doo.mhao.a. ~°o.oo«o.un woo-nvmu.ou woo.momo.nr aoounnflm.fl ~no.¢wmm.nu cu .oonov~v.~ woouoawo.nu «co-anno.~o «ocunoos.n «caumaoo.c. moonsnua.v acocnsn«.~ «couoasc.nc acouswos.~..ucoonusa.fls ow W m w . _ . .oz meOU m0?«~&<. MJm “fl“ an: 4mm¢4 zo~kuwm CC‘OCC‘IF I: "CC..- f“c c2 dl'C‘l. fin-I "Cr-'0‘Foc " "CI’II‘( “c ‘Ihflu - .' ...—*7---—A- . -pv .- ..a pn- an: ”,0 1L , ”co.momm.~. «ca.noow.m .eo.vvoc.n..~¢a-mvom.n. "eagnama.n. ~co.~m~5.m. _c°.~omo.~- mco.noDu.m "eo.onnm.u noo-«-o.n "oo.uovfi..- ~co.«mom.o ”ooeo°v¢.~. nco-cmom.o ”oo.nooo.m. ~co.o°~«.« ”ooran'n.« Nco.oo~m.u. .oo.oo~m.u. «oo-omoo.m ~o°.ov¢«.su mca.emnv.~. "90.noom.~. Noa.ou«o.fl .oo.ufiov.c «co-oono.a .a°.m5o..o. noo.nd~h.o. oo.nn~v.n uao‘nouu.fi .oaeuoun.» «co.osnn.v. oouoon~.s Noo.ouoc.w. oo.oooo.s neo.H~o«.~. _ , _ n .oz wonu ..OOUNDO'c'I VCOI'flhU1rI NGOINnflG-hI hCCIOVm-fl ( “cell‘s-.6 . httlfllut 1 (II 1111 1 111 IP11 1 ..Il 11l| 1": J\V§.\J >C~‘h\l1\lh I ... .01. “bay“ Vi” . -.-__—4-_--1_ "'-‘ .‘u . ...—1.... PH‘ _a a "-....- A—v—v— ...--.. onnnm.c oomu~.o “mane.“ nsnno.a. .oz wo nanoo.o vmowfl.c .nacoo.° vonoc.fi ~H0~H.° cocsv.u -ofis.a voono.m omnmd.fl monco.o mn~o=.~ aflceh.m vamnu.a ovoov.n nonv3.~ camofi.~ e anomo moo1m-m.n woo.ooe~.fl. ¢o013mm«.o. ~co.«omm.o. «0°.ooom.n. woa.ooaa.a. Nao1onofi.n. noo-o.nm.~ Ncogonoo.m. Noa.oaon.c Nootofimo.n Nooluomm.o Xqu00 whose.o monwn.o vdmna.o ... mmuz<.a¢_««\ m4c<~qu murqwqu wgmv_a Flu-(fl In: mec... >O~huwm c2 wC‘t ',,__‘,- .» ,-...._ 1169 AIHEHA. _ ..qul. n; 1. . . .. clung; lull-I'll? 1 Alrl‘iliufln-l . 911.. r4 ,waoccoo«.«1 «ocuomno.m «ocuomno.u «oo1mwno.n «aouhsoa.flc ~oo.~nan.~ moonmona.« acouooan.d1 moccveon.fi maocoou¢.fl1 a“ a “occnnuo.n1 neocoomm.n woo.o~ov.o. woo1meo~.« no01H0Hn.n moo.om~c.m «oo-noa«.« moo1mo¢n.au moognnmc.“ moc109en.n ha fl “oouovoo.au Ncotmono.n1 «coammno.n ~oounv~n.¢ Noounoom.nu ~oonnvoc.ma "conuncv.aa nooaooon.n magunean.u moouovmn.aa o“ « boomvoon.hl Noouomfiu.c Noounonw.ul NooanoOH.N Noocsfim~.n Noocnnma.ul Nooovmos.n Nomnmnma.mc maalvomc.wo Nocvmmnv.rl an « "eorunmv.v acoasovo.«m noormoou.ua «counono.ss woo.om¢o.m moo-nonn.c mccunnmv.w nconcaod.~u concocsc.v ~c=1nvvm.o vfl « "ac."vno.ou Heouusom.fi noo1aoov.m. mocunnmo.va «oo.~m¢o.fl ~oo150am.fl ~0o1nnoo.u floo.onvn.fi Hoavoonm.w1 ~c010m~w.m n” « .ocumvao.o. «o91nmva.~1 «oo.on~n.au nco.ofioo.m Hco.o~ma.mu Noo.mfios.¢ vao.oooo.o woo.m~Hm.nx ~c=.onflv.fi1 «go'sovd.n m“ « "oo1ouav.aa mcouaaos.o Noasnumn.~u ~o91ocno.o «aouocmc.~ moatnmon.fia ”co-soon.fi1 woo191ao.~ mosromnm.n mcc1omom.hs «a H ”oo.ooma.« ~901~mou.e. moc.cvvs.n ~o°1ncfic.w. woo.oco~.a. ~901nfidw.fi. NQQ.Vfl¢o.v. moo.n~nm.c mao.onm¢.u ~co.~mm~.fi a“ a “9°1mmon.og «co1vonv.o ~991~u~a.w noo1~w«~.h doctoamm.d noonnonm.m Hoosmovn.d woauoo°m.na qoogocom.n1 ~o01~hmo.nr o H ”oo1avvn.n ~co.~h~«.w Hoocnvnm.» «oo.ocao.a moaavomo.s. mooanfikm.wo «aoamoon.~. Noncovoa.o moosnnmc.n moo-engu.mc w fl «eocnoao.¢a noa.~o«~.n Hocsoouc.w «oo.omov.~ «calmm~«.m1 noouco¢o.m Hao1~vnn.fl1 moo1vqon.n accnonvc.fi wco1nfimw.v1 u H "aouooom.as «catcomn.“ Noog‘omo.v1 wooammwfi.mc ooooaoflo.d «901mnmw.flu «coauafiu.fl documam~.n aocamcvc.mu floocm.o~.—1 o H “coonwna.nu noounonu.m ~oo.nfi~m.flu mac-vo¢0.m Hoouono~.aa Hoo-~aon.fi wooccvmn.v “cocoon=.«1 naaunmmn.m «acumen..m m fl .aouvnvo.vu acoumovu.u H°o1mcon.~o floo1~vnn.«u Hoaauwan.a ~501¢¢mn.v cocooonv.« aoocoavn.nu qaeuvcom.fi1 mocnvfimm.fic v H Maoenhno.v Naatooou.no Noosmvoa.o NcOIVHoc.m Hoommcma.m doosaono.wl Hocnmcvn.nl HoatNNOH.x Noctocnm.wu maelmwom.n .n « aocawmv.s «canovom.~c woosnmmo.\ «ooxowvo.a Hoocmoec.n1 noocnmmn.m woo-coon.«o monsoono.nn «unaccom.o woounnmn.w N d Nooumumu.u woa155nc.su moo-vnvu.ou coo.nwmn.c. Hoo.~«o~.au ~ooc~amfi.n moo-«mm».«a acounuom.n «eacnsmn.~ «noccwmn.a a a s .02 wo¢a 2r: 4wm44 yonpuwm X~th<3 m07<~2<>00 -_.-_~—g~m. .02 09(3 glf70 --...‘. Q ‘T ""V’ —..-—‘-.1 .v. ..- ..., .-...1-.. -..—“-1 - ..H‘- .... V . "oaaoaso.Hc «cousvou.nu moocomHv.~ mac-HmsH.u moo.onmm.m. noouwsou.H Hc01onn.H1 monsoocH.n .mocxovoh.o mun-noHn.u1 Hw .,, H w3.2.3.0.. Hcounmcw.Hs coo-«Hmo.n1 mooanoo.o Hao.nmno.Hc moogmoNH.n HaoanHn.Hu nongovoo.oc HaooHoHc.H mausmcum.v om H 09.50qu... wconouougo moo-vaooJu «99.0.00.» 2.9-Hamid NooovHo~.mo Neonosoc; Nooemmn~.n naaumeho; N21502:... on H a ., ,1 ,. A... . . _ , : 1 W o .02 munb Vin: ndln4ufl minim Nlr,‘d \‘A.1.~\ |M\\~U\In )kk‘hxy '02 W0(l ...—vu—r ., «no-r-.. -... a - ...... .... ”..., ,. .oz mo~ ooc.amoo.H Hoconvoc.c =oc¢oHoo.H1 soc.cnmc.H NaaunHam.na =a=+~mOH.m =991~mom.m aco1ocoo.H aao.mnmc.H. soaoocHo.H H0510Hm~.nu oocoHeno.H oa=.nc¢H.v =o=.nnoc.H coo.=oo¢.n aao+ooo~.n cocoocvc.H soacaomH.m Nnnm ooo+mooo.H «ca-ecmn.u aooovmom.an oooomHHc.H waocomos.mu accounou.H oaooonom.m =co1mvoe.m oooaono¢.H1 oaoqmvon.H HcounumH.nc ooa+nnom.u accommmn.m coo¢cnac.H ooo.¢»mm.n ooaomcvn.n omoqnmmc.~ om°1omnn.~ Humm «40‘0‘¢J\Of\¢35 HfiHHfifiHriHflHflflflflflv-‘fi 30m 4wm14 2C~kuwm h (0(ntCDhO W Jmck 0.! 9C1 Wm.“ 'Ifwm halnlhw “Ir.c If. I Jrsbr‘lu >QNLU.K.~H -~'-——-—-'—-« ,‘ __ _T.¢V.‘.j - . .--,- ~7.H--_.., . ..-- . .. , u .- Y _w‘...__-' --..-_H_ ruc— .- ~...~——..- '- T... .... -~_.”Ja—a~_—~_ _ w-"’ 1372 on .02 mo«u Hoo.ommn.n. Hoo.nn~o+~1 H==.a:mo.w- Hco.mamm.n. Hm oco1mHno.H cocooamm.H =°o+mo~m.H ooocmomn.H om °e91n¢vc.~ ooc.v~Hn.~ ooo+H¢HH.~ oco.~c~H.« 0H Hflr‘ cunm nunm «unm .1 Hunm znx 4wm<4 2oH»umm 23: mst 20 m2~h CU omqool rmr—o-w-a-n»- NH Ocz w0(l| ‘l‘ Clcr" r: ‘lJ‘P Ff. . r.—r.“.——.~ _. . I 1 »~“'——~—C-w———- ! , _ _...___._I ‘ i 1 1 ‘ l 1 ' --~ —-..—-..-—. "‘“"‘,"~"‘. ’ "'T"“"'""".' 3174 7"" -.v» - -1- -..—y. wad... —H, - ~ f w ouqoou mmw: mwuHxFoo moochov.H «canoooo.H1 moouonoH.n Noouova.¢ Noocvnoo.H1 oao1~Hov.vu Noo1nn~v.n1 noauuomo.o moccnch.H ncouamHm.Hn Hm H .oosnoHH.o.‘~oa.o~HH-. moo.momo.H. Noo.ovsc.s Hoo.ovnH.H. noo.onHm.N. «H91H m . 1 . . . .. . - _ . . v a 5 mac «neH H Noacocmo a ncocman H cm H moo.uovn.n. Noo.HH¢H.n. «99.9»oo.n. «09.noom.n noouomo..~.‘nocuuHmn.H. woocucnn.Hu HHp.n=Ho.H vac-ONNH.m1 mnocoan.H. 3H H w . z, 1 1 1 . a ;; . 1 M NH .cz mu<¢ 20: 4mm¢4 onyuwm kakit WU2<~n<>QU Dm43nl Dd .02 muu(hu ‘I': "1"". t'..‘ ‘I'NI 2.11 JJIC~LUKW ’Y'flr ,_.., -_ _ ..-rv—__,..-....—.’ 1375 ---“ ---—u. - . . .- Hy- ‘7 WM . II lllll girl “5!... r. u 1. I .- ._ ... .. all .. 1.. . 13.,firwik— 23: mHI» ooo+mm~a.m oco1mnnm.m aaoovaH.~ aco1om~n.m mH H oooomcso.~ o¢p+0>79.H =33.mHv~.m ocoqonom.n HH H Hooovmmv.H1 Hoo10o=n.H1 Hno+nma¢.H1 Hcoonav.Ha oH H ooo+mnom.v ooo+un~o.n so:.oo¢v.v ocoomkcH.n mH H ocooonom.H Hop-~nmm.a Hoauman.n Hnoamomn.¢ «H H ‘, Hoo1H¢om.o Hoo.¢nH=.~ Hos-coon.o Hoos¢«HH.« ”H H ooo.~onh.~ coo.Han.~ soogmcox.m Hoo.nm~m.m NH H coooman.m coo.vvon.n s¢a¢ocoH.m ooo.~mom.m HH H oooowvcm.oa cocoomvc.au =o=+$omc.ou aao1momn.ou oH H mcocovom.v Hoo1Hanm.H ~091~m¢«.v1 mocuomce.n1 o H oco.~cwH.o coo.o~no.n aoo¢Hmmo.o ooc.Hn~o.o a H oc01nmmm.v accommoq.n ooo.mqac.m oncomcnm.m H H ocoovooh.v coo.o=oH.c =a=.ommm.v aoo1~oom.v o H HooonHn.~ Hoo1Huom.~ Hoc+ncom.w Hco1~Hm«.m n H coo.oonm.n acoowmnm.n ooo¢nmmo.n oc91HmHo.n e H ooo1oomc.s ocoonamh.H ooc.mc=c.m cocoonhm.u n H aoo1omnm.u coo1cnon.n oo=.Hn~<.n cao.nmvn.H m H Hooono~«.~ Hoo.oan.~ Hoa.~ncn.~ Hc=1onmm.m H H vunm n-am wnnm Harm 301 mec... zcaruwm zo m»:o~m1 u>~nC~LUKW --..._'—. - _ ..- *- .m— ..__ -..... --—- en .02 wu<1 omoo~.~HH ¢mkmo.HHH cmoom.VHH Hanna.HHH H7Hhm7ou vunm nunm Nana Hume aaoao . annexe HJH so“ wHHHHmzon Hoo.nmmc.o1 Hoo.mnOH.c. Haonoan.H. Hoc1mcon.c. Hm ,- coo.om°v.n oao.oHcH.n °°=.oo¢c.m ooaommmq.n ow He°.ovnv.¢ Hco.snm~.a Hoo1vmmo.m HcochcH.H oH “HM vunm , nunm munm Hnmm 3c: meHH 20H.” 2:: mek 20 msonwz u>Hpqtqnyoo no wqmcs 'flf.‘ 'Nfl1" rlIIIL . III 1 1 -11» 'c2 we‘t .~ :1 .1 1 ... 1.11 11111 . 11 11 .1 11.1. v. — .-- _-' ‘T'“" ..-_-__’ ”7 .--.- ._-_._ _, , r _ - ‘___._.-----1.. ...--._,-_, , 1377 ma .cz wo