The Socially ConsciousConsumer

W. THOMAS ANDERSON, JR.

and

WILLIAM H. CUNNINGHAM

Who are the socially conscious consumers?This article typologically classifies

socially conscious consumers and evaluatesthe relative sensitivity of demographic

and sociopsychological variables indiscriminating degree of social

consciousness.

• ABOUT THE AUTHORS.

W. Thomas Anderson, Jr. and William H. Cunninghamare assistant professors of marketing administrationin the College of Business Administration, The Univer-sity of Texas at Austin.

Journal ot Marketing. Vol. 36 (July, 1972), pp. 23-31.

A CENTRAL question confronting corporate man--^agement today concerns the viability of socialactivism as a short-run marketing strategy. Somecontinue to see the requirements of profitabilityand of social action as essentially irreconcilable.'Social activism, in their view, simply does notyield a high rate of return in any conventionaleconomic sense, or falls more properly within thedomain of governmental or regulatory responsi-bility. Others, however, argue that with furtheramplification in the demands for social and en-vironmental responsibility the cost to the firm ofignoring the social and environmental context inwhich it operates may not be profit; the cost maywell be survival.2

Thus, the issue has shifted from one of cor-porate social responsibility to a more conventionalmarket segmentation problem: Which consumersconstitute the market for products, services, or

1. Milton Friedman, "The Social Responsibility of Busi-ness is to Increase Its Profits," The New York TimesMagazine (September 13, 1970), pp. 32-33, 123.

2. See Lee Adler, "Symbiotic Marketing," Harvard Busi-ness Review. Vol. 44 (November-December, 1966). pp. 59-72; Robert W. Austin, "Responsibility for Social Change,"Harvard Business Review. Vol. 43 (July-August, 1965), pp.45-52; Daniel Bell, "The Corporation and Society in the197O's," The Public Interest. Vol. 24 (Summer, 1971), pp.5-32; John Davenport, "Bank of America is Not for Burn-ing," Fortune, Vol. 83 (January, 1971), pp. 90-93, 152; E. T.Grether, "Business Responsibility Toward the Market,"Ccdifomia Management Review. Vol. 12 (Fall, 1969), pp.33-42; Philip Kotler and Gerald Zaitman, "Social Market-ing: An Approach to Planned Social Change," JOURNAL OFMARKETING, Vol. 35 (July, 1971), pp. 3-12; Robert J. Lavidge,"The Growing Responsibihties of Marketing," JOURNAL OFMARKETING, Vol. 34 (January, 1970), pp. 25-28; WilliamLazer, "Marketing's Changing Social Relationships," JOUR-NAL OF MARKETING, Vol. 33 (January, 1969), pp. 3-9; TheodoreLevitt, "Why Business Always Loses," Harvard BusinessReview. Vol. 46 (March April, 1968), pp. 81-89; Sidney J.Levy and Philip Kotler, "Beyond Marketing: The Further-ing Concept," California Management Review. Vol. 12(Winter, 1969), pp. 67-73; Rodman C. Rockefeller, "TurnPublic Problems to Private Account," Harvard BusinessReview. Vol. 49 (January-February, 1971), pp. 131138; andDow Votaw and S. Prakash Sethi, "Do We Need a NewCorporate Respyonse to a Changing Social Environment?"California Management Review. Vol. 12 (Fall, 1969), pp.3-31.

23

24 Journal of Marketing, July, 1972

other corporate actions that promote social and/orenvironmental well-being? Who, in other words,are the socially conscious consumers? This studyprovides a partial answer to these questions.

Research Objectives

Markets are currently segmented principally onthe basis of demographic and behavioral vari-ables.3 Personality or sociopsychological attri-butes, however, have received increasing empha-sis as potentially more sensitive criteria for mar-ket segmentation, prompting research into therelative effectiveness of demographic, behavioral,and sociopsychological attributes in distinguish-ing product preference or choice. Results havebeen mixed.'*

With growing consumer sensitivity to social andenvironmental problems, market segmentationbased on consumers' societal orientation isemerging; markets will be evaluated [increas-ingly] according to the degree to which con-sumers accept the consumer-citizen conceptand buy as individuals concemed not only withtheir personal satisfactions, but also with so-cietal [and environmental] well-being. . . It isthrough the analysis of [social and] environ-mental developments and through new market-ing policies that management responds to thepressures and opportunities presented by social/environmental change.'Several studies have focused on consumer be-

havior in response to corporate actions in pollu-tion abatement or reclamation of wasted humanresources. Kassarjian's research, for example, re-vealed that "with a good product based on ecolog-ical concems, the potential for a marketer seemsto be impressive." Consumers in general appeared

3. Eugene J. Kelley, "Marketing's Changing Social/En-vironmental Role," JOURNAL OF MARKETING. Vol. 35 (July,1971), p. 1.

4. See Robert P. Brody and Scott M. Cunningham, "Per-sonality Variables and the Consumer Decision Process,"Journal of Marketing Research, Vol. 5 (February, 1968),pp. 50-57; James F. Engel, David T. Kollat, and Roger D.Blackwell, "Personality Measures and Marketing Segmen-tation," Business Horizons, Vol. 12 (June, 1969), pp. 61 70;Franklin B. Evans, "Psychological and Objective Factorsin the Prediction of Brand Choice: Ford versus Chevro-let," Journal of Business, Vol. 41 (October, 1968), pp. 445-459; Franklin B. Evans and Harry V. Roberts, "Fords,Chevrolets, and the Problem of Discrimination," Journalof Business, Vol. 36 (April, 1963), pp. 242-249; Robert Fer-ber, "Brand Choice and Social Stratification," QuarterlyReview of Economics and Business. Vol. 2 (February,1962), pp. 71 78; Ronald E. Frank, William E. Massy, andThomas M. Lodahl, "Purchasing Behavior and PersonalAttributes," Journal of Advertising Research, Vol. 2 (De-cember, 1969), pp. 15-24; and Stuart U. Rich and SubhashC. Jain, "Social Class and Life Cycle as Predictors ofShopping Behavior," Journal of Marketing Research, Vol.5 (February, 1968), pp. 41-49.

5. Same reference as footnote 3.

to be willing to try a pollution-reducing gasolineat premium prices.®

Henion echoed these conclusions:It would appear that a latent demand for [eco-logically relevant] buying information existsand that when it is presented [even] passively,buying behavior may be predictably modified.^Kassarjian further noted, however, that "it

[was] apparent from [the] study that there isno simple segmentation variable other than theattitude [toward pollution abatement] itself."«Demographic and sociopsychological variablesproved uniformly weak in discriminating degree ofconsumer concem over environmental pollution.

The important variable of concem to the mar-keter is not related to the usual segmentationcriteria, but rather the level of concem aboutthe issue at hand, whether it be nonreturnablebottles, high-phosphate detergents, aluminumcans, or excessive use of paper bags dispensedat supermarkets.^Although these and other studies confirm that

consumers differ in degree of concem over pollu-tion and social inequality, none is sufficient toallow market segmentation on the basis of demo-graphic and/or sociopsychological attributes orconsumption behavior associated with social andenvironmental consciousness. Indeed, Kassarjian'sstudy offers only discouragement on this point.Hence, the major objective of the present studywas to determine the extent to which consumerswho differ by degree of social consciousness maybe distinguished by selected demographic cind so-ciopsychological attributes, in order to provide afoundation for market segmentation and criteriato gauge the probable effectiveness of altemativemarketing strategies.

Given the above objective, the following re-search hypotheses were felt to be meaningful:

1. Consumers exhibiting a high degree of socialconsciousness differ significantly from con-sumers who do not on selected demographicattributes.

2. Consumers displaying a high degree of socialconsciousness differ significantly from con-sumers who do not on selected sociopsycho-logical attributes.

ProcedureSampling Frame

In April, 1971, a self-administered questionnairewas mailed to a random sample of 1,200 Austin,

6. Harold H. Kassarjian, "Incorporating Ecology intoMarketing Strategy: The Case of Air Pollution," JOURNALOF MARKETING, Vol. 35 (JuW, 1971), p. 65.

7. Karl E. Henion, "The Effect of Ecologically RelevantInformation on Detergent Sales," Journal of MarketingResearch, Vol. 9 (February, 1972), p. 14.

8. Same reference as footnote 6.9. Same reference as footnote 6.

The Socially Conscious Consumer 25

Texas households developed from the 1971 Austinmetropolitan area telephone directory; 412 ques-tionnaires were completed and returned in use-able form, with the lowest socioeconomic strataslightly underrepresented. The survey instrumentcontained questions not relating to the dependentand independent variables reported in the re-search. Subjects were told they were participat-ing in a study sponsored by the Department ofMarketing at The University of Texas. Anonymityof responses was assured. Given the scope of thequestionnaire and the time and financial con-straints under which the research was conducted,a mail survey was considered the only viableoption.

Dependent Variable

The dependent variable consisted of the eight-item Social Responsibility Scale developed byBerkowitz and Daniels and further tested byBerkowitz and Lutterman.'" The scale is reportedto measure an individual's traditional social re-sponsibility; i.e., the willingness of an individualto help other persons even when there is nothingto be gained for himself. Berkowitz and Lutter-man's research indicates that individuals whoscore high on the Social Responsibility Scale aremore likely to (1) make financial contributionsto religious and educational institutions; (2) beactive in community, church, or other organiza-tions or activities; (3) show intense interest innational and local political events; and (4) votein elections, and know the names of contendingcandidates. However, subjects who score high onthe test are also more likely to oppose govern-ment intervention in unemployment problems andare more likely to be opposed to extending socialsecurity.'1 Therefore, it can be assumed that so-cially conscious individuals, whose orientationsare reflected in a variety of socially responsiblebehaviors, would manifest social consciousness inconsumption decisions. Hence, it is reasonable toexpect that socially conscious individuals wouldbe more sensitive to and more likely to purchaseproducts geared toward the enhancement of socialor environmental welfare.

The Berkowitz-Daniels Social ResponsibilityScale required subjects to indicate their extent ofagreement along a five-point continuum from"strongly agree" to "strongly disagree" with each

of the below listed items. The socially responsibledirection is indicated in parentheses.

1. It is no use worrying about current eventsor public affairs; I can't do anything aboutthem anyway. (Disagree)

2. Every person should give some of his timefor the good of his town or country. (Agree)

3. Our country would be a lot better off if wedidn't have so many elections and peopledidn't have to vote so often. (Disagree)

4. Letting your friends down is not so bad be-cause you can't do good all the time foreverybody. (Disagree)

5. It is the duty of each person to do his jobthe very best he Ccin. (Agree)

6. People would be a lot better off if they couldlive far away from other people and neverhave to do anything for them. (Disagree)

7. At school I usually volunteered for specialprojects. (Agree)

8. I feel very bad when I have failed to finisha job I promised I would do. (Agree)

The Berkowitz-Daniels scale was derived in partfrom a personality scale developed by Harris,which was highly similar to a scale developed byGough, McClosky, and Meehl.i- The Harris sc:alewas designed to contrast the attitudinal responsesof school children who had a reputation for so-cially responsible behavior with students who didnot have such a reputation. The Gough et al.scale was a social responsibility index which uti-lized items similar to the Harris scale to measuresocially responsible attitudes among high schooland college students. The responsible studentswere characterized as having a "deep concernover broader ethical and moral problems, . . . astrong sense of justice, with a rather high, butsomewhat rigid, set of self demands and stan-dards . . . and a strong and unflagging sense ofconfidence in self and in the basic rightfulnessof the larger social world."'^ The validity of theGough et al. scale is substantiated by laboratoryfindings.1*

Berkowitz and Daniels' index was constructedby combining items from the Harris scale withnew test items developed by Berkowitz andDaniels. This pool of test items was subjected toseveral item analysis tests, using college studentsas a sample. Berkowitz and Daniels then selectedeight items which were administered to 766 Wis-

10. Leonard Berkowitz and Louise R. Daniels, "Affect-ing the Salience of the Social Responsibility Nornn," Jour-nat of .Abnormal and Social Psychology, Vol. 68 (March,1964), pp. 275-281; and Leonard Berkowitz and Kenneth G.Lutterman, "The Traditional Socially Responsible Person-ality," Public Opinion Quarterly, Vol. 32 (Summer, 1968),pp. 169-185.

11. Berkowitz and Lutterman, same reference as foot-note 10.

12. Dale B. Hards, "A Scale for Measuring Attitudes ofSocial Responsibility in Children," Journal of Abnormaland Social Psychology, Vol. 55 (November, 1957), pp. 322-326; and Harrison G. Gough, Herbert McClosky, and PaulE. Meehl, "A Personality Scale for Social Responsibility,"Journal of Abnormal and Social Psychology, Vol. 47 (Janu-ary, 1952), pp. 73-80.

13. Gough, McClosky, and Meehl, same reference as foot-note 12, at p. 77.

14. Berkowitz and Lutterman, same reference as foot-note 10, at p. 174.

26 Journal of Marketing, July, 1972

consin adults by the University of Wisconsin Sur-vey Research Laboratory. Analysis .of the dataindicated that the scale was internally quite con-sistent.'-'' The social responsibility test was scoredin a Likert manner. Respondents were classifiedas belonging to the upper or lower half of thesample according to their test scores.'^

Independent Variables

The independent variables consisted of six dem-ographic and six sociopsychological variables. Thedemographic variables consisted of (1) occupa-tion of the household head; (2) total family in-come for 1970; (3) education of the householdhead; (4) family socioeconomic status (theweighted average of occupation, education, in-come);' ' (5) age of the household head; and (6)stage in the family life cycle (indexed by the ageof oldest child). These were selected because theyrepresent generally accepted demographic seg-mentation criteria. The following sociopsycho-logical variables were included: (1) Alienation—a feeling of isolation from one's community, so-ciety, and or culture;"* (2) dogmatism—one's de-gree of open- or close-mindedness;'^ (3) conserva-tism—one's adherence to traditional attitudes andvalues;-'^ (4) status consciousness—a concern forsocial recognition, esteem, or prestige;^! (5) cos-mopolitanism—a global, nonparochial perspectiveand orientation;" and (6) personal competence—

15. Berkowitz and Lutterman, same reference as foot-note 14.

16. Precedents for dichotomizing the respondents in thismanner and for using linear discriminant analysis to an-alyze the dichotomized data may be found in John Har-vey, "What Makes a Best Seller?" in Motivation and Mar-ket Behavior, Robert Ferber and Hugh G. Wales, eds.(Homewood, Illinois: Richard D. Irwin, Inc., 1958), pp.361-381; in Robert Hogan, Donald Mankin, John Conway,and Sherman Fox, "Personality Correlates of Undergradu-ate Marijuana Use," Journal of Consulting and ClinicalPsychology, Vol. 35 (August, 1970), pp. 58-63; and in W. T.Tucker and John J. Painter, "Personality and ProductUse," Journal of Applied Psychology, Vol. 45 (October,1961), pp. 325-329.

17. U. S. Bureau of the Census, U.S. Census of Popula-tion: 1960. Subject Reports. Socioeconomic Status. FinalReport PC (2)-5C (Washington, D.C: U.S. GovernmentPrinting Office, 1967).

18. Russell Middleton, "Alienation, Race, and Educa-tion," American Sociological Review, Vol. 28 (December,1963), pp. 973-977.

19. Verling G. Troldahl and Fredric A. Powell, "A Short-form Dogmatism Scale for Use in Field Studies," SocialForces, Vol. 44 (December, 1965), pp. 211-214.

20. Herbert McClosky, "Conservatism and Personality,"American Political Science Review, Vol. 52 (March, 1958),pp. 2745.

21. Walter C. Kaufman, "Status, Authoritarianism, andAnti Semitism," American Journal of Sociology, Vol. 42(January, 1957), pp. 379-382.

22. Thomas R. Dye, "The Local-Cosmopolitan Dimensionand the Study of Urban Politics," Social Forces, Vol. 41(March, 1963), pp. 239-246.

a feeling of mastery of one's personal life andenvironment.^''

The sociopsychological variables ranged froma six-item alienation test to a ten-item dogmatismtest. Each of the six sociopsychological scaleswas scored in a Likert manner. These variableswere selected because several had been previouslyemployed in consumer behavior research^^ andappeared to the authors to be significantly relatedto one's likely level of social consciousness.

Data Analysis

Linear discriminant analysis was used to analyzethe data.25 This technique permits the analyst toexamine a set of independent variables to deter-mine which, if any, are able to distinguish be-tween two or more predetermined dependent vari-ables or classification categories. This is accom-plished by letting the individual's discriminantscore be a linear function of the independent vari-ables and then classifying the respondent as be-longing to one of the categories based on his re-spective discriminant score and the discriminantclassification boundary.^*

A strong upward bias, which results in an over-stated percentage of the respondents correctlyclassified, develops with discriminant analysis ifthe discriminant coefficients from one sample areused to compute the discriminant scores from thesame sample. This problem can be alleviated bydividing the sample into two subsamples. Thediscriminant coefficients are derived from the firstsubsample and are used to compute the discrimi-nant scores from the second subsample.^^

The predictive power of the several sets of in-dependent variables was tested by drawing a ran-dom subsample of 60% of the 412 respondents.From this sample discriminant coefficients weredeveloped which were then used to compute thediscriminant scores for the remaining 40% of the

23. Angus Campbell, Paul E. Converse, Warren E. Miller,and Donald E. Stokes, The American Voter (New York:John Wiley and Sons, Inc., 1960).

24. See Jacob Jacoby, "Personality and Innovation Prone-ness," Journal of Marketing Research, Vol. 8 (May, 1971),pp. 244-247; William H. Cunningham and William J. E.Crissy, "Market Segmentation by Motivation and Atti-tude," Journal of Marketing Research, Vol. 9 (February,1972), pp. 100-102; Douglas S. Longman and Henry O. Pru-den, "Alienation from the Market Place: A Study in Black,Brown and White," in Combined Proceedings 1971 Springand Fall Conferences, Fred C. AUvine, ed. (Chicago, 111.:American Marketing Association), pp. 616^19.

25. Donald J. Veldman, FORTRAN Programming for theBehavioral Sciences (New York: Holt, Rinehart, and Win-ston, 1%7), pp. 268-279.

26. Donald G. Morrison, "On the Interpretation of Dis-criminant Analysis," Journal of Marketing Research, Vol.6 (May, 1969), pp. 156-163.

27. Ronald E. Frank, William F. Massy, and Donald G.Morrison, "Bias in Multiple Discriminant Analysis," Jour-nal of Marketing Research, Vol. 2 (August, 1%5), pp. 250-258.

Tlie Socially Conscious Consumer 27

sample. In order to avoid the problem of <in un-representative random sample, this process wasrepeated four times for each set of independentvariables. The results of each replication are re-ported in Table 1. This table will be discussedin greater detail later in the article.

The demographic variables were examined first,the sociopsychological variables second, and thecombined demographic and sociopsychologicalvariables third. It was felt that an independentexamination of the demographic and sociopsycho-logical variables would make it possible to derivesome conclusions as to the research hypotheses.

Findings

Demographic Variables

The demographic variables were able to differ-entiate between the high and the low socially re-sponsible consumers. As Table 2 indicates theindependent variables "occupation of the house-hold head," "socioeconomic status," and "age ofthe household head" yielded consistently strongcorrelations with the discriminant scores for eachof the four replications. The correlation coeffi-cients show only how strongly the particular vari-ables relate to the discriminant score. They donot indicate whether a particular independentvariable is capable of discriminating between therespondent groups. Rather they merely indicatea measure of the part the respective variablesplay in whatever discrimination is achieved. Thedemographic variables were scaled such that acorrelation of > 0 implies that the variable isassociated with high social responsibility, whilea correlation of < 0 implies that the variable isassociated with low social responsibility.

The F-ratio column in Table 2 shows the resultsof an F-test of the differences between groupmeans for each of the six demographic variables.This is a test of individual attributes rather thana multivariate test. The same demographic vari-ables which had high correlations with the dis-criminant scores found statistically significant dif-ferences between the two respondent groups. Theonly exception was the variable "socioeconomicstatus" in the second replication which was notstatistically significant at the .05 level. The highsocially responsible group was characterized byhigher occupational attainment and socioeconom-ic status, and by younger household heads thanwas the low socially responsible group.

The results of the four cross-validated replica-tions of the discriminant analysis which were per-formed on the six demographic variables are pre-sented in Table l-I. The second column in Tablel-I gives the results of an F-test of the Wilks'lambda value, which is a test of the ability of thediscriminant function to significantly differentiatethe high and low socially responsible groups.

Statistically, Wilks' lambda takes the followingform:

A = |W| / |T|where W represents the pooled within-group devi-ation score cross-products matrix and T repre-sents the total sample deviation score cross-prod-ucts matrix.28 The first, second, and fourth repli-cations were significant at the .05 level, while thethird replication was significant at the .01 level.The Wilks' lambda tests were based on the 60%of the sample which was drawn independentlyfour times to determine the discriminant coeffi-cients for each replication.

Since the respondent groups were of equal sizeit would have been expected that 50% of the sub-jects in the cross-validated samples would havebeen correctly classified by chcince. The thirdcolumn in Table l-I presents the results of a t-testof the significance of the percent of the subjectscorrectly classified. Because the respondents inthe cross-validated samples were not included inthe 60% samples which made up the discriminjintcoefficients, the data is effectively cross validated.

The second and third cross-validated replica-tions were significant at the .01 level, while thefourth replication was significant at the .05 level.The first replication was not statistically signifi-cant. The findings of the t- and F-tests wouldseem to indicate that the demographic variableswere able to differentiate between the high andlow socially conscious respondents.

Sociopsychological Variables

The sociopsychological variables appear to bemore effective in differentiating between the highand low socially responsible consumers th<in werethe demographic variables. As Table 3 shows,dogmatism, conservatism, status consciousness,and cosmopolitanism all had strong correlationswith the discriminant scores. The sociopsycho-logical variables were scaled such that a correla-tion of > 0 implies that the variable is associatedwith high social responsibility, whereas a correla-tion of < 0 is associated with low social responsi-bility. Alienation and personal competence werenot as strongly correlated with the discriminantscores as were the other sociopsychological vari-ables.

The second column in Table 3 shows that eachof the six sociopsychological variables was ableto differentiate significantly between the high andthe low socially responsible groups. The F-testof the difference in group means indicates thatdogmatism, conservatism, status consciousness,and cosmopolitanism found significant differencesat the .001 level, while personal competence wassignificant at the .01 level, and alienation at the

28. William W. Cooley and Paul R. Lohnes, h4ultivariateProcedures for the Behavioral Sciences (New York: JohnWiley and Sons, Inc., 1962), p. 61.

28 Journal of Marketing, July, 1972

F-TEST AND

Repli-cation

1

23

4

AverageCorrectClassifi-cation

•Significant"Significant"Significant

F-ratio2.52'

2.65*

3.30"

2.64"

at theat theat the

T-TEST AND DISCRI.MINANT

/

DemographicVanables

PercentCorrectlyClassified

55.5%

62.8%

60.4%

58.5%

59.4%

.05 level

.01 level

.001 level

Valueof t"

1.42

329"

2.67*

2.19-

-t

TABLE 1ANALYSIS OF DEMOGRAPHIC

F-ratio

7.42'

5.62'

9.44'

724'

/ /

SociopsychologicalVariables

PercentCorrectlyClassified

62.8%

72.6%

70.7%

71J%

695%

AND SOCIOPSYCHOLOGICAL VARIABLES

Valueof t"

329"

5̂ 80*

5J2«

5.4«'

= proportion correctly classified -

J .5 (1 - .5)

F-ratio

4.64'

3.68'

5.84'

423'

-.5

/ / /Demographic PlusSociopsychological

Variables

PercentCorrectlyClassified

555%

72.6%

732%

62.8%

66.0%

Valueof t*

1.42

5.80«

5.96'

329»

TABLE 2

CORRELATION COEFFICIENTS WITH THE DISCRIMINANT SCORE AND F-TEST FOR THE DEMOGRAPHIC VARIABLES

Variables

Occupation of thehousehold head

Annual familyincome

Education of thehousehold head

Socioeconomicstatus

Age of thehousehold head

Stage in thefamily life cycle

Run 1

CorrelationCoefficient

.79

.30

.45

M

-.58

-.19

F-ratio

9.34"

1.35

3.03

422"

4.95"

.55

Run

CorrelationCoefficient

.72

24

Al

.48

-.70

-.14

2

F-ratio

8.19"

.88

237

3.60

7.68"

28

Run

CorrelationCoefficient

£2

32

56

.55

-.72

.11

3

F-ratio

7.40"

1.92

6.08'

5.73'

10.01"

21

Run

CorrelationCoefficient

ja

M

J9

.60

- .84

-.09

4

F-ratio

5.98'

222

234

5.56"

1139"

.13'Significant at the .05 level"Significant at the .01 level

TABLE 3CORRELATION COEFFICIENTS WITH THE DISCRIMINANT SCORE AND F TEST FOR THE SOCIOPSYCHOLOGICAL VARIABLES

Variables

Alienation

Dogmatism

Conservatism

Status Con-sciousness

PersonalCompetence

Cosmopolitanism

Run 1

CorrelationCoefficient

- .58

-.80

-.65

-.65

-.42

.76

F-ratio

13.48'

27.47'

17.30'

1727'

7.04"

24.51'

Run

CorrelationCoefficient

- .39

-.66

- .73

- .71

- .56

.79

2

F-ratio

4.51*

14.04'

17.67'

16.39'

9.95"

20.84'

Run

CorrelationCoefficient

- .59

-.85

-.64

-.65

-.46

.66

3

F-ratio

17.92'

39.67'

20.38'

2124'

10.34"

21.87'

Run 4

CorrelationCoefficient

- .48

- .81

- .71

- .65

- .46

.83

F-ratio

8.99"

26.97'

20.65'

16.96'

822"

28.68'

•Significant at the .05 level"Significant at the .01 level"^Significant at the .001 level

The Socially Conscious Consumer 29

TABLE 4

CORRELATION COEFFICIENTS WITH THE DISCRIMINANT SCORE ANDF-TEST FOR THE DEMOGRAPHIC AND SOCIOPSYCHOLOGICAL VARIABLES

Variables

Occupation of thehousehold head

Annual familyincome

Education of thehousehold head

Socioeconomicstatus

Age of thehousehold head

Stage in thefamily life cycle

Alienation

Dogmatism

Conservatism

StatusConsciousness

PersonalCompetence

Cosmopolitanism

Run

CorrelationCoefficient

.44

.17

25

52

-32

-.10

-.72

-.57

-59

- .38

-.69

JO

1

F-ratio

9.33'

1.35

3.02

13.48'

4.95'

.55

27.47'

17.30'

1727'

7.05"

24.51'

422*

Run

CorrelationCoefficient

.45

.15

.26

.34

- .44

- .08

- .58

-.65

-.62

-.49

-.70

.30

2

F-ratio

8.19"

.88

2.58

4.51*

7.67"

28

14.03'

17.68'

16.39'

9.95"

20.84'

3.60

Run

CorrelationCoefficient

.36

.19

.33

.54

-.41

.06

-.78

-.58

-59

-.42

-.60

.32

3

F-ratio

7.40"

1.92

6.08*

17.93'

10.01"

21

39.67'

20.39'

2124'

10.35"

21.87'

5.72'

Rtm 4

CorrelationCoefficient

37

23

23

.45

- .50

.05

-.74

-.66

-.60

-.43

-.76

35

F-ratio

5.98*

222

233

8.99*

1139»

.12

26.97'

20.65'

16.96'

822*

28.69'

536*

•Significant at the"Significant at theSignificant at the

.05 level

.01 level

.001 level

.05 level. The high socially responsible group wasmore cosmopolitan, but less alienated, less dog-matic, less conservative, less status conscious, andless personally competent than was the low so-cially responsible group.

Table l-II presents the results of the discrimi-nant analysis of the sociopsychological variables.As column one indicates, each replication's F-testof the respective Wilks' lambda value was sig-nificant at the .001 level. This would tend to indi-cate that the discriminant function was able todifferentiate between the two respondent groups.

The second column in Table MI indicates thatthe percentage of the cross-validated sample cor-rectly classified tended to be higher when thesociopsychological variables were used to distin-guish the groups than when the demographic vari-ables were used. The mean percent correctlyclassified for the sociopsychological variables was69.5, while the mean percent correctly classifiedfor the demographic variables was 59.4. The t-testof the percent correctly classified for the socio-psychological variables was significant at the .01level for the first replication and at the .001 levelfor the second, third, and fourth replications.

Sociopsychological and Demographic Variables

The sociopsychological and demographic vari-ables combined were not as effective in differen-

tiating high from low socially responsible groupsas were the sociopsychological variables alone.Table 4 lists the correlations of the variables withthe discriminant function and the F-test of thedifferences between the group means. It is ap-parent that the same variables which correlatedhighly with the discriminant scores in the previ-ous tests correlated highly with the discriminantscore when all 12 variables were utilized. Thestatistical differences between group means re-mained essentially unchanged from the previoustests.

Table MI I shows that the F-ratio tests of theWilks' lambda values were significant at the .001level for each of the four replications. The meanpercent correctly classified for the combined setsof independent variables was 66.0, which was lessthan when only the sociopsychological variableswere utilized. This was expected because it waspreviously found that the demographic variableswere not as effective in differentiating the highand low socially conscious respondents as were thesociopsychological variables. That is, combiningboth categories of independent variables resultedin a loss in the power of sociopsychological varia-bles to differentiate between high and low sociallyconscious consumers. The t-test of the significanceof the percentage correctly classified was signifi-cant at the .001 level for the second and third

30 Journal of Marketing, July, 1972

TABLE 5RELATIONSHIP BETWEEN SELECTED VARIABLES AND SOCIAL CONSCIOUSNESS

VariableHigh Social

ConsciousnessLow Social

ConsciousnessOccupation of the

household headAnnual family incomeEducation of the

household headSocioeconomic status

Age of the house-Iiold head

Stage in the familylife cycle

AlienationDogmatismConservatismStatus ConsciousnessPersonal Competence

Cosmopolitanism

Higher statusoccupations

N.S.

N.S.Above average socio-

economic status

Pre-middle age

N.S.Less alienatedLess dogmaticLess conservativeLess status consciousLess personally

competentMore cosmopolitan

Lower status occupationsN.S.

N.S.Average ajid lower socio-

economic status

Middle age and older

N.S.More alienatedMore dogmaticMore conservativeMore status consciousMore personally competent

Less cosmopolitan

cross-validated replications, and at the .01 levelfor the fourth replication. The first replicationwas not significant.

Conclusions

The demographic variables analyzed providedpartial support for the first research hypothesis.Socioeconomic status was found to be signifi-cant in discriminating between high and low so-cial responsibility; social consciousness tended tovary directly with socioeconomic status. Stage inthe family life cycle, by contrast, failed to signifi-cantly discriminate respondents as to degree ofsocial responsibility. Occupation and age of thehousehold head proved to be highly sensitive dis-criminators of social responsibility, social con-sciousness varying directly with occupational sta-tus and inversely with age. Education of thehousehold head produced mixed results with onlyone instance of significant discrimination. Annualfamily income was uniformly poor as a discrimi-nator of social responsibility. In general, it wouldappear that socioeconomic status, occupation, andage of the household head provide significant dis-criminators of social consciousness.

Of the sociopsychological variables analyzed,dogmatism, conservatism, cosmopolitanism, andstatus concern proved equally effective as dis-criminators of social responsibility, tending tosubstantiate the second hypothesis. Social con-sciousness tended to vary inversely with dog-matism, conservatism, and status consciousness,and directly with cosmopolitanism. Alienation and

personal competence were only slightly less effec-tive in discriminating degree of social responsi-bility; both varied inversely with social conscious-ness. Overall, the sociopsychological variablesanalyzed yielded more sensitive discriminators ofsocial consciousness than either the demographicvariables alone or the demographic and socio-psychological variables combined.

Briefly, the image of the socially conscious con-sumer emerging from the research is that of apre-middle age adult of relatively high occupationalattainment and socioeconomic status (see Table5). He is typically more cosmopolitan, but lessdogmatic, less conservative, less status conscious,less alienated, and less personally competent thanhis less socially conscious counterpart. Alterna-tively, consumers displaying low social conscious-ness may be characterized as intermediate orlower in occupational attainment and in socioeco-nomic status, and are of middle age or older.They are characteristically more dogmatic, moreconservative, more status conscious, more alien-ated, more personally competent, and less cosmo-politan than are socially conscious consumers.

The findings of the research support the con-clusion that markets can be segmented on thebasis of consumers' social consciousness. Bothdemographic and sociopsychological attributesprovide criteria for market segmentation, al-though it would appear that sociopsychologicalvariables are more sensitive discriminators ofsocial consciousness.

The present findings suggest some directionsfor further research. It would be useful to de-

The Socially Conscious Consumer 31

termine whether consumption pattems are dif-ferent between high and low scorers on the Berko-witz-Daniels Social Responsibility Scale, particu-larly with respect to products and/or brandswhich claim environmental benefits. Specific hy-potheses between degree of social consciousnessand various aspects of buyer behavior may bederived from the sociopsychological and demo-graphic correlates of social consciousness. For ex-ample, it was found that socially conscious con-sumers appear to be open-minded, aware, and

exhibit a general orientation toward progress orchange. But are they willing to pay a higher pricefor products and services which enhance socialor environmental well-being? Does their open-mindedness and progressiveness result in recep>-tivity to new products which are compatible withthe environment? What types of promotional ajvpeals, information channels, and distribution sys-tems are best suited for this particular market?These questions remain unanswered at present,and additional research is needed.

-MARKETING MEMO'Consumer Responsibility, a Two-way Street . . .

Business, to suryive, must merit the confidence of its customers. If the con-sumer activists have done nothing else, they have aroused business men andwomen all over the country to a greater awareness of their responsibility toeliminate the shady practices which flourish on the fringes of the husiness com-munity. As husiness people, we cannot afford to close our eyes to violationsof sound husiness principles, and there is much evidence that the business com-munity is accepting increasing responsibility in this direction. . . .

But consumers, too, share the responsibility for ethical standards in themarketplace. Some "perfectly honest" people seem to think it's all right tocheat Business. Defraud the telephone company, for instance, by placing longdistance calls for fictitious names which, by pre-arrangement, convey theirmessage. There's the grocery shopper who deliberately damages vegetables;and the shoplifters, who each year, pilfer two and a half billion dollars worthof merchandise from stores in the United States. All such acts cost you andme money as Consumers. The losses from these petty crimes are necessarily partof the cost of doing business and they add to the price we all must pay for thegoods and services we buy.

Honesty and integrity thus are mandatory on both sides of the counter ifwe are to maintain a healthy, mutually beneficial business-consumer relationship.

—Mercedes S. Wood, "Business and theConsumer," The Journal of Business(published by the Bureau of BusinessResearch, Seton Hall T^niversity), Vol.10 (December, 1971), pp. 21-25, atp. 23.