THE MATURE FEMALE CLOTHING SHOPPER : PROFILES AND SHOPPING BEHAVIOUR

This study was desib'1led to profile the mature female clothing shopper. More speciflCatly rertain \'ariables that could be attributed to differences in consumer behaviour were investigated. Mature female clothing shoppers were segmented as clothing modera tes, clothing enthusi3sts and the clothing unconcerned. These three duster groups significantly differed regarding clothing involvement, clothing orientation, activities, inten..>sts, opinions, family orientation. needs and media usage. No significant differenct>S were found among the three groups reg<lrding evaluative criteria for clothing and clothing store attributes. Profiles of the clusters were develoJX'<I, along 'vith applicable marketing implications. Recommendations for further research are made. [n the USA the most important consumer market in terms of growth is the mature market (consumers of 55 years and older). Contrary to rcseilTch on the clothing preferences and problems of the elderly or mature consumer where specific trends have repeatedly been confirmed by v<1rious reSC<1rchers, there is a need for comprehensive research on the mnrketplace behaviour of the mature consumer (Huddleston, Ford & Bickle, 1993, p. 27; Lumpkin, 1985, p. 271; Shim & Bickle, 1993, p. 55). Very little, if any, rese<1rch on this grolVing rnnrket segment has oc'Cn done in South Africa. Americ3n research fi ndings cannot without question be accepted as valid for South African circumstances, but have to be tested empirically. Lumpkin (1985, p. 286) wrote: " Research should continue to focus on this important group of consumers to develop a more complete profile of the sub-segments, so tha t marketing strategies can be matched to their needs." PURPOSES OF TH E STUDY The broad research aim of th is study is to enhance the existing body of knowledge on the mature female clothing shopper. This C<1n be achieved by setting the following goals: (3) to invcstigate differences in the consumer behaviour of elderly/mature female clothing shoppers, (b) to invest igate certa in variables that could be attributed to these differences, i.e. clothing involvement, clothing orientation, clothing evaluative criteria, clothing store attributes, lifestyle <1nd medi<1 usage, (c) to profile differen t clusters/groups of consumers that sh<1re the same dl<1r<1cte ristics according to the differentiation variables and (d) to examine the implications for marketers in reference to the marketing mix, target markets (colhllmer profiles) and evalu<1lion criteria set by elderly consumers in the evaluation of clothing stOfes nnd -items. 14"1".--1< ~J(" fI7"""r.J . ... III I." .~'. I,.....,." /(> r \I \',.,.T /1'lot"' •• , ,, ,>{ 1/ •• ,. h •• ult ..... (I""'"NI.¥ "f .""1k·" I • ...-I' P,m,I<" itoI),' \; 1 \ L"rrl."kl ,,""2 LITERA TIJRE REVIEW The mature market constitutes a heterogeneous group of consumers and lends itself to the basic requirements for effective market s eg m entation, i.e. measurability, size, reachability, identification, sufficiency, s t<1bility and 3ccessibility (Greco, 1984, p. 7; Schiffmnl1 & Kanuk, 1994, pp. 82-85). Rescnrch on segmenting mature consumers has been reported by scveral researchers. They h<1ve invcstigated vnrious bnsis for this segmentation. Lumpkin (1 985, pp. 272-289) identified three market se~,'ments o f consumers who have unique shopping orientation profilcs, i.e. active shoppers, economic shoppers and uninvolved or apnthetic shoppers. Based on information sources Lumpkin, Greenberg and Goldstucker (1985, pp. 75-105) d ivided the market into three segments: confident comparison shoppers, inactive/apathetic, critical shoppers as well as price and qualityfgunrantce-conscious, personalized shoppers. Lifestyle ch<1r<1cteristics and the impo rtance of retail sto re attributes served as b.lses for segmen ting the market into five groups, i.e. shoppers, posi tive thinkers, the educ.ltionally orient.lted, SOCially active consumers nnd credi t prone consumers (Huddleston et 111., 1993, pp. 26-31). RescnTch by Grew and Paksoy (1989, PI'. 7-23) wns developed into psychographic nnd perceived informnlion source importance profiles of m<1ture f<1shion -conscious 3nd non-fashion-conscious shopper segments. This prO\~dcs evidence that clothing consumers may be segmented on the basis of fashion consciousness and not nge alone. ~ I ail ordl'r catalogue uscrs could be \~e\\"ed <1S <1 market segment. Shim and Bickle (1993, Pl'. 53-64) di\~ded these shoppers into three groups <1ccording to perceived height, i.e. pe ti te, medium and tnll . A profile of elderly heavy catalogue lisen< can be found in Shim and Mnhoney's rese11rch (1992, Pl'. 4lJ-58). Clothing in volvem en t nnd clo thing orie nt'llion of the m3ture consumer seem to be unexplored by resL'nrchers. Clothing involvement is defined as "<1 person's perceived reb-nne\.' of the obiL'Ct bascJ on inherent m .. '"t.'Cis, v<1luL'S, and 2 VISSER,. DU PREEZ,. DU TOIT interests" (Zaichkow'sky, 1985, p. 342). Laurent and Kapfcrcr (1985, p. 43) refer to fh-e facets of involvement, i.e. the perceived importance of the product; the perceived risk associated with the product including the perceived importance of negative consequences in case of 11 poor choice; as well as the perceived probability of making such a mistake; the symbolic or sign value attributed to the product; and the hedonist value of the product. Clothing involvement has been determined in segmentation studies of tuxedo customers (May, Shim & Kotsiopulos, 1992, pp. 31 -38) and big and tall men (Shim & Kotsiopu!os, 1991, pp. 16-24) but no such information regarding the mature consumer could be found. Oothing orientation could be described in a similar way as shopping orientation, Le. a shopper's style that places particular emphasis on certain activi ties, interests and opinions regarding clothing (definition adapted from Shim & : .. Iahoney, 1992,. p. 57). Clothing orientation hilS been included as a variable in a segmentation study by Shim and Kotsiopulos (1991. p. 18). They used an adapted form of Solomon's (1987, pp. 11 0-128) and Tigert, Ring and King's (1975, pp. 46-52) questionnaires to identify clothing orientation. However, no research could be found that deteonined clothing orientation of mature consumers. Fashion innovativeness, fashion interest and fashion awareness are dimensions of clothing orientation but research on the mature consumer and fashion is scant, especially in connection with market scgmentation. lifestyle characteristics such as social activity, positive thinking. shopping orientation and credit-pronenes sct'm to be predictors of fashion opinion leadership of mature consumers (Huddleston et aI., 1993, pp. 26, 28-30; Greco & Paksoy, 1989, p. 20). Lumpkin and his co-workers investigated specific aspects of the elderly consumers' behaviour. Patronage behaviour, use of infoonation sources, store and product attributes as well as age and retirement as key factors in deteonining shopping behaviour arc examples of variables that were tested. Quite a number of attributes have been identified. When compared with research findings of Huddleston and her co-workers certain attributes seem to be of importance, e.g. product quality, attractive prices, sizes and styles suited to age, store reputation, variety in one store and availability of advertised products. Retailers' and elderly consumers' vie\vpDints of the importance of store attributes were also compared (Huddleston, Ford & Mahoney, 1990, p. 75; Lumpkin & Hite, 1988, pp. 315-323; Lumpkin et aI., 1985, pp. 75, 87, ·101, 102; Lumpkin, 1985, pp. 271, 276, 280-284; Lumpkin, 1984, pp. 622 -626; Lumpkin & Greenberg. 1982, pp. 69, 79, 82-84). Psychographies and demographiCS alone are not sufficient to typify consumers to the extent that marketers can use the typologics to channel their marketing efforts. [t is therefore necessary to focus allention on Iifesty[e in order to foon a more comprehensive view of consumer characteristics (Fox. 1989, pp. 80-83; Huddleston et at , 1990, p. 73). Lifestyle can be described as a composite of perceptible and non-perceptible variables that culminates in a person's daily living. The different components of lifestyle are shown in Figure 1. LIFESTYLE I +------+ Activities Interests DemographiCS Social Class I ~~s and Motives Values Family Orientation Personality Attitudes Opinions Perceptible variables Non-perceptible variables Figure 1: The lifestyle-component model (Fox, 1989, p. 81) Due to the difficulty and complexity of measuring the non perceptible variables more emphasis was placed on the perceptible variables in the research without totally disregarding the non-perceptible variables. The influence of retirement on lifestyle and consumer behaviour is mentioned by Lumpkin and Greenberg (1 982, p. 84) as well as Hudd[eston et at (1993, p. 30) . The importance placed on retail attributes in relationship to lifestyle characteristics is pointed out by Huddleston et al. (1990, p. 73). A rela tionship was found between the lifestyle characteristics of mature female consumers and the importance of retail store attributes (convenience attributes, age-related attributes, quality/reputation, price attributes and credit attributes) (Lumpkin et al., 1985, pp. 95-99). Lifestyle characteristics such as social activity, positive thinking. shopping orientation and credit-proneness seem to be predictors of fashion opinion leadership (Huddleston et al., 1993, pp. 26, 28-30; Greco &: I'aksoy, 1989, p. 20). TraditionaHy consumers' age has been considered an im portant basis for market segmentation e.g. children, teens and adults. According to Martin (1976, p. 453) major differences exist among generations with particular emphasis on alternatives considered, predispositions fooned prior to shopping and reliance on media for infoonation. Lifestyle patterns of this group are of interest too because lifestyles directly impact on preferences and consumption (Shim &: Bickle, 1993, p. 53). Ac

[n the USA the most important consumer market in terms of growth is the mature market (consumers of 55 years and older).
Contrary to rcseilTch on the clothing preferences and problems of the elderly or mature consumer where specific trends have repeatedly been confirmed by v<1rious reSC<1 rchers, there is a need for comprehensive research on the mnrketplace behaviour of the ma ture consumer (H uddleston, Ford & Bickle, 1993, p. 27;Lumpkin, 1985, p. 271;Shim & Bickl e, 1993, p. 55).Very little, if any, rese<1rch on this grolVing rnnrket segment has oc'Cn done in South Africa.Americ3n research fi ndings cannot without question be accepted as valid for South African circumstances, but have to be tested empirically.Lumpkin (1985, p. 286) wrote: " Research should continue to focus on this important group of consumers to develop a more complete profile of the sub-segments, so tha t marketing strategies can be matched to their needs."

PURPOSES OF TH E STUDY
The broad research aim of th is study is to enhance the existing body of knowledge on the mature female clothing shopper.This C<1n be achieved by setting the following goals: (3) to invcstigate differences in the consumer behaviour of elderly/mature female clo thing shoppers, (b) to investigate certa in variables that could be attributed to these differences, i.e. clothing involvement, clothing orientation, clothing evaluative criteria, clothing store attributes, lifestyle <1nd medi<1 usage, (c) to profile differen t clusters/groups of consumers that sh<1re the same dl<1r<1cte ristics accordin g to the differentiation variables and (d) to examine the implications for marketers in reference to the marketing mix, target markets (colhllmer profiles) and evalu<1lion criteria set by elderly consumers in the evaluation of clothing stOfes nnd -items.
Clothing in volve m en t nnd clo thing orie nt'llion of the m3 ture consumer seem to be unexplored by resL'nrchers.
Clothing involvement is defined as "<1 person's perceived reb-nne\.' of the obiL'Ct bascJ on inherent m .. '"t.'Cis, v<1luL'S, and interests" (Zaichkow'sky, 1985, p. 342).Laurent and Kapfcrcr (1985, p. 43) refer to fh-e facets of involvement, i.e. the perceived importance of the product; the perceived risk associated with the product including the perceived importance of negative consequences in case of 11 poor choice; as well as the perceived probability of making such a mistake; the symbolic or sign value attributed to the product; and the hedonist value of the product.Clothing involvement has been determined in segmentation studies of tuxedo customers (May, Shim & Kotsiopulos, 1992, pp. 31 -38) and big and tall men (Shim & Kotsiopu!os, 1991, pp. 16-24) but no such information regarding the mature consumer could be found.Oothing orientation could be described in a similar way as shopping orientation, Le. a shopper's style that places particular emphasis on certain activi ties, in terests and opinions regarding clothing (definition adapted from Shim & : .. Iahoney, 1992,.p. 57).Clothing orientation hilS been included as a variable in a segmentation study by Shim and Kotsiopulos (1991. p. 18).They used an adapted form of Solomon's (1987, pp. 11 0-128) and Tigert, Ring and King's (1975, pp. 46-52) questionnaires to identify clothing orientation.However, no research could be found that deteonined clothing orientation of mature consumers.Fashion innovativeness, fashion interest and fashion awareness are dimensions of clothing orientation but research on the mature consumer and fashion is scant, especially in connection with market scgmentation.lifestyle characteristics such as social activity, positive thinking.shopping orientation and credit-pronenes sct'm to be predictors of fashion opinion leadership of mature consumers (Huddleston et aI., 1993, pp. 26, 28-30;Greco & Paksoy, 1989, p. 20).
Lumpkin and his co-workers investigated specific aspects of the elderly consumers' behaviour.Patronage behaviour, use of infoonation sources, store and product attributes as well as age and retirement as key factors in deteonining shopping behaviour arc examples of variables that were tested.Quite a number of attributes have been identified.When compared with research findings of Huddleston and her co-workers certain attributes seem to be of importance, e.g.product quality, attractive prices, sizes and styles suited to age, store reputation, variety in one store and availability of advertised products.Retailers' and elderly consumers' vie\vpDints of the importance of store attributes were also compared (Huddleston, Ford & Mahoney, 1990, p. 75; Lumpkin & Hite, 1988, pp.  315-323; Lumpkin et aI., 1985, pp.75, 87, •101, 102; Lumpkin,  1985, pp.271, 276, 280-284; Lumpkin, 1984, pp.622 -626;   Lumpkin & Greenberg.1982, pp.69, 79, 82-84).
Psychographies and demographiCS alone are not sufficient to typify consumers to the extent that marketers can use the typologics to channel their marketing efforts.[t is therefore necessary to focus allention on Iifesty[e in order to foon a more comprehensive view of consumer characteristics (Fox. 1989, pp. 80-83;Huddleston et at , 1990, p. 73).Lifestyle can be described as a composite of perceptible and non-perceptible variables that culminates in a person's daily living.The different components of lifestyle are shown in Figure 1.Due to the difficulty and complexity of measuring the nonperceptible variables more emphasis was placed on the perceptible variables in the research without totally disregarding the non-perceptible variables.
The influence of retirement on lifestyle and consumer behaviour is mentioned by Lumpkin and Greenberg (1 982,p. 84) as well as Hudd [eston et at (1993, p. 30) .The importance placed on retail attributes in relationship to lifestyle characteristics is pointed out by Huddleston et al. (1990, p. 73).A rela tionship was found between the lifestyle characteristics of mature female consumers and the importance of retail store attributes (convenience attributes, age-related attributes, quality/reputation, price attributes and credit attributes) (Lumpkin et al., 1985, pp. 95-99).Lifestyle characteristics such as social activity, positive thinking.shopping orientation and credit-proneness seem to be predictors of fashion opinion leadership (Huddleston et al., 1993, pp. 26, 28-30;Greco &: I'aksoy, 1989, p. 20).
TraditionaHy consumers' age has been considered an important basis for market segmentation e.g.children, teens and adults.According to Martin (1976, p. 453) major differences exist among generations wit h particular emphasis on alternatives considered, predispositions fooned prior to shopping and reliance o n media for infoonation.Lifestyle patterns of this group are of interest too because lifestyles directly impact on preferences and consumption (S him &: Bickle, 1993, p. 53).
According to Smith and Moschis (1985, p. 75) elderly consumers differ from younger consumers and there may be more differences among elderly consumers than among younger market segments.
Knowledge of consumer's media usage enables the marketer to target the prospective consumer and predict media usage.11 is therefore common practice in lifestyle and typology rescarch to include media items (Fox, 1989, pp. 180-184;Terblanche, 1990, pp. 218-221;Shim & Kotsiopulus, 1991, p. 19).Media exposure of the elderly and their relationship to fashionability were examined.The results indicated that significant differences existed between fashionability and media exposure for female styles.Those who selected classic styles used more sources of fashion infoonation than those who selected fashionable styles (Chowdhary, 1988, pp. 23-30).

RESEARCH METHODOLOGY Sample selection and p rofile
The population of interest for this s tudy is mature female cloth ing shoppers in clothing s tores/boutiques.The target population was defined as a ll female clothing shoppers over the age of 55 living in the municipal area of Ste llenbosch during the time of the study (November 1993).Data was obtained from vo ters lists of the voter areas, i.e.Stellenbosch and Helderberg.These lists were provided by the Department of Internal Affairs.From the to tal of 16 945 people a unive rse of 1144 (6,75%) ladies (55 years and older) were drawn using their birth dates.These names were cross-con trolled with the telephone directory for the area.Kaiser and Chandler (1984, p. 207) sta tes that " telephone directories have been noted as providing fa irly good sampling frames, although a given directory may not include new s ubscribers or those who have requested unlisted numbers".Fo r the pre-test a systematic random stratified sample was taken of all the residential a reas and a reply rale of 65,38% (n '" 68) was obtained.Using the same method of sample generation, the main study sample size was N '" 367 of which 212 questionnaires were usable for statistical analysis.This reflects a reply rate of 57, 76 % The majority of the sample participants (50%) were in the age grouping 55 to 64.The rest of the sample represented the age groupings 65 to 74 (32%) and 75 and older (17%).Of these participants 65% were still married and lived in their own houses.
Total annual income varied between less than R999 and more than R300000 with an average income of beh .... een RWooo and R59000 (57%).The qualifications of the participants were relatively high with 79% having standard 10 or higher diplomas.Forty three percent arc retired and a further 33% never followed any career.
Clothing expenditure ranged between less than R99 per month and 79% of the sample spend between RO to R299 per month on clothing.These purchases were not only made in the Stellenbosch municipal area but also in other centres.
In summary the sample profile reflected a young-older female buyer, with (a) relatively high education qualifications; (b) a moderate income; and (c) living in her own home.The sample profile is influenced by the relatively large grouping of females who were retired and not employt.-d.Due to the moderate income very little is spend on clothing on a monthly basis.

Measure ment of constructs
As the lifestyle and clothing shopping behaviour of consumers were studied, a questionnaire had to be designed wi th the objectives of the study as guideline.
Section A of the compiled questionnaire reflected drmograpl.icinfOlmation and questions on the geographic buying orientation.
The Personal Involvement Index (PII) of Zaichko\vsky (1985) was used as base document to construct Section B, namely clotlring involvement.Twenty word pairs were used on a 7-point semantic differential scale.This method has been extenSively used by May et aI., (1992, p. 33) as well as Shim and Kotsiopulos (1991, p. 18).The standardized alpha coefficients reported by these studies were 0,87 and 0,90 respectively.Shim and Kotsiopulus (1 991, p. 18) used eight out of the original 20 items and reported high intemal consistency (Cronbach's alpha coefficient = 0,90).May et al. (1992. p.33) used four items and reported an alpha coefficient of 0,87.
Clothillg oriell/a/ion represented Section C of the questionnaire, making use of the work reported in Tigert et al. (1975, p. 47).A 5 o point Likert-type scale was used to measure the construct of clothing orientation.Tigert et .11. (1975) report extensively on .... alidating the construct wi th an independent fashion specific lifestyle/A IO factor analytical technique.Both questionnaires for sections Band C were slightly adjusted for language.
Sections D and E respectively measured rol.mtive criteria rrgardil/g clo/hillg (Cassill & Drake, 1987, p. 24) and s/oI1' nttrib'ltes regarding clolhing stores (Lumpkin el aI., 1985, p. 83) making use of a 5-point Likert type•scale.The main purpose of these sections was to detennine those attributes respondents considered as important when selecting clothing or chOOSing a clothing store.
The questionnaire was subjected to an item analysis.The refined questionnaire \vas conslructed for usc in the main study.Questionnaires were available in Afrikaans and English.
Q uestionnaire administration and gathering of data A pilot study was performed in order to refine the questionnaire for further usage.Field workers were trained to gather the data by personally handing out the questionnaires and also re-collecting them.Their training included handling posSible questions regarding the meaning of questionnaire items and helping vel)' elderly and sick respondents to fill out the questionnaire.The main study was undertaken in the same manner but due to the greater sample size 23 field workers were employed and trained.

Statistical Analysis
The SPSS-X statistical package (release 4.1) was used in all stiltisticill procedures (SI'SS-X User Guide, 1988).
The pilot study and the main study datil WilS submitted to descriptive statistical analysis and reliability testing after which all items with item-total correlation ratings 10IVer than 0,20 and -0,20 were eliminated.
Due to the high item-total correlation on the various sub• scales composite variables were computed from the sub-scale items for further analysiS.
The Ward clustering method was used to cluster participants based on the squilred Euclidian distance between subjects in theW-dimensional computer variable space (srss-x User Guide, 1988).
A multivariate analysis of variance (MANOVA) was performed to detennine whether mean differences on the vector of subscale/composite variables among clusters aTe attributable to chance.

RESU LTS AN D INT ERPRETATION
The aims of the study were to investigate certain variables that could attribute to differences in the consumer behaviour of mature fema le clothing shoppers.The following results, in accordance with the goals of the study, were obtained.

Reliability analysis of the main study
The results of the reliability analysiS, after elimination of items, are given in Table 1.Cluster analysis and MANOVA Cluster analysis was perfonned on all the sub-scales of the questionnaire using the SPSS-X program (4.1) to tl)' and partition the total sample of respondents into a number of relatively homogeneous sub-groups that are large enough to be profitably cultivated (SI'SS•X User Guide, 1988).A graphical summaI)' (icicle plot) of the cluster analysis, combined with the distance bcr..veen the two most dissimilar points contained in the clusters tha t are combined at each stage as provided by the agglomeration schedule, was used to determine the appropriate number of clusters to extract for further analysiS.
Based on the criteria of duster sb•..e and homogeneity the extraction of three dusters was deemed appropriate, wi th Cluster 1 (n = 113), Cluster 2 (n = 71) and Cluster 3 (n = 28).
This substantiates the fact thai mature consumers cannot be considered a homogeneous group.
Description of the extracted clusters centered on those composite variables on which cluster membership produces statistically significant differences.To determine the dimensions along which clusters differ a one-way MANOVA was performed via SI'SS with cluster membership as independent variable and the 10 composite variables as dependent variables.Wilk's lambda was used to lest the s tatistical significance of the cluster membership main effect.The results of the multivaria te tests of Significance afe shown in Table 2.As can be seen from Table 2 the multivariate null hypothesis of no cluster membership main effect can be rejected (p < 0,05) .
From table 2 it follows that approximately 77% (1/ 2 "" 1 -i.== 1 -0.23248) of the variance in the optimal linear combination of the 10 composite variables is accounted for by cluster membership.Rejection of the multivariate null hypotheses of no cluster membership effect on an optimal linear continuation of the dependent variables led to a series of univariate AJ'\!OVAs to determine the precise dimensions along which clusters differ Significantly.The results of the univariate F-tests are reported in Table 3.The variJb1es clothing eVJluation criteria and clothing store Jltribules did not significantly differentiate among the three clusters.These variables were eliminated in the typology of the clusters.
The typology of the clusters was bJsed on the mean scores of composite variables which significantly discriminate among the various clusters.
The average cluster profiles defined in tenns of the composite variables are graphically portrayed in Figure 2.
To clarify the identity of the different clusters, the original variables comprising each composite v.::rriable were Jlso stud ied.Differences on the original variables tend to give J clearer picture of the nature of the differences among clusters.

CLUSTER PROFILES
(mean scores of significant variables) Clot/rillg oricntation: The opinion leadership and opinion receiving of this group are limited towards the buying of clothing.The women arc not actively involved in fashion news and consequently do not read a lot about fashion.The clothing moderates are uncertain whether they are keeping their wardrobe up to dJte as well as whether they keep up with fashion changes.
Activitirs: This group is the most active of all the clusters.
When buying clothes at a sJle the women tend to compare prices, which can be a reflection of the groups knowledge of clothing, as they enjoy needlework and making their own clothes.They Jre also very active in the church and community.
Jl1Icrrsts: The interests of cluster 1 arc on par with the activities of the group.The cluster members enjoy cooking and entertaining.They have an in terest in nature, health and dressing stylishly (without taking Jny fJshion risks).
Opinions: A moderately positive opinion towards South Africa and its future is endorsed by this group.Some neutrJI opinions are recorded on whether yesteryear's clothing WJS better thJn today's as wen as scepticism towards change.This can be seen as J reflection of cluster 1 's buying behaviour towards clothing.
Family orimtatiolt: All the clusters arc characterized by J very positive orientation towards family gatherings, relationships Jt1d the vJlue of the f.::rmily in the community.This can to a large extent be explained by the nJture of the sample of this study.
NCl'ds: All the clusters reflected a great need for financial security, self-respect and security.This can be seen as a reflection of the age of the sample and the need for a secure old age.In cluster one there is a neutral response towards whether the respondent is willing to spend money on clothing items.
Mrdia USll,\,'I': Radio is the medium of communication used most by cluster 1.This is reflected in the responses as they have a low viewing rate on television entertainment programmes.Cluster members also enjoy watching some television and reading magazines.Information regarding clothing is mainly drawn from magazines and very little is gathered from television, radio and newspapers.The magazine used by cluster 1 to gather clothing information is the Sarie (a fam ily magazine with a primarily female target audience).'The clothing moderates have very limited exposure to any form of fashion speciality magazines.
Cluster 2 profile (II = 71): THE CLOTHING ENTHUS IASTS Clothing invollJ('llu'l1t: The clothing enthusiasts are typified by a very strong positive clothing involvement on word pairs.The duster members see clothing as very important, of high concern to them, very relevant and useful, valuable, beneficial and very interesting and desirable.All these positive perceptions abou t clothing show a good amount of enthusiasm for clothing in this cluster.
Clo/hillg oriFll ta/ioll: None of the three clusters are early buyers of fashion clothing or opinion leaders in fashion.Members of cluster 2 are, notwithstanding the abovementioned, the earliest buyers of fashions and ranked the highest in the sharing of clothing information.The clothing enthusiasts have an above-average exposure to fash ion magazines.They use these magazines to try and keep their wardrobes up to date with fashion trends but not necessarily to try to dress in that manner.
Activities: The clothing enthusiasts do not compare many prices when buying clothes, do not enjoy buying on a sale and do not make any of their own clothes.This may be due to the fact that they enjoy bying quality fashion items and are therefore not very price-conscious.These cluster members are also involved in church and community activities.
lll/Frcs/S: The clothing enthusiasts have an active interest in fashion and entertaining.Nature and gardening are also recorded as fields of interest.The fact that the clothing enthusiasts are very health-conscious is also reflected in their concern for dressing stylishly.
Opinions: Cluster 2 has the most positive orientation towards life and likes the changes and challenges that life offers.The opinion that todays clothing is better than yesteryear's is substantiated in the clothing buying behaviour.
Family oriellta/ioll: Of all the clusters, this cluster has the strongest family tics and interest in the family.
Nerds: Strong needs towards financial security, self-respect and security are prevalent.The clothing enthusiasts do not mind spending money on clothing and they find doing so most enjoyable.
Media llsage: Cluster members watch a fair amount of television and entertainment programmes.Magazine readership is the highest of all clusters and they therefore get most of their clothing information from magazines.Radio listening is relatively high but they do no t use this media vehicle for gathering clothing information.The Sarie and Fair Lady are the two magazines mostly read by cluster members.
Clothing is seen as items which are not very important, relevant, essential or exciting.
C/othillg orit'llialion: Ouster 3 members do not have a positive orientation towards fashion.This is reflected in their behaviour, i.e. not buying fashion items, not I,>iving any fashion infonnation, not being interested in fashion trends and not reading fashion news.Fashion changes are not adhered to and keeping a wardrobe up 10 date is of little concern to the duster members.
Activities: On the activity sub-scale cluster 3 members have the lowest scores.This is a reflection of their apathy towards doing needlework or making their own clothes.The activities toward clothing show that they do not compare prices when buying clothes and do not enjoy buying at sales.Activity in the church and community is also very limited.
"arrests: Gardening and health are of interest to the clothingunconcerned.They have moderate interests in nature and entertainment.To dress stylishly has an average appeal and is therefore a further reflection of the cluster profile.
Opinim/s: The statement that clothes of yesteryear were better than today's received a strong positive confinnation.Neutral res!XJnses characterized cluster 3 towards change in the new South Africa, the complexities of life and scepticism towards change.
Family orim/atioll: As in all clusters, members of cluster 3 have strong fami ly ties and place a high premium on family gatherings and family relationships.
Nl'eds: Needs of finan cial security, security and self-respect are strong.as one could forecast from the sample profile.This group, however, does not enjoy spending money on clothing.which reflects the general attitude of the clothing-unconcerned.
Media usage: Watching television and specifically entertainment programmes are rated highest for this group.The clothing-unconcerned docs not spend a lot of time reading magaZines or listening to radio programmes and is neutral towards gathering clothing information !Tom these media vehicles.None or very little exposure is recorded to fashion speciality magazines and women's magazines.

CONCLUSION AND IMPLICATIONS
Clothing marketers need to comprehend that not all mature consumers are similar in their cloth ing involvement, clothing orientation, lifestyle and media usage.It is therefore justifiable to segment the elderly market and to ta rge t a specific market segment by using a unique marketing and promotional mix , 1992, pp. 242-262).
TIle Clothillg Moderates : 11le garments targeted at th is segment should reflect a stylish conservative look without being a fashion risk.The gathering of clothing information is done through magazines and radio, and promotional material should therefore be channelled through these mediums.Very little opinion leadership is noted.The lifestyle that should be portrayed in campaigns is an active, socially integrated individual who enjoys nature, needlework and community activities.
TIle Clotliillg Ellfllll siasls: 'This market segment is most probably the easiest to market to as the women are enthusiastic about clothing and can be reached through fashion magazines.They buy quality items and do not see price as a persuasive power in deciding to buy or not to buy.Therefore advertising messages can be focused on the quality and desirability of the ganncnts.The portrayal of lifestyle should focus on the concern for the community, nature and health.
TIl e ClotJlillg-Ullcollcemed: This market segment needs to be converted to ensure a grmving percentage of mature consumers who fall into the two above-mentioned categories.
To encourage this conversion the marketing effort should reflect clothing as an integral part of dressing stylishly and portraying a healthy lifestyle.This market segment can be reached through te levision entertainment programmes as they have thc highest viewership.The endorsement or testimonial approach portraying one of thc popular characters of a television programme migh t be successful as they can fOITll thc reference group for Ihe dothing-unconcerned.
Mature female clothing shoppers sC<!m to hold similar clothing and clothi ng store cri teria for evaluation, as these va ri ables did not significantly differ among thc respective groups of mature female consumers.

LI MITATIONS AND RECOMMENDATIONS FOR FURTHER RESEARC H
This study could be used as a point of departure for further research on the mature clothing consumer in South Africa.
There are, however, certain limitations which should be kept in mind when generalizing and interpreting thc find ings of this study.A reliltively small silmple from il restricted geographic areil hilS been studied.Underprivileged b'fOUPS wi th low incomes ilnd limited eduC<lti on were underrepresented.This is also the case wi th reference to the severely handicapped and the frilil elderly.Lumpkin e t aI., (1985, p. 79) refer to this problem.Another vie\"'J'Oint is thilt the reliltively better educated and more ilffluent elderly represent the active consumers.Marketers and milnufacturers are usually more concerned or interested in the ilctive consumer.A study of the needs ilnd problems regilrding the clothing of underprivileged groups in Stellenbosch has already been conducted by Visser, Viln Wyk. Senekal and Uys (1993).
In-depth reseilrch is needed to clilrify the extent to which mature consumers use eVilluative criteriil for clothing and clothing stores as well as whilt these aiteriil ilre.Reseilrch by Lumpkin et al. (1985) to ilSsesS the importance of store il tt ri butes to mature consumers ilnd by CassiU ilnd Drake (1987) to identify females' clothing selection criteriil could be of value in this regard.
Segmenting the mature market by clothing involvement and explaining the relationships to consumer chilracteristics and shopping orientation are also of importance.Research on the milture consumer and fashion is scant ilnd the need for research is cleilr (Visser, 1994).The results of this research indicate thilt clothing involvement ilnd clothing orientation Ciln be meilsured successfully by means of inventories developed ilnd tested abrOild.