Social Class and Income Influences on External Search Processes of Adolescents

Citation:

Stephen K. Keiser and Philip G. Kuehl (1972) ,"Social Class and Income Influences on External Search Processes of Adolescents", in SV - Proceedings of the Third Annual Conference of the Association for Consumer Research, eds. M. Venkatesan, Chicago, IL : Association for Consumer Research, Pages: 602-631.

Proceedings of the Third Annual Conference of the Association for Consumer Research, 1972      Pages 602-631

SOCIAL CLASS AND INCOME INFLUENCES ON EXTERNAL SEARCH PROCESSES OF ADOLESCENTS

Stephen K. Keiser, University of Delaware

Philip G. Kuehl, University of Maryland

[Both authors are assistant Professors of Marketing at their respective schools.]

INTRODUCTION

In recent years, the adolescent segment of the population of the United States has received increased attention and study from many societal institutions. For example, political parties, religious groups, educational institutions, and governmental agencies have attempted to study the impact of adolescent behavior and attitudinal structures on the performance of their tasks and activities. In a similar manner, producers of economic goods and services have recognized the importance of understanding the "adolescent market" for the following four reasons:

1. Adolescents comprise an important, unique market segment for many types of goods and services with an estimated spending level in excess of $20 billion annually (National Industrial Conference Board, 1969).

2. Adolescents are estimated to affect the expenditure of $60 billion annually through their influence on the consumption patterns of other family members (Senior Scholastic, 1965).

3. The market behavior of adolescent groups, such as shopping behavior and brand preference, affects their consumption behavior as adults.

4. Adolescent behavior has affected the consumption behavior of all members of the American society. For example, their influence has been observed in clothing and hair stoles in recent years.

One important dimension of the consumption behavior of adolescents which is of interest to researchers and practitioners are the processes by which adolescents acquire and utilize information about their environment. Unfortunately, previous research on the information search behavior of adolescents is limited in the following manner:

1. Research has been limited to the investigation of only one or two information sources (Ward & Robertson, 1970; Bowerman & Rinch, 1959; and Editor and Publisher, 1966). As a result the information obtained from these studies is of little use in explaining the relationships between the use of all possible sources of information.

2. Many of the previous studies have concerned aspects of adolescent behavior other than consumption behavior (Remmers & Radler, 1957).

3. Past research has treated adolescents as a homogeneous group. As a result the differences in the consumption behavior of adolescents with different economic resources or social status have been ignored (Gilkinson, 1965).

The general objective of this study, then, is to examine the processes by which information about economic goods and services-is obtained by adolescent consumers. These processes are referred to as external search behavior. The nature of these processes are examined from the viewpoint of (a) the relations between different information sources and (b) the effect of the socio-economic characteristics of adolescents on search behavior. Two characteristics of consumers that have been shown to influence external search processes are-social class and income. The strong research tradition of these two predictor variables suggests that they should be effective predictors of external search behavior of adolescents

RESEARCH OBJECTIVES

The objective of this study is to investigate the extent to which two predictor variables, income and social class, are related to the external search processes of the adolescent segment of American society. In specific terms, the research focused on:

1. Examining whether or not external search processes can be understood in terms of differences between adolescents with different incomes.

2. Examining whether or not external search processes can be understood in terms of differences between adolescents in different social class categories .

As indicated above, income and social class are the two predictor variables investigated in this study. Using the Engel, Kollat, and Blackwell conceptual model as a framework for the research, the following sources of information served as the major criterion variables in this study:

1. Mass media sources: including the degree of utilization of broadcast and print media sources; and, the relationship of brand awareness to external search processes.

2. Personal sources: including the examination of opinion leadership, family, and reference groups as sources of information.

3. Marketer-dominated sources: including the impact which retail store shopping behavior has on external search activity.

The degree of utilization and importance of these information sources for adolescents constitutes the major empirical thrust of this study. In order to accomplish these objectives, a questionnaire was designed, pretested, and administered to a sample of 1,200 junior high and senior high students in Columbus, Ohio in March, 1971.

LIMITATIONS OF THE STUDY

The use of Columbus, Ohio students as the sampling frame represents two obvious limitations of the study in terms of generalizing these results to adolescents in other geographic areas. First, the researchers did not attempt to include (a) adolescents not enrolled in Columbus, Ohio secondary schools, or (b) students absent on the day the questionnaire was administered to the sample. In this respect, the sample tends to be representative only of students enrolled and attending class on the day the questionnaire was administered. However, the questionnaire was administered on a tay when absenteeism tented to be lowest (Thursday) and State of Ohio law requires that adolescents must be enrolled in school until age 17. These considerations tend to minimize the seriousness of this first limitation.

The second limitation in the study evolves from the fact Columbus, Ohio is not representative of the national population. However, age, income, educational, and occupational characteristics of Columbus, Ohio are similar to those of the United States as a whole. As a result, a carefully constructed sample of Columbus, Ohio adolescents does tend to be somewhat representative of adolescents in general.

Another limitation of this study evolves from the nature of the research design used in the research. Inasmuch as the research design used in the present study was cross-sectional, rather than longitudinal, the results of the study should be used cautiously when predicting changes in behavior overtime.

Finally, the results cannot be generalized to specific products because length constraints on the questionnaire eliminated the opportunity to ask adolescents about their external search processes for specific products. As a result, respondents were asked about "things they buy" rather than specific products.

CONCEPTUAL FRAMEWORK

This section discusses the conceptual framework and variables used in the study. The Engel, et al. consumer decision-making model, which was adopted as the conceptual framework, is discussed first. External search behavior is then defined and discussed. An explanation of the dimensions of external search behavior examined in the study are included in this discussion. Next, income and social class, the predictor variables, are operationally defined. This is followed by an explanation of the means by which the different research variables were measured.

CONSUMER DECISION MAKING AS A SYSTEM

As the consumer behavior "research tradition" develops and gains in methodological sophistication, the need for conceptual models to guide such research becomes apparent. In order to incorporate the results of the present study within the context of the existing consumer behavior "research tradition", the Engel et al. consumer decision-making model was adopted as the conceptual framework

Their model, as shown in Figure I, describes the consumer decision process as a system with outputs that result from the processing of received inputs. This cognitive processing of environmental inputs is directed by the component labelled the "central control unit" which is the memory and thinking of the consumer. The output of consumer cognitive processes is visualized as consisting of five "decision steps"; (l) problem recognition; (2) external search for alternatives; (3) evaluation of alternatives; (4) purchasing processes; and (5) postpurchase behavior. In the present study, the external search for alternatives stage of the decision process is examined in relation to the effects of the adolescents income and social class on the central control unit.

External Search Behavior--Defined

External search behavior is defined in this study as:

...processes and activities whereby the consumer uses various sources of information, including mass media, personal sources, and marketer-dominated sources...to learn about the number of alternative solutions to the perceived problem, the characteristics and attributes of those alternatives, and their relative desirability (Engel, Kollat, & Blackwell, 1968, p. 378),

FIGURE I

A COMPLETE MODEL SHOWING OUTCOMES OF THE PURCHASING DECISION

In specific terms, the primary focus of this study is the examination of the following sources of information used by adolescents in external research.

1. Mass media sources: print and broadcast media which reach large number of consumers.

2. Personal sources: members of the extended family, friends, school teachers, and strangers who contribute information to any given consumer.

3. Marketer-dominated sources: the marketing activities of variouS firms by which they attempt to communicate with consumers, such as sales presentations by shopkeepers or clerks, in-store displays, consumer visits to retail outlets, and catalogs mailed to prospective buyers.

Figure II, on the following page, presents a detailed summary of these information sources.

Dimensions of External Search Behavior Investigated

The use of information sources in external search can take many forms. These information sources may (a) initiate the decision-making process; (b) influence the exact nature of the decision-making process such as time of purchase; or, (c) place of purchase. Several of these dimensions of external search behavior are investigated in the present study.

All Information Sources

Some of the dimensions of external search behavior investigated involved all of the categories of information sources described above. These included the use of information sources to --

a. initiate decision-making processes related to consumption. This aspect involves the sources which are the first to make the consumer aware of the product. The relative frequency with which the different information sources were used for initial consumption information was investigated.

b. to provide information subsequent to the initiation of the decision-making process. Similar to initial product information, respondents' relative use of the information sources for subsequent product information was examined.

c. provide decision rules used to solve the various aspects of the decision-making process. This role of information sources is one of "decisive effectiveness" (Engel, Kollat, & Blackwell, 1968, p. 404). The present study investigates the information sources individuals consider the most influential and the least influential.

Mass Media Sources

One aspect of external search processes investigated here relates to mass media. This aspect was measured by the amount of exposure to mass media. Mass media exposure is divided into (a) broadcast media exposure and (b) print media exposure. Broadcast media exposure is measured as the number of hours spent by a respondent on an average day watching television or listening to radio. In anticipation of variations in broadcase exposure patterns for different times during the week, a respondent"s exposure to television and radio was investigated for two time periods--weekdays and weekends. Print media exposure was defined for purposes of the study as a combination of the (a) number of newspapers and magazines read by a respondent and the (b) frequency with which he read the media. Mass media exposure provides one indicant of the use of information sources in external search.

FIGURE II

A DELINEATION OF INFORMATION SOURCES USED FOR EXTERNAL PRODUCT SEARCH BEHAVIOR

Personal Sources

Many types of personal information sources are available to adolescent consumers during the decision process. For example, one important source of information influencing the consumption of adolescents is friends. In view of the importance of friends on adolescent consumption behavior, the role of friends as opinion leaders was investigated. Opinion leaders are defined as "those to whom others turn for information and advice" (Robertson. 1970. P. 83).

Marketer-dominated Sources

Two dimensions of the role of marketer-dominated sources in external search were investigated. One of these dimensions investigated the overall effectiveness of messages sent by marketers through other information sources. For example, external search behavior may result in the learning of brands ant advertising slogans. As a result, "brand awareness" was adopted as another indicator of the nature of external search behavior. For purposes of this study, "brand awareness" is defined as: (1) the ability to correctly associate brand names and producers, and (2) correctly complete or identify the sponsor of popular advertising slogans.

The other aspect of marketer-dominated information sources investigAted involved retail store patronage. It is especially pertinent for the retailer to understand how adolescents choose one store rather than another as a means to seek information about products. As a result of the importance of store visitation patterns of adolescents to the retailer, the relative importance of the following characteristics in the store choice decision were investigated

1. Convenience of the store.

2. Trustworthiness of store personnel.

3. Friendliness of store personnel.

4. Quality of store offering.

5. Availability of latest styles.

6. Service provided by the store.

7. Assortment available.

8. Low prices.

External Search--Limitations

The processes of external search behavior are examined in this study within the following constraints.

1. No assumptions were made regarding whether the respondents in the study are involved in: a. extensive problem solving, b. limited problem solving, or c. routinized response behavior (Howard & Sheth, 1967).

2. No assumptions were made regarding the source of the consumers' problems, such as assortment deficiency, expanded consumer desires, or expanded consumer demands (Walters & Paul, 1970).

3. No assumptions were made concerning the a. perceived value of the external search, b. the appropriateness of stored information, c. the ability to recall stored information, d. the degree of perceived financial, social, or physical risks in purchasing the products (Engel, Kollat, & Blackwell, 1968, p. 382-385).

4. No assumptions were made concerning the real or perceived cost of the external search (Engel, Kollat, & Blackwell, 1968, p. 882).

5. The sources of information used by consumers in the external search process are researched within the context of the Engel et al. (1968, p. 34-53).

Income as a Variable

One of the first variables used by early researchers to explain consumer behavior was the buyer's income. In an effort to measure the adolescent's spending power regardless of its source, income of adolescents is defined as: the amount of money the adolescent perceives himself receiving during an average week from all possible sources of earnings including allowances and occupation pursuits.

Social Class as a Variable

The use of social class as a market variable was first suggested by Martineau in 1958, and it has been treated as a major market variable by many writers for the past decade. Early researchers in the use of social class claimed or implied that social class was a better variable than income to use when explaining buyer behavior. For example. Coleman wrote. concerning furniture purchases of consumers in different social classes, that: "The correlation between prices of goods purchased and social class is relatively quite high in these product areas while the correlation between price paid and annual income is lower than one might expect (Coleman, 1960). Subsequent research has cast doubt on the assumption that social class is a better predictor of consumer behavior than income (Slocum & Mathews, 1970; Rich & Jain, 1968; and Myers, Stanton, & Haug, 1971). The reason social class may be losing its predictive power is that: ", .social class distinctions have been obscured by rising income and educational levels" (Rich & Jain, 1968) Since previous studies suggest that income and social class have different effects on consumer decision making, the two variables are treated as separate independent variables in this study.

Social classes are defined for this study as: ".,.relatively permanent and homogeneous divisions in society into which individuals or families can be categorized when being compared with other individuals or families in the society (Engel, Kollat, & Blackwell, 1968, p. 264).

Conceptual Framework--Summary

In summary, the study was conducted within the Engel, et al. decision process framework and involves an investigation of the relationships between selected predictor and criterion variables, The criterion variables are:

1. income.

2. social class.

The predictor variables, which encompass different aspects of external search processes, are:

1. the use of different information sources to initiate decision-making processes related to consumption.

2. the use of different information sources to provide information subsequent to the initiation of the decision-making process.

3. the types of information sources considered to be the most and the least influential in consumption-related decisions.

4. amount of exposure to mass media.

5. opinion leadership.

6. brand awareness.

7. retail store patronage motives

Measurement of Variables

The types of questions framed to measure each of the variables are shown in Figure III on the following page. A mixture of open-ended and multiple-choice questions were employed. The exact nature of these questions is summarized in the third column of Figure III.

METHODOLOGY

Data for the sample were collected in March, 1971 through a questionnaire form completed by a sample of 1,280 Junior high and senior high school students in Columbus, Ohio. The questionnaire form was pretested in York County, Pennsylvania and Prince Georges County, Maryland, before its use in Columbus, Ohio. It was administered to the respondents with the cooperation and assistance of local school guidance counselors and administrators The questionnaire form took approximately 45 minutes (one class period) to complete.

FIGURE III

SUMMARY OF MEASURES

The respondents were selected through the use of a multi-stage sampling process using individual schools and grade sections as the primary sampling units. In this sampling process, all secondary schools in the Columbus, Ohio system were divided into junior high and senior high school stratifications. Next, five schools were randomly selected from each strata. Then each of the 10 schools included in the sample were stratified by grade level and 2 home room sections of students were randomly selected from each grade strata. The questionnaire was administered to all students present in the selected home rooms during the home room period at each school.

The data provided by the 1,280 respondents were used to accomplish the objectives of the study. In order to determine if the differences observed in the external search processes of adolescents were attributable to sampling error or to actual differences in behavior, several statistical tests were adopted. The three statistical tests used in this study were the (a) Chi-square test, (b) Wilcoxon matched-pairs signed-ranks test, and (c) Friedman two-way analysis of variance. All three tests are nonparametric tests. This type of statistical test was deemed more appropriate than parametric tests because the data were assumed to be abnormally distributed and ordinal in nature.

The Chi-square test was used to examine the relationships existing between the predictor variables and most of the criterion variables. The relative rank of retail patronage motives was the only criterion variable which was not examined with the Chi-square test. The relationships existing for this criterion variable were tested with the use of the Friedman statistic. The Wilcoxon test was used to determine if the relative variations in use of different information sources were significant.

The significance of the observed relationships for each of the test was determined by either the value of the Chi-square distribution or values of the Z distribution. The Chi-square distribution was used to determine the significance of the Chi-square and two-way analysis of variance statistics. The Z distribution was used in the case of the Wilcoxon test

An important issue involved in statistical analysis is the definition of the "region of rejection". The region of rejection is a subset of the sampling distribution that the researcher feels is so extreme that when the research hypothesis is true the probability is very small that the sample actually observed will yield a value which falls into this region. Inasmuch as there is little theory to guide the researcher in establishing the size of the rejection region the level established by research tradition, 5 percent, was adopted for the statistical analyses of this data.

RESEARCH RESULTS

As indicated in the preceding discussion, three general types of information sources (mass media, personal, and marketer-dominated) were investigated in this study. The results are organized using these three types of information sources. For each of these types of information sources, the dimensions of external search processes (initial, subsequent, and decisive impact) are discussed as well as their relationships with income and social class.

Mass Media

The first class of information sources investigated were mass media sources. As previously discussed, television, radio, newspapers, and magazines were the four specific sources included in this investigation.

TABLE I

RELATIVE FREQUENCY OF USE OF INFORMATION SOURCES FOR INITIAL PRODUCT INFORMATION

TABLE II

RESULTS FOR HYPOTHESIS H1 COMPARING RELATIVE UTILIZATION OF INFORMATION SOURCES FOR INITIAL PRODUCT INFORMATION

Initial Product Information

The average frequency with which respondents used the different information sources for initial product information were used to rank the different types of sources as shown in Table I. The average for each source was determined by scaling responses on a one (never) to five (always) scale and summing these values for all respondents. Then, this sum was divided by the size of the sample.

As indicated in Table I, television is the mass media source used most frequently for initial product information while magazines were used least frequently. In order to determine if the differences in frequency of use of different information sources were significant, two statistics, the Wilcoxon T and Z, were computed for pairs of information sources. These information sources were paired on the basis of their relative rankings reported in Table I. The values of the T and Z statistics are reported in Table II. It is concluded that television is used significantly more frequently for initial product information than other mass media sources with the Z value computed for the pair of mass media sources, television and newspaper, significant at a .001 alpha level (See Table II). Conversely, magazines were not used significantly less frequently for initial product information than was radio although the difference between magazines and the other mass media source, newspapers, was significant.

Income. The Chi-square values which resulted when the effect of income was related to the use of information sources for initial product information reported in Table III. In addition the results are reported for the criterion variable of social class in this table.

Income does not have a significant effect on respondents' utilization of mass media sources for product information. Subsequent analysis of the two sample strata, junior high and senior high school, revealed that the higher the income of junior high respondents the more they utilized magazines for initial product information (X - 14.16, level of significance = .05). In all other cases the use of mass media for initial product information was not related to the amount of respondent income.

Social class. Mass media are not utilized differently for initial product information by adolescents who are in different social classes (See Table III). This lack of relationships was also found to be the case when junior high and senior high respondents were considered separately.

Subsequent Product Information

Similar to the analysis of initial product information behavior, average frequency of use of information sources for subsequent product information was used to rank the sources, as shown in Table IV, and these rankings were paired prior to performance of the Wilcoxon test (these results are shown in Table V).

Examination of Tables IV and V result in the conclusion that television is the mass media source utilized most frequently for subsequent as well as initial product information. Similarly, magazines were utilized less frequently than any other mass media.

Income. Income of respondents was not related to the frequency of use of mass media for subsequent product information (See Table VI). When junior high respondents were isolated, income was found to be related to the frequency of use of magazines for subsequent product information in a manner similar to that observed for use of magazines for initial product information. As junior high respondents earned higher incomes their use of magazines for subsequent product information increased (X2 - 15.73, level of significance - .02). As a result magazine acquisition by junior high adolescents may depend on the amount of income the adolescent earns.

TABLE III

INFORMATION SOURCES USED FOR INITIAL PRODUCT INFORMATION AND INCOME AND SOCIAL CLASS OF RESPONDENT

Social class. As shown in Table VI social class was not significantly re lated to subsequent use of mass media for product information. This lack of significant relationships also held for both junior high and senior high respondents. As a result, the behavior of adolescents does not concur with that found for adults in the middle class who have been found to be more dependent on mass media for subsequent product information than members of other classes (Rainwater. Coleman & Handel, 1959).

TABLE IV

RELATIVE FREQUENCY OF USE OF INFORMATION SOURCES FOR SUBSEQUENT PRODUCT INFORMATION

TABLE V

RESULTS FOR HYPOTHESIS H1 COMPARING RELATIVE UTILIZATION OF INFORMATION SOURCES FOR SUBSEQUENT PRODUCT INFORMATION

Decisive Product Information

Television was found to be the mass media which had the most influence on external search behavior of adolescents. This finding agrees with the primary position found for television for initial and subsequent product information. Conversely the source indicated as the least important mass media source was not magazines (as expected from observation of initial and subsequent use of mass media sources). Instead, radio was considered the least important mass media source by more respondents than any other mass media source. As would be anticipated, fewer respondents considered television as the least important mass media source than any other mass media source.

Income. No significant relationships were found between income of respondents and the importance attached to mass media sources. As a result the general profile of the sample would seem to fit respondents with different incomes. This is also true for junior high and senior high respondents.

Social Class. Similarly a lack of significant relationships between social class and the influence of mass media lead to the conclusion that adolescents regardless of social class are most influenced by television and least influenced by radio. The significant differences were that senior high respondents in the upper class placed more importance on newspapers than those in other social classes. While senior high, lower class respondents were more influenced by radio than were other respondents.

Mass Media Exposure

One aspect of external search behavior examined in this study applies only to mass media information sources. This aspect is a measurement of the amount of exposure to mass media.

The correlation coefficients derived for each of the relationships for exposure to the different mass media are shown in Table VII. As noted above six types of mass media exposure were studied as both television and radio exposure were examined for two time periods--weekdays and weekends. An examination of Table VII shows that there is very little correlation between exposure to television and exposure to radio. Only when radio and television exposure for weekdays is considered does significant correlation result. Low correlation is also observed between print media exposure and broadcast media exposure. One difference is noted when the relationship between newspaper and magazine exposure is examined. Readership of these two types of print media is significantly correlated.

Income and social class. The exposure habits of respondents were related to their income and social class and analyzed using Chi-square tests. The results of these analyses show that exposure to broadcast media is dependent on the income but not the social class of the respondent (See Table VIII) The number of hours spent watching television and listening to radio on weekends decreased as the income of the respondents increased. The relationship found between income and television viewing of adolescents is consistent with research examining adult behavior (King & Summers, 1971). Conversely an expected relationship of increased radio listening by respondents with higher incomes only materialized in the patterns found for weekdays.

TABLE VI

INFORMATION SOURCES USED FOR SUBSEQUENT PRODUCT INFORMATION AND INCOME AND SOCIAL CLASS OF RESPONDENT

Interestingly, the significant relationships between income and television exposure were found to be true for senior high but not junior high respondents. This may represent a shift in source of adolescent income as the adolescent becomes older. Senior high adolescents may rely more on after-school Jobs for income while junior high respondents depend on income provided by parents. As a result senior high respondents may have to give up hours of television exposure to acquire more income.

TABLE VII

CORRELATION BETWEEN TYPES OF MASS MEDIA EXPOSURE

TABLE VIII

RESULTS OF COMPARING MASS MEDIA EXPOSURE TO INCOME AND SOCIAL CLASS

Print media exposure was found to be significantly related to both income and social class (See Table VIII). As suggested by previous research, adolescents with high incomes and in the upper class spend more time reading newspapers and magazines than other adolescents (Greenburg & Dervin, 1970). It would appear that print media readership depends on income available to buy such media and social status motivation to read such media.

PERSONAL SOURCES

Family information sources as well as friends and teachers were personal information sources investigated in this study.

Initial Product Information

Mother was the personal information source used most frequently for initial product information (See Table I and Table II). In fact, mother was used more frequently for initial product information than any other information source. It is important to note that friends were the second most frequently used personal source of initial product information. Teachers and strangers were the personal sources used least frequently for initial product information.

Income. Examination of Table III discloses that neither income nor social class are significantly related to utilization of most personal sources of information for initial product information. The only significant relationship observed for income involves the use of father. The observed relationship indicates increased utilization of father for initial product information by respondents with higher incomes. The examination of junior and senior high school strata show that the relationship is significant for junior high but not senior high adolescents. Possibly junior high respondents increase their use of father as income increases because their father is the source of income. Another significant relationship is observed for income when junior high respondents are investigated. Junior high respondents with zero or high income ($10 or more weekly) used relatives outside the household more frequently for initial product information than did those junior high respondents who had incomes of less than $10 a week.

Social Class

Social class was found significantly related to use of friends for initial product information (See Table III). In this case increased levels of social class were found matched with increased frequency of use of friends for initial product information. No other significant relationships were found for social class and initial product information even when junior high and senior high strata were considered

Subsequent Product Information

As observed for initial product information mother was the personal information source used most frequently for subsequent product information (See Tables IV and V) Similarly, teachers and strangers were the personal sources used least frequently.

Income

Income of respondents was found significantly related to frequency of use of mother and friends for subsequent product information. As found by Ward and Robertson (1970) adolescents with low or medium income utilized mother more frequently as a source of subsequent product information than respondents who earned no income or high income.

The relationship between income and the use of father for subsequent product information reaches the .05 level of significance in the case of junior high respondents. The observed relationship is in agreement with that previously discussed for use of father for initial product information. Junior high respondents earning a high amount of money utilized father more frequently for subsequent product information than those who earned less money. The other significant relationship observed between income and subsequent use of personal sources was found for friends. Friends were used more for subsequent product information by respondents with high incomes than by those who earned less money.

Separate consideration of junior high and senior high respondents resulted in the findings of no significant difference between social class and use of personal information sources for subsequent product information.

The fact that no significant relationships were found between social class and use of parents for product information agrees with the findings of Ward and Robertson (1970).

Decisive Product Information

Mother was considered to be the most important personal information source by more respondents than any other source. Friends was the second most important source of personal information. The primary importance of mother and friends agrees with the high frequency of usage of those sources for initial and subsequent product information. Agreement was also found for the sources considered the least important as strangers and teachers were designated as least important by more respondents than any other personal source.

Income. Income was found to be significantly related to importance attached to different personal information sources (x2 = 37.41, level of significance - .001). Again the observed relationships parallel those found for initial and subsequent product information. The higher the income of respondents the more likely it was that respondents considered father and friends to be the most influential sources of personal information. Conversely family members other than father were more important personal information sources to respondents earning low incomes. No other significant relationships were found for income and the importance attached to personal information sources even when junior high and senior high respondents were considered separately.

Social class. The social class of respondents did not affect the importance attached to different personal information sources. As a result adolescents of any social class are more likely to be influenced by mother than any other personal information source. This is true for junior high as well as senior high respondents.

Opinion Leadership

Opinion leadership is an aspect of external search behavior which is applicable to personal information sources. Several aspects related to opinion leadership are examined in this study: (1) the relationship of opinion leadership to income and to social class, (2) the relationship of opinion leadership to mass media exposure, and (3) the relationship of opinion leadership to the relative age and family income of the recipient of the interpersonal communication.

Analysis of the relationship of the opinion leadership score, derived from a set of self-designation items, to income using Chi-square analysis indicates that opinion leaders are more likely to have high amount of earnings than lower earnings (See Table IX). As a result, it would appear that the amount of income earned by adolescents affects their ability to influence the consumption behavior of others. This finding is in agreement with those of Summers (1970) and Corey (1971).

TABLE IX

RESULTS COMPARING OPINION LEADERSHIP WITH INCOME AND SOCIAL CLASS

Relating opinion leadership to social class reveals that adolescent opinion leaders seem to be found in approximately equal proportions at all strata of society (See Table IX). As in the case of income and opinion leadership, the relationship found between social class and opinion leadership is consistent with previous research of Katz and Lazarsfeld (1955). Alternately, these findings are contrary to those reported by Summers (1970), Corey (1971) and Fenton and Leggett (1971). They found that opinion leaders were more likely to be in the higher social classes than nonopinion leaders

Mass media exposure. In order to test the relationship between opinion leadership and amount of mass media exposure, Chi-square analysis was employed.

Except for weekend television viewing, no significant relationships were found between opinion leadership and amount of exposure to television and radio (See Table X). The significant finding for amount of exposure to television on weekends is contrary to that reported by Katz and Lazarsfeld. Self-designated opinion leaders spent fewer hours watching television on weekends than nonopinion leaders. Possibly, adolescents who consider themselves opinion leaders spend their spare time on weekends influencing friends rather than watching television. Unfortunately, this reasoning does not explain why nonopinion leaders would watch more television on weekends if they are the ones who are being influenced.

TABLE X

OPINION LEADERSHIP AND EXPOSURE TO MASS MEDIA

Conversely, the significant relationships between opinion leadership and amount of exposure to newspapers and magazines do agree with those reported by Katz and Lazarsfeld. Respondents who considered themselves opinion leaders spent more time reading newspapers and magazines than those who did not consider themselves opinion leaders. Newspapers and magazines appear to be a more important source of information for adolescent opinion leaders than television and radio.

Relative age and family income. Reynolds and Myers (1966) suggested that the older the adolescent the higher the status accorded him by younger adolescents and the more likely it is that he is able to influence their attitudes. Similarly, previous studies by Summers (1970) and Corey (1971) have reported positive relationships between the incidence of opinion leadership and income

Results from the present study provide support for the idea that opinion leadership is a horizontal phenomenon within groups of adolescents who are similar ages and are in families with the same incomes. As shown in Table VI, the majority of respondents felt that influential friends were the same age (88 percent) and that the incomes (78 percent) of their families were comparable

Marketer-Dominated Sources

Initial Product Information

Store visits were used more frequently for initial product information than other marketer-dominated sources (See Table I and Table II). The least frequently used marketer-dominated source is shopkeepers or clerk.

TABLE XI

RELATIVE AGE AND FAMILY INCOME OF INFLUENTIAL FRIENDS

Income. As shown in Table III income did not affect frequency of use of marketer-dominated sources for initial product information. Neither junior nor senior high adolescents with different incomes varied significantly in their frequency of use of marketer-dominated sources for initial product information.

Social class. Further examination of Table III results in the conclusion that social class affects use of store visits for initial product information. A positive relationship was observed between social class and use of store visits as upper class respondents used store visits more frequently than those in other social classes. No other effects of social class were observed in the case of use of marketer-dominated sources.

Subsequent Product Information

Store visits were utilized more frequently for subsequent product information than any other marketer-dominated source As shown in Table IV shopkeepers or clerks were the least frequently utilized marketer-dominated source.

Income. Income was significantly related to the use of two marketer-dominated information sources, shopkeepers or clerks and store visits, for subsequent product information (See Table VI). In both cases the relationship was significant for junior high but not senior high respondents. Income may determine if junior high respondents are able to visit the store and use the clerks in these stores as increased usage of these sources for subsequent product information was coupled with high levels of income.

Social class. The only significant relationship between social class and use of marketer-dominated sources for subsequent product information is for the source of store visits. As encountered for use of store visits for initial product information adolescents in the upper class utilized store visits more frequently for subsequent product information than did those in the middle or lower classes.

Decisive Product Information

Store visits were not only the most frequently used marketer-dominated information sources but also were the most influential. As experienced for mass media sources the least used marketer-dominated source, shopkeepers or clerks, were not the least influential. Mail catalogs rather than shopkeepers or clerks were assigned the role of least important marketer-dominated source .

Income

No significant relationships were found in the importance assigned to the different marketer-dominated information sources and income of respondents. As a result the increased use of store visits for subsequent product information by respondents with high income does not mean that these earners of high incomes are more influenced by store visits than those earning lower incomes

Social class. Significant differences exist between social class and marketer-dominated information sources considered the most important by senior high respondents. Senior high respondents in the upper class were more likely to consider store visits as the most important marketer-dominated source of product information.

Brand Identification and Slogan Recall

Another aspect of external search processes involving marketer-dominated sources investigated in this study is brand identification and slogan recall. In order to measure these two aspects of brand awareness, respondents were asked to associate 20 brand names with generic product categories and to identify the sponsor of 10 different advertising slogans. The brand names and advertising slogans used were selected for their variety and recency of appearance in the media or retail stores.

The total number of brands correctly identified and slogans correctly recalled were determined for each respondent. The results of statistically relating these two variables to income and social class are shown in Table XII

TABLE XII

RESULTS COMPARING BRAND AWARENESS WITH INCOME AND SOCIAL CLASS

Brand identification is statistically related to income and social class. Although income and social class variables have been found to be statistically related to slogan recall of adults, present research results do not support a similar conclusion for adolescents.

For both income and social class, rising levels of the variables were found to be positively related to brand identification. Adolescents with high earnings and in the upper class were able to identify more brands than other adolescents. Evidently, increased income and social status are linked to increased external search behavior. It may be that the other roles of promotion--liking, conviction, etc.--should be emphasized when aiming at adolescents with high income and social status. The significant relationship observed between social status and brand awareness supports that found for adolescents by Guest (1942) three decades ago.

Retail Patronage Motives

All marketers regardless of their position in the channel are concerned with external search behavior of consumers. This is especially true of the retailer since he is the first to feel the effect of a consumer's decision. The retailer has many ways to send messages to the potential consumer. A substantial portion of his messages are sent from his retail store and will not be received unless the consumer visits his store. As a result of the importance of store visitation patterns of consumers to the retailer, the reasons for adolescent store choice were investigated.

The results of statistically examining the relationship between income and social class and retail store patronage motives with the use of Friedman two-way analysis of variance are presented in Table XIII. Neither income nor social class affect the importance ascribed to these different motives.

Subsequent Friedman analysis on the differences in the ranks of the different motives showed that significant differences exist between the importance respondents attach to different motives. Quality is considered the most important reason for patronizing a store by more respondents than any other motive. As shown in Table XIV assortment and style were also considered to be relatively important factors in the store selection decision. Conversely, respondents considered friendliness and convenience as relatively unimportant reasons for choosing a store. Interestingly, the low importance attached to friendliness and convenience aspects is contrary to behavior normally ascribed to low income earners and lower class buyers. The behavior observed for adolescents is more like that of adults with high amounts of income and in the upper social class. The relatively discretionary nature of the income of adolescents may explain why they are more concerned with the quality of store patronized rather than their price or friendliness of personnel.

TABLE XIII

RESULTS COMPARING RETAIL STORE PATRONAGE MOTIVES WITH INCOME AND SOCIAL CLASS

TABLE XIV

RELATIVE IMPORTANCE OF RETAIL STORE PATRONAGE MOTIVES

SUMMARY

The purpose of this study was to examine the effect of social class and income on the external search behavior of adolescents The aspects of external search behavior examined were: (l) use of mass media information sources; (2) use of interpersonal information sources; and, (3) use of marketer-dominated sources.

Data utilized in this analysis were collected through the use of a questionnaire which was administered during March, 1971 to 1,300 adolescents attending school in the Columbus, Ohio public school system. Of this total, 1,280 usable questionnaires were obtained. The analyses were performed with the use of the Chi-square test, Wilcoxon matched-pairs signed-ranks test, and Friedman two-way analysis of variance. For all statistical tests, an alpha level of .05 a lower was used.

SUMMARY OF FINDINGS

For mass media information sources--

1. Television was the mass media source used most frequently for "initial", "subsequent", and "decisive" product information.

2. Income and social class were not significantly related to adolescents utilization of mass media sources of product information.

3. Exposure to mass media was related to income but not to the social class of the respondents.

For interpersonal sources of information--

1. The respondent's mother was the most frequently utilized interpersonal source of information.

2. Income but not social class was significantly related to adolescents utilization of inter-personal sources_of product information for "initial", "subsequent", and "decisive" product information.

3. Income and exposure to print media were related to opinion leadership.

4. Opinion leaders were found to be similar to non-opinion leaders with regard to their age and family income.

For marketer-dominated sources of information--

1. Store visits were the most frequently used and most influential marketer-dominated sources of product information.

2. Income was significantly related to use of marketer-dominated sources for "subsequent", but not "initial" or "decisive" product information while social class was significantly related to all three dimensions of marketer-dominated sources of product information.

3. Brand identification, but not slogan awareness, was related to both income and social class of adolescents.

4. Adolescents' patronage motives did not vary as a result of income or social class.

AREAS FOR FUTURE RESEARCH

External search behavior of adolescents should be subjected to further investigation. A need for further research stems from the requirements of marketers for a better understanding of the consumer behavior of adolescents.

This need for further research encompasses two broad areas in the consumption decision-making process of adolescents. Future research should further investigate the nature of adolescents' external search behavior as well as the behavior that follows external search behavior

Increased information needs about the nature of external search behavior include the reasons why this behavior occurs, how the search behavior patterns of adolescents change as they mature, how external search behavior of adolescents in different geographical locations compare, and the type of consumption_related information that adolescents acquire from each information source.

In addition to further research on the nature of external search behavior of adolescents, future research should focus upon adolescent consumption behavior after external search has taken place. The most pressing need is to link aspects of external search behavior to subsequent consumption decisions. The usefulness of the information provided by this research effort will be enhanced if the different aspects of external search behavior are related to aspects of buyer behavior, such as brands purchased, product usage rates, repurchase patterns, stores patronized, and prices paid. The issue is whether adolescents with different external search behavior have different consumption behavior.

REFERENCES

Bowerman, Charles E. & Kinch, John W. Changes in Family and Peer Orientation of Children Between Fourth and Tenth Grades. Social Forces, 1959, 37, 206-211.

Coleman, James S. The Adolescent Society. New York: Free Press of Glencoe, 1961.

Collazzo, Charles J., Jr. Effects of Income Upon Shopping Attitudes and Frustrations. Journal of Retailing, 1966, 42. l-7

Corey, Lawrence G. People Who Claim to be Opinion Leaders: Identifying Their Characteristics by Self Report. Journal of Marketing, 1971, 35, 48-53.

Douvan, Elizabeth & Adelson, John. The Adolescent Experience. New York: John Wiley and Sons, Inc., 1966.

Engel, James F.; Kollat, David T.; & Blackwell, Roger D. Consumer Behavior. New York: Holt, Rinehart, and Winston, Inc., 1968.

Fenton, James S. & Leggett, Thomas R. A Way to Find Opinion Leaders. Journal of Advertising Research. 1971, 11, 21-25.

Gilkinson, Paul. What Influences the Buying Decisions of Teenagers. Journal of Retailing, 1965, 41, 31-41, 48.

Greenburg, Bradley, & Dervin, Brenda. Mass Communications Among the Poor. Public Opinion Quarterly, 1970, 34, 224-235.

Guest, Lester P. Brand Loyalty: Twelve Years Later. Journal of Applied Psychology, 1955, 39, 405-408.

Guest, Lester P. The Genesis of Brand Awareness. Journal of Applied Psychology, 1942, 26, 800-808.

Howard, John A. & Sheth, J. N. Theory of Buyer Behavior. Changing Marketing Systems. Edited by Reed Moyer. Chicago: American Marketing Association, 1967.

Katz, Elihu, & Lazarsfeld, Paul F. Personal Influence. New York: Free Press of Glencoe, 1955.

Katz, Michael, & Rose, Jan. Is Your Slogan Identifiable. Journal of Advertising Research, 1969, 9, 21-26.

King, Charles W., & Summers, John 0. Attitudes and Media Exposure, Journal of Advertising Research, 1971, 11, 26-32.

Larsen, Carl M. & Wales, Hugh G. Slogan Awareness in the Chicago Market. Journal of Advertising Research, 1970, 10, 38-41.

McCarthy, E. Jerome. Basic Marketing. 4th ed. Homewood, Ill.: Richard D. Irwin, Inc., 1971.

Martineau, Pierre. The Pattern o, Social Classes. Marketing's Role in Scientific Marketing. Edited by Richard Clewett. Chicago: American Marketing Association, 1957.

Myers, James H.; Stanton, Roger R.; & Haug, Arne F. Correlates of Buying Behavior: Social Class vs. Income. Journal of Marketing, 1971, 35, 8-15.

National Industrial Conference Board. A Graphic Guide to Consumer Markets. New York: National Industrial Conference Board, 1969.

Rainwater, Lee; Coleman, Richard P.; Handel, Gerald. Workingman's Wife. New York: Oceana Publications, Inc. 1959.

Reynolds, William H., Myers, James H. Marketing and the American Family. Business Topics, 1966, 14, 57-66.

Rich, Stuart U., & Jain, Subhash C. Social Class and Life Cycle as Predictors of Shopping Behavior, Journal of Marketing Research, 1968, 5, 41-49.

Robertson, Thomas S. Consumer Behavior, Glenview, Ill.: Scott, Foresman and Company. 1970.

Siegel, Sidney. Nonparametric Statistics for the Behavioral Sciences New York: McGraw-Hill Book Company, 1956.

Slocum, John W., Jr., & Mathews, H. Lee. Social Class and Income as Indicators of Consumer Credit Behavior. Journal of Marketing, 1970, 34, 71-78.

Solomon, Daniel. Influences on the Decisions of Adolescents. Human Relations, 1963, 16, 45-60.

Summers, John O. The Identity of Women's Clothing Fashion Opinion Leaders. Journal of Marketing Research. 1970, 7, 178-185.

Teen-Age Consumer: Dynamic Force in the U. S. Economy, Senior Scholastic, February 25, 1965, p. 5.

The Los Angeles Chamber of Commerce. The Dynamics of the Youth Explosion: A Look Ahead. Los Angeles: Los Angeles Chamber of Commerce, 1967.

Thompson, Bryan. An Analysis of Supermarket Shopping Habits in Worchester, Massachusetts. Journal of Retailing, 1967, 43, 17-29.

Walters, C. Glenn & Paul, Gordon W. Consumer Behavior: An Integrated Framework. Homewood, Ill.: Richard D. Irwin, Inc., 1970.

Ward, Scott, & Robertson, Thomas S. Family Influences on Adolescent Consumer Behavior. Paper presented to the Association for Consumer Research, Amherst, Mass., August, 1970.

Wasson, Chester R. Is it Time to Quit Thinking of Income Classes? Journal of Marketing, 1969, 33, 54-69.

White, R. Clyde. Social Class Differences in the Uses of Leisure. American Journal of Sociology, 1955, 61, 145-150.

"Youth Appeal...A Creative Booby Trap?" Grey Matter, 1971, 17.

----------------------------------------