Focus Groups: a Review of Some Contradictory Evidence, Implications, and Suggestions For Future Research

ABSTRACT - Several commonly made assumptions about focus group methodology were reviewed in this article. Empirical evidence from studies on group problem-solving, group brainstorming, discussion groups and psychotherapy groups seems to contradict many of these assumptions. Since these unresolved issues have strong implications for the conduct of qualitative research, directions for future research on the focus group technique are suggested.

Citation:

Edward F. Fern (1983) ,"Focus Groups: a Review of Some Contradictory Evidence, Implications, and Suggestions For Future Research", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 121-126.

Advances in Consumer Research Volume 10, 1983      Pages 121-126

FOCUS GROUPS: A REVIEW OF SOME CONTRADICTORY EVIDENCE, IMPLICATIONS, AND SUGGESTIONS FOR FUTURE RESEARCH

Edward F. Fern, Virginia Polytechnic Institute and State University

ABSTRACT -

Several commonly made assumptions about focus group methodology were reviewed in this article. Empirical evidence from studies on group problem-solving, group brainstorming, discussion groups and psychotherapy groups seems to contradict many of these assumptions. Since these unresolved issues have strong implications for the conduct of qualitative research, directions for future research on the focus group technique are suggested.

INTRODUCTION

In the last few years focus group interviewing has become a widely used technique for conducting qualitative research. However, some concern has been expressed about the potential misuse of this convenient interview form (Caruso 1976). This concern cakes on added significance in view of the fact that very little research has been reported on the use of this technique. Therefore, this paper will provide (1) a review of the focus group literature, (2) some contradictory empirical evidence from related disciplines, and (3) suggestions for future research in this area.

Ac the outset it should be recognized that the very definition of focus groups depends upon the moderator's preferred methodology. Most methods of focus group interviewing are consistent with Davidson's (1975) definition: "A focused group interview is a qualitative tool for collecting information in which a number of respondents simultaneously discuss a given topic under the guidance or a moderator." This definition appears to be broad enough to cover the wide range of uses that have been reported (Fern 1982).

Assumptions About Focus Group Research

Although there is not unanimous agreement among group moderators on precise methodology, some areas of agreement do exist. In this section several assumptions or common beliefs about focus group methodology will be presented along with chose views that appear to be divergent.

Authors on group interviewing assume that focus groups provide more information that is qualitatively better than individual interviews. Moreover, the advocates of this research technique see group output as providing greater synergism (Hess 1968), stimulation (Goldman 1962), spontaneity and cander (Hess 1968), release of inhibition (Merton 1956), and feelings of security (Hess 1968). However, Fern (1982) in a study which compared focus groups with individual interviews found that individual interviews produced more ideas and ideas of nigher quality than focus groups.

A second assumption is that the ideal group size is between eight and twelve members. However, Sampson (1972) believes there is no "correct" size for any group. The value or a discussion group is independent of its size. size may vary from 5 to 12 members depending on how articulate and fluent the members are. According to Wells (1974), the ideal group size depends on the seating arrangements and the interviewer's personal style. Hess (1968) prefers twenty or more people because large groups have a greater synergistic effect. Smith (1977) thinks 8 to 10 members are adequate because larger numbers tend to become unmanageable, and smaller numbers may not produce a wide enough range of ideas. According to Payne (1976), no group discussion should ever have more than eight respondents--6 or 7 is perfectly adequate. Fern (1982) found that the incremental number of ideas per person decreased as group size increased from one to eight members.

A third assumption is that the moderator plays a crucial role in obtaining the desired information from focus groups. The moderator's expertise, personality traits, and focus group procedure are critical to promoting group interaction (Axelrod 1976, Caruso 1976). Expertise includes ability, knowledge in social or clinical psychology, past focus group experience, and product or problem knowledge. Traits deemed important are sensitivity, outward personality, and a genuine interest in people. However, Fern (1982) raises serious questions about the importance of moderators. In his study, moderated focus groups did not generate significantly more or better ideas than unmoderated groups.

A fourth assumption is that within-group homogeneity of members should be maintained. According to Goldman (1962), the group should share some common interest (e.g., shoe buyers, drug manufacturers, or purchasers of luxury items). Merton (1956) thinks the group will be more productive if members are socially and intellectually homogeneous--specifically in terms of age, education, and occupation. Axelrod (1976) calls for as much commonality as possible and suggests needs, goals, and life styles of group members should be similar. Wells (1974) sees homogeneity with respect to social class, stage in the family life cycle, and product use as being particularly desirable.

These are not the only assumptions made about focus groups. Some authors indicate more open and intimate reports of personal experiences and sentiments will be achieved under group conditions than when individuals are interviewed separately (Merton 1956). Sampson (1972), on the other hand, finds that if it is necessary to obtain considerable detail about people's past and current behavior, their beliefs, opinions, and attitudes, then an individual intensive interview may be preferred. Still others assume respondents forge: about cape recorders (Wells 1974), respondents do not notice one-way mirrors (Wells 1974), the interview setting is crucial (Peterson 1975), and a warm-up or general discussion is necessary to "sneak" into the focal topic.

In summary, several assumptions about focus group methodology and the rationale behind them have been reviewed. None of the assumptions are supported by empirical evidence. As a matter of fact, only one study has tested focus group assumptions. Therefore, a review of literature or brainstorming, group problem solving, psychotherapy, self-disclosure, and other related fields was undertaken to see if these assumptions were reasonable. However, before exploring this literature it should be noted that these studies may not be directly applicable to focus groups.

A COMPARISON AMONG BRAINSTORMING, PSYCHOTHERAPY, AND FOCUS GROUPS

Focus group researchers are quick to point out that focus groups are different from brainstorming and psychotherapy groups. On the other hand, Calder (1977, p. 357) draws on group therapy sessions in his discussion of the clinical approach and Fern (1982) sees brainstorming casks as being similar to specific casks used in the exploratory approach. The position taken in this paper is that there are fundamental differences among the three types of groups but nevertheless much can be learned by exploring empirical research on brainstorming, group problem-solving, and group psychotherapy.

One major difference that distinguishes these types of group phenomena is the nature of the cask. Brainstorming groups rely on the use of creative tasks which call for the suspension of critical judgement (e.g., providing new uses for existing products) while problem-solving groups are requested to arrive a: a group solution to a specific problem (e.g., changing work methods). Psychotherapy tasks usually involved verbal interaction, as well as, nonverbal activities (e.g., group discussion and/or listening activities). Focus group tasks can be characterized as exploratory, clinical, and phenomenological (Calder 1977) and range from uncovering theoretical constructs to simply hearing consumers talk. In many ways focus group tasks are different from the others yet, in some ways they overlap.

Another difference is in the explicitness of the group's objectives. In both brainstorming and problem-solving groups the goals are quite explicit. Therefore, group members are externally motivated to be productive. This is not the case in psychotherapy and focus groups. In tact, the therapist and moderators usually are the only participants in the group interaction to know the groups' goals.

Unlike brainstorming and problem-solving groups the therapists and moderators provide structure for psychotherapy and focus groups. Therapy and focus group sessions usually adhere to a set of preplanned activities and/or a discussion guide. Occasionally, therapy sessions rely on written instructions or tape recordings for structure. In brainstorming and problem solving groups written instructions provide the only structureC formal leaders or moderators are rarely used.

Finally, the unit of analysis differs according to the type of group phenomena. Recently, brainstorming and problem-solving studies view the group as the unit of analysis. Researchers in these areas are primarily interested in group rather than individual productivity. Similarly focus groups treat the group as the unit of analysis. On the other hand, therapists are primarily interested in changing individuals' attitudes and behaviors through group discussion.

Notwithstanding the above differences, these groups have a common dimension--group interaction. Group members are par: of a social setting and their contributions are highly dependent upon feedback from other group members. In brainstorming and problem solving groups much is known about the nature of this interaction and the critical variables affect it. In psychotherapy, there is a growing body of knowledge about factors affecting group interaction. Ye:, knowledge about interacting focus groups is meager. As a result the existing knowledge in these other divergent areas was reviewed to uncover research paradigms, variables, and hypotheses that could be tested in a focus group context.

Individuals Versus Groups

Focus groups are commonly assumed to provide advantages not available in individual interviews. Most of these advantages seem to indicate that group interaction facilitates the participation of individuals. Yet studies in creative problem solving and brainstorming have consistently shown that interacting groups are inferior to an equal number of individuals working alone.

Taylor (1958) and his colleagues used a brainstorming procedure developed by Osborn (1953) to compare the number of ideas generated by male undergraduate students working alone with those generated by groups. In this study ideas generated by individuals were pooled as though they had been generated by a group without interaction among the members. Repeated ideas were not counted so that the resulting "nominal groups" were directly comparable to the interacting groups. The most important finding was that the nominal groups out-performed real groups in terms of number of ideas produced and number of unique ideas produced. The mean quality for real groups was also inferior to that of the nominal groups. Taylor concluded that group participation when using brainstorming inhibits creative thinking. Dunnette et al. (1963), using research scientists and advertising men rather than students replicated the Taylor findings.

Managers from a public utility were used in a study to determine how differences in personal characteristics were related to individual problem-solving (Campbell 1968). All respondents were mailed a personal history and personality questionnaire along with a problem which they were to solve individually. The subJects were then assembled into 12 four-man experimental groups and 8 four-man control groups The experimental conditions were: (1) individual solution while working alone, (2) an individual solution after hearing and participating in a group discussion, and (3) a group solution after discussion. Campbell found the judged quality of real group solutions to be inferior to nominal groups. None of the personality main effects nor their interactions with the experimental treatments were significant.

Several attempts were made to determine why individual brainstorming is superior to group brainstorming and how group brainstorming may be made more effective. Bouchard (1972) used a procedural rule in the group conditions. Each respondent contributed ideas in sequence with each saying "pass" if he had nothing to contribute. Nominal groups out-performed real groups on two of the three problems and real groups provided more good ideas than nominal on one of the problems. Bouchard attributed this finding to the sequence procedural rule employed by the real groups in this study. In a later study (Bouchard, et al., 1974), the procedural rule had no effect

Group Size

Although the individual versus group issue has received the most attention in studies on brainstorming, the effects of group size on group productivity has also received much attention. The review to follow is by no means complete. However, i: does provide evidence that: (1) group size is a significant factor in accounting for productivity, and (2) the optimal group size for creative tasks is somewhat less than 8 members.

Although various studies purport to show the group size that is optimum, there is little agreement on an operational definition of optimum group size. Slater 1958) found that group participants considered five members to be optimum for the task of discussing a human relations problem. He also found members of larger groups less satisfied with the amount of time available for discussion, with their opportunity to participate, and with the group meeting or its decision.

Bales and Borgatta (1955) coded every observable act of verbal and non-verbal behavior for groups of size 2 through 7. They noted that as group size increased each individual had relatively less time to talk and had more people to talk with. This resulted in the observation that as group size increased there was an increase in giving suggestions and a decrease in asking for and giving opinion. Bales and Borgatta offer two lines of reasoning as justification: (l) when time is at a premium members feel pressure to cake the most direct approach (i.e., making suggestions without justifying) and (2) the larger groups may pose a more formidable sanctioning system and thereby inhibit evaluative statements.

Another result of time pressures is that the number of people who actively participate in the group will decrease Reluctant participants will tend to be forced to engage in behaviors which can be performed simultaneously such as listening, showing tension (i.e., withdrawal, nervous mannerisms, or awkward pauses that occur for the group as a whole), and showing tension release (primarily through laughter of the group as a whole). Surprisingly however, in the Bales and Borgatta study as group size increased the observed tension decreased.

Bouchard and Hare (1970) have suggested there may be more rapidly diminishing returns in nominal groups than real groups. Using male students, with group size (5, 7, and 9) as one of the independent variables, they found just the opposite. In fact, as real group size increased there was no appreciable increase in the number of ideas produced. However, for nominal groups the number of ideas produced by nine member groups was almost double the number Produced by five member grouPs.

The Moderator

Since focus groups evolved from clinical psychology, this area seemed to be a logical place to look for research on the role of the group counselor or facilitator. It was surprising to find group studies in psychotherapy that compared professionally led groups with leaderless groups and groups led by lay persons. Even more surprising was the fact that many of these studies seriously question that trained counselors are necessary for effective group therapy. The implications of this issue for focus group methodology will be discussed later.

Trained therapists were used by Poser (1966) to compare differences in behavior change between patients created by lay [The lay persons were eleven young women undergraduates with no background in psychology and two inpatients--one an alcoholic and the other suffering from hysteria. Among the professional therapists were seven certified psychiatrists with five to seventeen years experience, six psychiatric social workers, and two occupational therapists.] persons and chose created by professional therapists. Each of the 343 hospitalized male chronic schizophrenic patients was assigned to a ten person group using age, severity of illness and length of hospitalization as criteria for achieving within-group homogeneity. The resulting groups were then randomly assigned to the control condition or to one of the 98 therapists. Each group met one hour daily five days a week over a five month period. Before and after each therapy session patients were tested with a battery of side psychological tests. Comparisons of post-therapy test behavior showed: (1) patients treated by lay therapists were significantly superior to the untreated control on four of the six tests, (2) those treated by professionals were superior to the untreated controls in two of the six, and (3) patients treated by lay therapists performed significantly better than those treated by professionals on three of the six tests and there were no significant differences on the other three. Poser concluded that traditional training in the mental health profession may be neither optimal nor even necessary for achieving behavior change in mental hospital patients.

The above evidence indicates that other means of structuring group interaction may be just as effective as using professional therapists. Why then has the role of the leader been depicted as a central force in theories behind personal change in groups? Lieberman (1976) suggests several reasons: (1) most theories have been developed by highly charismatic leaders who have been myopic in overestimating their contribution to the curative process, (2) supercharged feelings toward the leader are often generated in groups, and (3) professionalization, distinctive languages, and fee structures tend to emphasize the centrality, prominence, and indispensability of the leader's role. Lieberman's reasoning may be equally applicable to the perceived importance of focus group moderators.

Member Homogeneity

Another question that has received attention in group problem-solving studies is how heterogeneous should the group's membership be¦ On the one hand variation in personal experience, perceptions, and cognitive structures seem desirable when a wide range of information or creativity is required. On the other hand heterogeneity of individuals' backgrounds may lead to conflict resulting in less group productivity and a narrower range of information. Several studies have investigated this dilemma.

Hoffman (1959) used personality characteristics as the basis dividing undergraduate male and female students into homogeneous and heterogeneous four-person groups. In the first of two problem solving casks, nonhomogeneous groups provided significantly higher quality solutions than homogeneous groups In a second task the difference between homogeneous and heterogeneous groups

was not significant. Additionally, homogeneous groups were satisfied with their "eminently poor" solutions while no relationship between satisfaction and solution quality was found in the heterogeneous groups.

The degree of member homogeneity desired may be best determined in light of the task or problem the group is asked to address. Cohen, Whitmyre, and Funk (1960) hypothesized that cohesive brainstorming groups would out-perform both nominal groups and noncohesive groups on neutral problems with increased superiority occurring on ego-involving problems Cohesive groups were comprised of two individuals who a priori preferred each other as partners and noncohesive pairs were formed by pairing those who least preferred each other. In this study only unique ideas were counted--duplicates or ideas within a group were thrown out. The results indicate that there were no significant differences be tween groups on the two neutral problems and that the cohesive groups produced significantly more unique ideas than the other groups for the ego-involving problem Cohen's group concludes that for creative tasks consideration should be given to the desire of group members to work together.

A wide variety of problems was used by Hoffman and Maier (1961) to test the generalizability of Hoffman's (1959) earlier finding. As in the previous study correlations among students responses to personality scales were used to determine four-person group homogeneity and heterogeneity. Both group conditions received the five problems at intervals during laboratory sections of an undergraduate psychology course. In general, the quality of solutions by heterogeneous groups were judged to be significantly higher or did not differ in quality from solutions by homogeneous groups. A post hoc analysis of homogeneity in terms of sex composition offered additional support for the superiority of heterogeneous groups. Mixed-sex groups tended to produce higher quality solutions than did all-male groups (there was only one all-female group). The authors concluded that solutions with high quality and high acceptance among members can be obtained from groups in which the members have substantially different perspectives on the problem, particularly if these differences are expressed and used by the group in arriving a: the final decision.

Self Disclosure

The degree to which individuals reveal personal information about themselves to another person may be related to their affection, love, and trust for that other person as well as the willingness of the other person to divulge personal information.

Jourard and Friedman (1970) in the first of two experiments, expected reduced social distance between the subject and the male experimenter to decrease self-disclosure. One group of students disclosed information to a tape recorder while the experimenter was out of the room. A second group disclosed information to the experimenter who did not look into the subjects' eyes In the third condition the experimenter maintained eye contact as the subjects disclosed information. The dependent measures were time spent disclosing on various topics, and attitudes towards the experimenter. The male students tended to maintain their level of disclosure under all three conditions of interaction. On the other hand, women students reduced their amount of disclosure with the experimenter's reduction in distance (i.e., eye contact). The attitude items exhibited a complex relationship to disclosure which differed depending on the situation and the subjects' sex.

A second study was done to further decrease distance and to refine the attitude measures. Fifty male and 50 female undergraduate psychology students were assigned to one of four groups representing differing levels of interpersonal distance. Distance (4 levels) ranged from nodding and saying "yes" (greatest distance) to touching and disclosing personal information (least distance). Each subject was then presented eight cards on which was typed a self-disclosure question (4 high and 4 low disclosure questions). Then the interviewer asked them to discuss each question. Time spent disclosing was the Primary dependent variable. The difference in time spent disclosing between groups was significant with the exception of the two groups representing the greatest social distance. There were no significant differences due to the subjects' sex.

Others have noted what appears to be a norm of reciprocity in self-disclosure studies. Several correlational studies such as the Jourard (1959) study show a relatively high correlation between the amount of disclosure output given to a subject and the amount of disclosure input reciprocated by the subject. However, there is no reason to believe this relationship is linear. Jourard (1959) noted char two people liked least by their co-workers were the highest and lowest disclosers. It is possible that too little or too much disclosure suggests deviance or mental health problems and results in less liking or affection for the discloser.

Cozby (1972) tested the reciprocity hypothesis and the disclosure/liking relationship over a range of intimacy values. Thirty female undergraduate volunteers were asked to play the role of the "other subject" in this study. The subjects were assigned to either the high, medium, or low disclosure condition and given 10 disclosure statements which "one of the girls decided to tell." The subject then chose ten statements from a list of 70 that she would reveal about herself. The results support a curvilinear relationship between disclosure and liking. Medium disclosure resulted in greater liking than high or low disclosure. Reciprocated intimacy was not a linear function of the self-disclosure of others (i.e., both the linear and quadratic components were significant). Partial support for the mental health self-disclosure hypothesis was found. In general, subjects rated the other person who disclosed at a high level as more maladjusted (on 7 impression formation rating scales) than in either of the other conditions.

SUMMARY

Several assumptions about focus group methodology were reviewed. Empirical evidence was presented which casts doubt on the validity of some of these assumptions. Yet, other evidence seems to support these assumptions.

A review of the brainstorming literature suggests that individuals working alone produce more ideas than individuals working in groups. Although the evidence is not as consistent, studies in the brainstorming area also suggest that the optimal group size is somewhat less than eight members. Several alternatives to moderated group discussions were tested in studies on psychotherapy. The results of these tests suggested the moderator may not be crucial for efficient group performance. Other studies were reviewed which show that member homogeneity does have an impact on group productivity although the magnitude and direction of impact may depend on the specific situation and the dimensions of homogeneity chosen for study. Finally, no research was found to support the notion that more self-disclosure occurs in groups than in individual interviews--all of the disclosure studies involved groups. However, these studies did provide insight into factors that enhance self-disclosure and the effects of self-disclosure on the group. If the above findings are generalizable to focus groups then they may have implications for the structure of a focus group methodology.

Suggestions For Future Research

Many casks performed by focus groups are similar in nature to brainstorming, problem-solving, and psychotherapy tasks. Although the studies reviewed from these areas are no: directly applicable to the conduct of focus groups, they do raise questions which cannot be easily ignored. The purpose of this section is to propose hypotheses for research on focus groups To many readers these hypotheses may be intuitively obvious, self-evident, tautological, or otherwise inappropriate. However, caution should be exercised in accepting without question the assumptions which have evolved from reports on the use of this technique. To date there is little empirical evidence to support any of the commonly accepted tenets of focus group methodology. The interested researcher can test the hypotheses that follow or derive others from the existing literature to conduct research on this popular data collection method.

Focus groups are widely used to generate lists of product attributes, belief statements, new product ideas, new uses for existing products, new products for existing uses, and items for questionnaires, yet, the brainstorming studies most consistent finding suggests that individuals are better than groups for thought generation. Analogously individual interviews may be more appropriate than focus groups in some situations Additional research needs to be done to determine what situations call for groups and what situations call for individual interviews A general hypothesis that can be tested empirically is:

H1: When the research task involves eliciting thoughts, more information will be provided by the pooled output of individual interviews than by the output of focus groups.

The effect of group size on idea generation has implications for research costs. The addition of group members beyond the optimal group size is not only inefficient in terms of resource utilization but may be counterproductive as well. Recruiting costs might be lower for smaller groups with no attendant loss of information. However, several moderators reported overrecruiting [This information was made available to the author through private telephone conversations with several group moderators.] to insure a large enough turnout (8-12 members). If everyone that was recruited showed up (e.g., 16 people), extra chairs were brought in and the group was conducted as usual. This practice increases recruiting costs and may result in inhibited responses due to the size of the group. Efficient resource utilization and productivity may also be adversely affected if fewer than the ideal number or participants show up. If only half the desired number show up, the results from the session may not be very useful but the researcher still has to bear the cost of the session. Hypotheses two follows:

H2: More information will be collected per moderator hour in two 4-person focus groups than one 8-person focus group.

Research on the crucial role of the group moderator raises several questions. First, is the role of the moderator essential to the conduct of focus groups? The second, is the magnitude of the moderator's contribution worth the cost? Thirdly, what other alternatives are available for leading group discussion? Throughout the group counseling literature evidence points to the noncrucial role of facilitators. However, it can be argued that the objectives of psychotherapy and focus group research are quite different and therefore the roles of the group leaders are quite different. Again, it would appear this issue is researchable. As pointed out in the group counseling review, alternatives to group moderators do exist. Written instructions have been used in prior research to provide structure for group discussion. .store frequently, caped recorded messages have provided structure. Tapes provided the additional advantage of budgeting the available time among the various activities However, both of these methods are inflexible compared to group moderators Moderators can change the course of the group as the situation requires it. A final alternative is informal or peer leadership such as Poser (1966) used. The problem with this approach is that focus groups, unlike psychotherapy groups, are used to collect some predetermined type of information. Therefore the moderator needs to be aware of the group's objectives or the type of information wanted and prepare a strategy for getting it. This may be asking coo much of an untrained peer or informal leader. Some of these ideas are incorporated in the following hypothesis:

H3: When information is to be collected unobtrusively, written instructions, cape recorded messages, or peer leaders are more appropriate than professional moderators.

It appears that in some situations within group heterogeneity is desirable, keeping in mind that too much heterogeneity may result in conflict leading to decrements in group productivity. In determining how much heterogeneity is desired the researcher faces a dilemma. On the one hand, a wide range of experiences and backgrounds among group members may provide a wider and richer array of information. On the other hand heterogeneity may adversely affect the ability of the moderator to manage the group's discussion. It would appear that the degree and dimensions of homogeneity should be determined by the nature of the cask. If group members are being asked to share similar product experiences, perhaps groups should be as homogeneous as possible. Conversely, heterogeneity may be desirable when the task is to generate an exhaustive list of product attributes. An issue that overrides these concerns is the constraints posed by measurement and recruiting problems. The costs of recruiting would be expected to escalate as the homogeneity dimensions become more difficult to measure and respondents occupying extreme positions on those dimensions become more difficult to recruit.

H4: When the research objective is to elicit shared experiences from a heterogeneous sample of respondents, heterogeneity should be achieved among groups of similar respondents or among individual interviews.

H5: When a wide range of divergent thoughts is desirable, groups or individual interviews should consist of respondents with heterogeneous experiences and backgrounds.

Studies on the intimacy of self-disclosure suggest that a group member's reciprocation of intimate self-disclosures may be related to the other person's sex, liking the other person, and the social distance between disclosures. Moreover, self-disclosure increased throughout the duration of the group discussion. These findings suggest what some group moderators no doubt already know. If the group discussion is to deal with personally sensitive issues then the moderator might set the tone by disclosing intimate information during the warm-up part of the interview. It makes intuitive sense that one's propensity to disclose personal information is also related to other factors such as group size, the moderator's style, member homogeneity, and whether or not the other group members are friends.

H6: Individual interviews provide more intimate self-disclosure than focus groups.

H7:As group size increases, intimate self-disclosure will increase at a decreasing race (i.e., diminishing returns will occur).

H8: Respondents' self-disclosures will be facilitated if the focus group moderator (or interviewer) discloses intimate personal information first.

The implications and propositions just discussed are predicated on the generalizability of the research that was reviewed. In most cases the studies were done in situations quite different from typical focus group interviews. Focus groups are not normally recruited from a population of chronic schizophrenics. And, personality is not usually a variable upon which within group homogeneity is sought. Yet, if the relationships reviewed in this report are not situation and sample bound then perhaps focus group researchers should be concerned about chem. One way to satisfy our skepticism or doubts is to formulate these findings as hypotheses and test them in the focus group context.

It is hoped that this paper will provide the stimulus, for those interested in focus groups, to do research in that area. There are problems unique to the study of small groups but they are not insurmountable. Effort spent in researching this widely used technique should aid in providing a much needed methodological structure.

REFERENCES

Aronson, E., and J.M. Carlsmith (1968), "Experimentation in Social Psychology," in Gardner Lindzey and Eliot Aronson, The Handbook of Social Psychology, 2, Reading: Addison Wesley, pp. 1-79.

Axelrod, Myril D. (1976), "The Dynamics of the Group Interview," in Beverlee Anderson (ed.), Advances in Consumer Research, 3, Ann Arbor: University of Michigan, pp. 437-441.

Axelrod, Myril D. (1975), "10 Essentials for Good Qualitative Research," Marketing News, March 14, p. 6.

Bales, Robert F., and Edgar F. Borgatta (1955), "Size of Group as a Factor in the Interaction Profile," in Paul A. Hare, Edgar F. Borgatta and Robert F. Bales (eds.), Small Group Studies in Social Interaction, New York: Alfred A. Knopf, pp. 396-413.

Berent, Paul H. (1966), "The Depth Interviews," Journal of Advertising Research, 6, pp. 32-39.

Bouchard, Thomas J., Jr. (1972), "A Comparison of Two Group Brainstorming Procedures," Journal of Applied Psychology, 56, pp. 418-421.

Bouchard, Thomas J., Jr., Jean Barsaloux, and Gail Drauden (1974), "Brainstorming Procedure, Group Size, and Sex as Determinants of the Problem-Solving Effectiveness of Groups and Individuals," Journal of Applied Psychology, 59, pp. 135-138.

Bouchard, Thomas J., Jr., and Melane Hare (1970), "Size, Performance, and Potential in Brainstorming Groups," Journal of Applied Psychology, 54, pp. 51-55.

Calder, Bobby, J. (1977), "Focus Groups and the Nature or Qualitative Marketing Research," Journal of Marketing Research, 14, pp. 353-364.

Campbell, John P. (1968), "Individual Versus Group Problem Solving in an Industrial Sample," Journal of Applied Psychology, 52, pp. 205-210.

Caruso, Thomas E. (1976), "Moderators Focus on Groups: Session Yields 7 Hypotheses Covering Technology Trend, Professionalism, Training Techniques, Reports, etc. (sic)," Marketing News.

Cohen, David, John W. Whitmyre, and Wilmer H. Funk (1960), "Effect or Group Cohesiveness and Training Upon Creative Thinking," Journal of Applied Psychology, 44, pp. 319-322.

Cozby, Paul C. (1972), "Self-disclosure, Reciprocity and Liking," Sociometry, 35, pp. 151-60.

Davidson, Thomas Lea (1975), "When ... if ever ... are Focused Groups a Valid Research Tool? in Edward M. Mazze (ed.), 1975 Combined Proceedings, American Marketing Association Conference, Series No. 37.

Dunnette, Marvin D., John Campbell, and Kay Jaastad (1963), "The Effect of Group Participation on Brainstorming Effectiveness for Two Industrial Samples," Journal of Applied Psychology, 47, pp. 30-37.

Fern, Edward F. (1982), "The Use of Focus Groups For Idea Generation: The Effects of Group Size, Group Type, Acquaintanceship, and the Moderator on Response Quantity and Quality," Journal of Marketing Research, Vol. 19 (Winter 1982).

Goldman, Alfred E. (1962), "The Group Depth Interview," Journal of Marketing, pp. 61-68.

Hess, John M. (1968), "Group Interviewing" in Robert L. King (ed.), 1968 ACR Fall Conference Proceedings, Chicago, Illinois: American Marketing Association, pp. 193-196.

Hoffman, L. Richard (1959), "Homogeneity of Member Personality and Its Effect on Group Problem-Solving," Journal of Abnormal Social Psychology, 58, pp. 27-39.

Hoffman, L. Richard, and Norman R.F. Maier (1961), "Quality and Acceptance of Problem Solutions by Members of Homogeneous and Heterogenous Groups," Journal of Abnormal Social Psychology, 62, pp. 401-407.

Jourard, Sidney M. (1959), "Self-disclosure and Other Cathexis," Journal of Abnormal and Social Psychology, 59, pp. 428-31.

Jourard, Sidney M. and Robert Friedman (1970), "Experimenter-Subject 'Distance' and Self-Disclosure," Journal of Personality and Social Psychology, 15, pp. 278-82.

Lieberman, Morton A. (1976), "Change Induction in Small Groups," in Mark R. Rosenweig, and Lyman W. Porter (eds.), Annual Review of Psychology, 27, Palo Alto, California: Annual Review. Inc., pp. 217-250.

Merton, Robert E;., Marjorie Fiske, and Patricia Kendall (1956), "The Group Interview," The Focused Interview, Glencoe, Illinois: The Free Press.

Osborn, Alex F. (1953), Applied Imagination, New York: Charles Scribner's Sons.

Payne, Melanie S. (1976), "Preparing for Group Interview," in Beverlee Anderson (ed.), Advances in Consumer Research, Ann Arbor, Michigan: University of Michigan, pp. 434-436.

Peterson, Karen Ida (1975), "The Influence of the Researcher and his Procedures on the Validity of Group Sessions," in Edward M. Mazze (ed.), 1975 Combined Proceedings of American Marketing Association Conference, Series No. 37, Chicago, Illinois: American Marketing Association.

Poser, Ernest G. (1966), "The Effect of Therapist's Training on Group Therapeutic Outcome," Journal of Consulting Psychology, 30, pp. 283-289.

Smith, George H. (1954), "Interviewing in Groups," Motivation Research in Advertising and Marketing, New York: McGraw-Hill Book Company, Inc.

Taylor, Donald W., Paul C. Berry, and Clifford H. Block (1958), "Does Group Participation when using Brainstorming, Facilitate or Inhibit Creative Thinking?." Administrative Science Quarterly, pp. 93-47.

Wells, William D. (1974), "Group Interviewing" in Robert Ferber (ed.), Hand-book of Marketing Research, New York: McGraw-Hill Co., pp. 136-146.

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