The Impact of Culture on Group Behavior: A Comparison of Three Ethnic Groups. (Research)

Article excerpt

Multicultural counseling has become an important discipline and a primary source for explaining human development and functioning (Lee, 1997; Pedersen, 1991); therefore, cultural differences should be actively considered in mental health interventions. The purpose of this study was to investigate the impact of culture on group behavior and on outcomes. Specifically, the study compared self-disclosure, group interaction, and outcomes in counseling groups for Arab (Moslem and Druze) and Jewish adolescents in Israel.

MULTICULTURALISM IN THE ISRAELI CONTEXT

Arabs in Israel encompass several ethnic groups, among which Moslems and Druze are examined in the present study. Both groups share a basic Arab culture characterized as collectivist and authoritarian (Dwairy, 1998). Religion, tradition, and family are central components in this culture (Barakat, 1993; Dwairy, 1998; Jackson, 1997). Accordingly, social relationships are built on duty and faithfulness to family and friends rather than on self-needs. The individual is dependent on his or her family at large (Nydell, 1987), and the family's reputation depends on the individual member's behavior. The concepts of self-esteem and respect are interdependent in Arab families, which operate as close units--cohesive, loving, and warm, but at the same time suppressing personal feelings, opinions, experiences, and needs. As a result, the Arab individual typically avoids overt emotional expression (El-Rufaie & Absood, 1993). The need to conform to collective norms has led to "pleasing" as a major coping mechanism (Griefat & Katriel, 1989).

Within this common Arab culture, there are cultural variations between groups, depending on their degree of acculturation (Lee, 1997). Dwairy (1998) has identified three major groups among Arabs: traditional, bicultural, and Westernized. Traditional Arab identity is the common cultural identity of Arabs in rural areas, where people live with traditional collectivist values and norms within their extended families and social life. Bicultural identity is common among middle-class, educated Arabs, but even in this category Westernization is evident, mainly in the emphasis on materialistic aspects of living rather than on their social relationships (Al-Sabaie, 1989). The Westernized group is usually composed of Arabs who immigrated to Western countries.

The major difference between Moslems and Druze is the level of acculturation. Close to 1 million Moslems live in Israel, and most of them may be considered bicultural. The Druze, in contrast, numbering about 94,000 in Israel, are known as an ethnic group that is furthest removed from the mainstream lifestyle in Israel and considered to have the greatest solidarity in the Middle East. They differ from the Moslems in their religion, ethnicity, geography, and political stands. They are a minority in several Arab countries in the Middle East, including Israel, and believe that keeping up with their tradition is the basis for their survival (Dana, 1998). Therefore, they are very strict in their socialization of the young generation (Phalet & Claeys, 1993).

The differences between Arab (Moslem, Druze) and Western cultures are reflected also in Arab attitudes toward counseling. Arabs tend to believe in external natural factors or supernatural factors as causes for pathology (Timimi, 1995). On the whole, only extreme cases of pathology are treated, in an attempt to avoid stigmatizing, and Arabs wait a longer time before seeking help (Dwairy, 1998; Okasha, 1993). When they do pursue counseling, they expect it to be direct, to be short-term, and to offer advice. The traditional focus on verbal exchange and insight is often experienced as intrusive. Therefore, outreach counseling and short-term, goal-oriented, systematic eclecticism is offered for counseling Arabs (Budman, Lipson, & Meleis, 1992; Dwairy, 1998). Yet, owing in part to social oppression, social change, and family transformation, Arabs tend to display high levels of psychosocial disorders (Ibrahim & Ibrahim, 1993; Okasha, 1993). The Israeli Ministry of Education has opened many counseling centers in Arab residential areas, and increasing numbers of school counselors and other helping professionals work with the Arab population. Dwairy, a clinical psychologist, has noticed a dramatic change from the 1980s to the present in referrals to counseling, including children.

On the basis of these cultural restrictions, we expected that Arab adolescents would disclose private information less than would Jewish adolescents, particularly in matters related to family and close friends. Indeed, a previous study that compared attitudes of Arab and Jewish adolescents toward self-disclosure in groups (Hadad, Nechas, & Shechtman, 1998) pointed to a difference on two dimensions: family and friends; for both, Arab adolescents held less favorable attitudes toward self-disclosure.

SELF-DISCLOSURE IN GROUP

Self-disclosure is a key element in group counseling and group psychotherapy (Corey & Corey, 1992; Dies, 1993; Yalom, 1995). Group counseling is defined as a therapeutic intervention for persons with special psychosocial needs that is aimed at assisting in accomplishing developmental tasks and increasing personal development and adjustment. It is distinguishable both from educational groups, aimed at prevention and training, and from therapy groups, designed for persons with severe behavior problems (Conyne, Wilson, & Ward, 1997; Gazda, 1989; Gazda, Ginter, & Horne, 2001; Gladding, 1995).

Counseling groups are particularly recommended for adolescents struggling with developmental tasks, such as identity formation, establishing independence, and dealing with peer relations (Erikson, 1963; MacLennan & Dies, 1992; Rose, 1998). In the group process, adolescents share personal concerns in a supportive climate, realizing that many of their problems are universal and can be worked out. They learn to know themselves better through the interaction with group members and the feedback provided by others. They gain strength from the cohesiveness of the group and improve their problem-solving skills through guidance, modeling, and the development of insight (Crouch, Bloch, & Wanlass, 1994).

Because psychological factors interact with the social context in determining behavior in group counseling (Dwairy, 1998; Lee, 1997), basic behaviors such as self-disclosure and group interaction are affected by cultural factors. The basic assumptions regarding group counseling processes have been established on the foundation of Western concepts. Accordingly, the individual is the center of treatment, spontaneous verbal interaction is encouraged, and risk-taking behavior is expected, to name but a few examples of necessary behavior (i.e., behavior that helps the group process move ahead) in the group. All these codes of behavior may be in conflict with cultural values of some groups (DeLucia-Waack, Coleman, & Jensen-Scott, 1992; Sue, Arredondo, & McDavis, 1992). Therefore, race or ethnicity provides a primary context for understanding the concept of culture within counseling practice.

RESEARCH HYPOTHESES

The unique cultural background of Arabs in Israel has led us to believe that Arab adolescents would express some difficulties with self-disclosure, particularly in the areas studied here. Moreover, because of the ethnic and cultural difference between the Moslems and the Druze, also found in studies in Israel on future orientation (Seginer & Halabi, 1991), we expected differences between them as well.

Several hypotheses were generated from these expected differences: (a) In a comparison of the three ethnic groups, Jewish adolescents would show the highest levels of self-disclosure (in regard to self, family, friends, and feelings); Moslem adolescents would be next; and the Druze would rank lowest. Similarly, we anticipated that (b) Jewish adolescents would best respond to self-disclosure; Moslems would be next; and the Druze would be last. Because behavior in the group is supposed to reflect on outcomes (DeLucia-Waack et al., 1992), we further expected that (c) Jewish adolescents would gain the most from the group experience; Moslems would be second; and the Druze would gain the least. Finally, gender differences were investigated, because the literature suggests that gender may be considered yet another culture that influences behavior (Duck & Wright, 1983). However, although research suggests that girls in general disclose more in close relationships than boys do (Sharabany, 1994; Shechtman, 1994), in an earlier study, both sexes were found to display similar levels of self-disclosure in counseling groups (Leichtentritt & Shechtman, 1998). Hence, we hypothesized that (d) no gender differences would be found in self-disclosure or in group interaction in all groups.

METHOD

Participants

Participants were 54 Moslems (30 girls and 24 boys), 48 Druze (34 girls and 14 boys), and 52 Jews (33 girls and 19 boys), all in the 9th and 10th grades (ages 14-15 years). There were five counseling groups per ethnicity, and each group was run by the school counselor. The Moslem groups were held in three schools belonging to villages near major cities in northern Israel; the Druze groups were held in four schools in four different villages; and the Jewish groups were held in five schools in the north of Israel.

All participants were referred to the school counseling services for a variety of psychosocial difficulties such as isolation, rejection, aggression, and discipline problems. The referral procedure was similar in all the schools, which are all under the auspices of the Israeli Ministry of Education. Furthermore, all the counselors involved were graduates of the same university school counseling program and had been trained in interpersonal and supportive child group counseling. Altogether, there were 12 counselors (some led two groups)--11 female counselors and 1 male counselor; 6 Jews, 3 Arabs, and 3 Druze--ranging in age from 28 to 46 years and in experience from 4 to 12 years.

Variables and Measures

Self-disclosure. We defined self-disclosure as any verbal interaction in which a person reveals some private information about his or her life or feelings. The Rating Scale for Self-Disclosure in Preadolescents (Vondracek & Vondracek, 1971) was used to measure the frequency and level of self-disclosure. The original scale has eight dimensions: family, friends, self, transgression, evaluation of own performance, expression of feelings, activities, and attitudes. In an attempt to reduce the number of variables, and following what was done in an earlier study (Hadad et al., 1998), we selected only four for the present study: family, friends, self, and feelings, which seemed most relevant to the cultural difference under investigation. The selection of just some of the dimensions, however, was not expected to affect the psychometric properties of the instrument, because its strength relies primarily on interrater agreement. Each dimension is coded on two levels of self-disclosure: simple (Level 1) and intimate (Level 2). The levels for each variable are described in the following sections.

Family-Level 1 includes statements that provide routine information about family members and their tastes, interests, activities, and possessions (e.g., location of home, father's profession). Family-Level 2 includes statements yielding more personal data about family members, their personalities, misbehavior, and personal problems (e.g., father's illness, mother's temper, difficulties with the law).

Friends-Level 1 includes statements incorporating routine information about friends, their tastes, interests, activities and possessions (e.g., "my friend lives close by," "my friend has a dog"). Friends-Level 2 includes statements yielding more private disclosures, especially information about personal characteristics, descriptions of incidents, and references to antisocial behavior.

Self-Level 1 includes routine data about the self, including description of personal possessions and of plans and desires of a nonintimate nature (e.g., school schedule, plans to join a club). Self-Level 2 includes descriptions of private fantasies and dreams, personal wishes of an intimate nature, and reports of undesirable habits or improper behavior (e.g., difficulty to control temper, admitting being unpopular).

Feelings-Level 1 includes the expression of moderate or mildly embarrassing feelings (e.g., "I like my teacher," "I feel restricted at home"). Feelings-Level 2 includes reports of major humiliations that seem to have made a permanent impression or caused considerable suffering, major grief experiences, and major disappointments (e.g., "teased by peers about appearance," "unhappy because of sense of failure," "feeling abused"). It should be noted that this particular subscale was originally divided into three levels of disclosure. For this study, we eliminated the simplest level; hence, the first level already entails a moderate level of disclosure.

The analysis, by frequency and level of self-disclosure in each dimension, was based on transcripts of all sessions. The unit of analysis was a paragraph that described one self-disclosure episode. Reliability was established through the comparison of ratings between two independent and trained raters on some of the sessions (2 sessions of six groups--2 Moslem, 2 Druze, and 2 Jewish--randomly selected; a total of 12 sessions). Agreement between raters was high ([kappa] = .95, p < .001).

Response to self-disclosure. This variable refers to the immediate verbal reaction to a disclosure statement. The Structured Analysis of Social Behavior (SADB; Lora Benjamin; see MacKenzie, 1990, pp. 294-297) was used for the analyses. It consists of two principal axes: positive/negative affiliation and independence/interdependence. Because we were interested only in the relationship focus, we used the affiliation rating. Positive affiliation included affirming and understanding, loving and approaching, and nurturing and protecting; the three negative affiliations were ignoring and neglecting, attacking and rejecting, and belittling and blaming. The reported internal consistency of the scale ranged from .82 to .97. In our sample, interrater agreement was high ([kappa] = .98, p < .001).

Adjustment problems. We measured adjustment problems through the Child Behavior Checklist (CBCL; Achenbach & Adelbrock, 1991). This self-report scale contains 113 items in eight dimensions: withdrawal, anxiety, concentration, somatic disorders, delinquency, social problems, thinking, and aggression. Responses were on a 3-point scale (0, 1, 2), the higher score representing more adjustment problems. This is a popular scale, with excellent psychometric properties, that has been widely used in Israel (e.g., Shechtman, 2000). The internal consistency for this scale was high in both the Arab population (Moslems and Druze combined; Cronbach's alpha = .92) and the Jewish population (.91). The Arab adolescents completed the Hebrew Version of the CBCL with no difficulty, because they feel comfortable with the Hebrew language at this stage in their education. Somatic behavior was eliminated from the analyses because of missing data.

Procedure

The counseling groups (N = 15) met in the schools for 7 to 8 months in weekly 45-minute sessions, for a total of 15 sessions. The central mode of communication in the group was verbal, although certain helping devices such as stories and therapeutic games were used to promote group interaction. The style was basically interpersonal, with a strong focus on self-expressiveness and group support. All groups were recorded with permission and transcribed by their group leaders as part of their course duties.

Two independent raters, not involved in leading the groups, received training from two experienced researchers. In the training process, the two raters analyzed the sessions of one counseling group with the help of a third person experienced in coding transcribed data from a previous study. In case of disagreement, a fourth person (the first author) joined the discussion. In general, we followed the training procedure suggested by Clara Hill CHill & O'Brien, 1999). After the training, the two independent raters analyzed all groups. Interrater agreement was high.

RESULTS

Self-Disclosure

Our first hypothesis suggested differences in self-disclosure between the three groups: Jews were expected to disclose more than Arabs (both Moslems and Druze), and Moslems were expected to disclose more than the Druze. Self-disclosure results were obtained by calculating for each group participant the proportion of all his or her verbal statements that involved self-disclosure. Proportions were used to control for individual differences in general talk level. Table 1 presents the means of these proportions for each of the four disclosure dimensions and by level of disclosure (simple vs. intimate).

A 3 x 2 multivariate analysis of variance (MANOVA), with group and gender as between-group variables and the four dimensions as the dependent variables, pointed to a multivariate group effect, MF(8, 290) = 12.11, p < .001; a nonsignificant gender effect, MF(4, 145) = .85, p = ns; and a Group x Gender effect, MF(8, 290) = 2.74, p < .01. Univariate effects were significant for all four dimensions of self-disclosure (see Table 1, the Total on each dimension).

Two additional MANOVA procedures were conducted separately for each level of disclosure. For Level 1, only a multivariate group effect was found, MF(8, 290) = 3.41, p < .001; the gender effect and the Group x Gender effect were both nonsignificant, MF(4, 145) = 0.92 and MF(8, 290) = 0.62, respectively. In addition, univariate effects were all nonsignificant except for Self (Table 1).

Findings for Level 2 indicate a significant group effect, MF(8, 290) = 11.09, p < .001; a nonsignificant gender effect, MF(4, 145) = 0.79, p = ns; and a significant Group x Gender effect, MF(8, 290) = 3.72, p < .001. In contrast to Level 1, all the univariate group effects for Level 2 were significant (see Table 1). However, in respect to the Group x Gender effect, only one interaction was found, for Feelings, F(2, 148) = 8.55, p < .001; not in Table 1.

Because gender difference was consistently nonsignificant, and the group by gender effect was based on only one dimension (Feelings), we concluded that gender difference was marginal and so did not report it. Table 1, therefore, presents only the means of the proportion of self-disclosure per group (and not gender).

The data in Table 1 suggest, in general, that a high proportion of talk was of a disclosing nature. For Moslem adolescents, 60% of the talk involved self-disclosure; for the Jewish adolescents, it was 40%; and for the Druze, it was 29%. The highest rate of self-disclosure was in Self for all three ethnic groups.

The findings also suggest that self-disclosure on Level 2 was higher than on Level 1 in almost all dimensions and for all ethnic groups. For example, disclosure of Self among Moslems was 3 times more frequent on Level 2 than on Level 1 (18% vs. 6%); for Family, it was 13% compared with 3%. Even for the Druze, the difference is in a similar direction for these two dimensions. Only on Friends and Feelings was there no difference for the Druze group.

We conducted a series of 3 x 2 analysis of variance (ANOVA) procedures on each of the four dimensions to test the difference between the two levels of self-disclosure for the three ethnic groups. Results indicated a significant difference in the level of disclosure for each dimension, with interaction effects between level and group. For Self: F(1, 148) = 146.34, p < .001; interaction effect, F(2, 148) = 6.85, p < .01. For Family: F(1,148) = 102.62, p < .001; interaction effect, F(2, 148) = 10.83, p < .001. For Friends: F(1, 148) = 51.53, p < .001; interaction effect, F(2, 148) = 9.11, p < .001. For Feelings: F(1,148) = 21.07, p < .001; interaction effect, F(2, 148) = 10.02, p < .001. In sum, the group participants expressed themselves on an intimate level more than on a simple level on all dimensions, although variations existed between the ethnic groups.

To address our first hypothesis, the revealed between-group difference was insufficient. The post hoc analyses indicated the direction of difference and rejected our hypothesis that the Jewish adolescents would rate higher on self-disclosure. In contrast to our expectations, it was the Moslem group that showed the highest level of self-disclosure. Only on a few variables were they equal to the Jews. The second part of the hypothesis, suggesting that the Druze would display the lowest level of self-disclosure, was mostly supported. Yet, on several of the variables, the Druze were not lower than the Jews in self-disclosure, although they were consistently lower than the Moslems (see Table 1).

Response to Self-Disclosure (Affiliation)

Our second hypothesis suggested similar group differences in participants' verbal reactions to self-disclosure. It was expected that Jews would react more frequently and positively than would Moslems and Druze and that Druze would be the least positively interactive. Again, we calculated proportions of all verbal statements (positive and negative) that were responses to self-disclosure. Table 2 presents the means of proportions and the differences between the ethnic groups.

In general, positive responses (ranging from 63% to 76%) were more frequent than negative ones (ranging from 12% to 20% of all responses). A 3 x 2 ANOVA procedure confirmed the difference between positive and negative affiliation responses, F(1, 148) = 164, p <. 001, with no Affiliation x Group interaction, F(2, 148) = .70, p = ns. In other words, all of the groups were more positive than negative.

A 3 x 2 MANOVA for group and gender, on the six responses, revealed significant group differences, MF(4, 294) = 2.48, p < .05, but no gender or Gender x Group differences. Only two univariate group differences were found: in loving and approaching and in ignoring and neglecting (see Table 2). In the general positive affiliation, the difference was marginal (p = .06).

Again, the direction of difference was contrary to our expectations: The Moslem adolescents were both more loving and approaching and more ignoring and neglecting. There were no differences between Jewish and Druze adolescents on these dimensions. In sum, the second hypothesis was rejected as well; results were in line with those found regarding the first hypothesis.

Gains Resulting From Group Experience

Our third question referred to outcomes. It was hypothesized that Jews would gain the most, Moslems would be second, and the Druze last. Table 3 reports gains for each group and between-group differences for both prescores and gains.

It is interesting that there were several prescore differences (i.e., Moslems and Druze showed higher rates of adjustment problems than did the Jews), whereas only one difference was found in terms of gains (Moslems and Jews gained more than the Druze regarding delinquency). Thus, the groups gained similarly from the group experience, even though Moslems and Druze seemed to have had more adjustment problems initially. It is also interesting to note that anxiety and social problems were affected positively in all three groups; attention and aggression in two groups; withdrawal only among the Druze; and cognitive problems only for the Moslems.

Finally, to investigate the prediction of outcomes by self-disclosure, a stepwise regression analysis was conducted on the total population. The prediction was marginal, F(1, 133) = 3.00, p = .08; only Self contributed significantly to it, F(1, 133) = 3.63, p = .05. This same procedure was also conducted for each group separately. Only for the Moslems did self-disclosure predict gains, F(1, 53) = 7.61, p < .01, with both Self and Friends contributing significantly to the prediction, F(1, 53) = 7.61, p < .01, and F(1, 53) = 4.24, p < .05, respectively.

DISCUSSION

Self-disclosure is a central behavior in group counseling, which is fundamental to all therapeutic factors in group work (Corey & Corey, 1992; Dies, 1993; Yalom, 1995). Yet, this behavior is particularly sensitive to cultural differences when comparing collectivist, authoritarian societies with individualist, democratic ones (Dwairy, 1998). It was logical to expect Jewish adolescents, who grow up in an individualist culture, to self-disclose more than Arabs, who grow up in a collectivist and authoritarian society and are expected to repress their feelings and remain loyal to their families and friends. This rationale was particularly applicable to sensitive areas such as family and friends, in which the requirement to disclose would conflict with basic Arab values (Hadad et al., 1998).

However, results of the present study rejected the basic hypotheses; in fact, it was the Moslem adolescents who showed the highest level of self-disclosure in all four dimensions. The Druze, on the other hand, indeed showed the lowest rates of self-disclosure.

This difference between the two groups of Arab adolescents may be ascribed to differences in their level of acculturation. The Druze are known as the most traditional group and least acculturated of Arabs in Israel, and perhaps in the Middle East (Dana, 1998). They are geographically remote and culturally removed from the Israeli mainstream. As a result, they have little exposure to individualist and democratic values and are more reserved, particularly in expressing feelings and discussing friends.

Moslems, in contrast, are closer to the Jewish culture and more bicultural as a group. Under such circumstances, their unexpected higher level of self-disclosure than that of the Jews may be attributed to both their strong need for support with developmental tasks and the enabling climate of the group. All adolescents struggle with developmental issues and often need the support of a peer group (Rose, 1998), but Moslem adolescents, living in a bicultural context, go through transition processes in their family, school, and community and may have greater need of assistance (Dwairy, 1998). The behavior of the adolescents in this study does not seem to reflect a linear pattern of self-disclosure--from traditional to bicultural to Westernized. If cultural restrictions are very strong (as they are among the Druze), they do not dare to resist them; but, if they are moderately strong, owing to a bicultural identity, participants take advantage of the new possibilities offered by group counseling.

It is interesting that self was the most frequent topic of discussion in all ethnic groups. This reflects both the developmental needs of adolescents to form a sense of self-identity (Erikson, 1963) and the trend of personal growth typical of group counseling (Conyne et al., 1997; Gazda, 1989; Gladding, 1995). The high frequency at which self was discussed in the Moslem adolescent group (24% of all their disclosures) may reflect the identity conflicts of bicultural Arab Israelis in transition stages as well as their conflict between tradition and other social (i.e., Western) influences.

In contrast, talk about feelings was relatively low in general and extremely low among the Druze. The difficulty of speaking about feelings in front of strangers has been acknowledged in the Arab culture, particularly in the more traditional ethnic groups (Dwairy, 1998).

Beyond between-group differences, it should be noted that all groups preferred to disclose on a deeper, intimate level than on a superficial one. Overall, the disclosure on Level 2 was significantly higher than on Level 1. This suggests that participants took the counseling group seriously and were prepared to invest effort in it. It also reinforces the frequent recommendation to work with adolescents in groups (Cramer-Azima, 1989; MacLennan & Dies, 1992; Rose, 1998).

The effective performance of all participants was supported by analyses of the responses to self-disclosure. Overall, positive responses outnumbered negative ones, and group differences were rather minimal. Thus, although not all adolescents were able to self-disclose, their response to others' disclosure reflects a positive attitude.

Support for the positive performance of these groups can also be drawn from the gains resulting from treatment. Although the Arab adolescents (both Moslem and Druze) initially exhibited more adjustment problems, there were almost no differences in gains. These findings suggest that all three ethnic groups, despite differences in self-disclosure, are capable of both interacting effectively in a counseling group and gaining from the group experience. In other words, the higher level of difficulty to disclose did not seem to affect either performance or outcomes. It seems that other processes or therapeutic factors operate, even when self-disclosure is relatively low. Yet, the finding that Moslem group outcomes were predicted by self-disclosure, particularly by the discussion of self, suggests that disclosure does play an important role in counseling groups.

A final word on gender differences. Despite the vast literature that suggests such differences in intimacy and self-disclosure, we did not find them in the group process, which is in line with the results of a previous study (Leichtentritt & Shechtman, 1998). We attribute this finding to the unique characteristics of groups. Group process creates different group norms, such that boys and girls feel equally free to express themselves. These norms are very different from the real world of adolescents. These findings suggest that boys can adapt to norms of self-disclosure and positive affiliation once the climate permits and encourages it.

IMPLICATIONS, LIMITATIONS, AND CONCLUSIONS

Implications

One of the important implications of the current study is the danger of overgeneralizing cultural characteristics. Culturally sensitive counseling has to foresee possible conflicts between ethnic and therapeutic values. Self-disclosure, a necessary behavior in group counseling, may conflict with the restrictions imposed by tradition and values of collectivist societies, particularly in such sensitive areas as family, friends, and expression of feelings. In the United States as well as in other Western countries, many ethnic groups still hold the values of collectivist cultures (e.g., Asians; Conyne, Wilson, & Tang, 2000). But this study has shown that they are clearly not the same. Careful attention should be given to the unique characteristics of each subculture that are often associated with the level of acculturation (Dwairy, 1998; Lee, 1997). Indeed, being aware of the client's worldview is one of the three major principles of multicultural counseling competencies (Sue et al., 1992).

A second implication of our study is related to the connection between developmental needs and group counseling in collectivist cultures. How should professional counselors deal with cultural sensitivity and restrictions against group counseling? Should we avoid the discussion of areas of need for adolescents in such cultures? The results of this study suggest that group counseling is not only needed but is also effective in collectivist societies. Even though such adolescents often express reservations regarding group counseling (Hadad et al., 1998), they do become engaged in the group process and even gain from the experience. Because it is expected that these adolescents are not as likely to reach out for counseling, counseling should be offered to them. Group processes seem to be stronger than cultural restrictions in helping adolescents to deal with their developmental tasks.

Nonetheless, there are certain practical aspects that must be taken into account. For one thing, in groups that express reservations regarding self-disclosure, it is necessary to adjust leadership style, methods of intervention, and techniques (Conyne et al., 2000; Sue et al., 1992). In the more authoritarian ethnic groups, the leader is perceived as the authority figure and responsible for all group processes. He or she must be aware of such expectations and address the issue through clarification of goals and expectations. In some groups, the leader may use more structure in the form of varying therapeutic activities to help participants engage in an open conversation. In any event, it is important to approach the issue of self-disclosure with caution, to resist applying pressure to disclose and especially to avoid mislabeling such difficulties in performing as "resistance."

In addition, group composition may have an effect on participants' reservations regarding self-disclosure. In this study, groups were homogeneous in ethnicity, including the group leader. Such composition could of course ease the difficulties to disclose in a group setting. Indeed, homogeneous groups showed higher rates of cohesion and greater satisfaction with counseling (Perrone & Sedlacek, 2000). Future research should explore the impact of culture on counseling sessions with groups in which the participants are multicultural.

A third implication of the current study is the relevance of group factors other than self-disclosure in counseling sessions. Indeed, the limited connection between self-disclosure and outcomes and the lack of differences in the expression of affiliation suggest that there are other therapeutic factors operating in the group process to influence outcomes. In other words, notwithstanding the high appreciation for self-disclosure in group counseling, it is quite possible that, in groups where self-disclosure is restricted by cultural values, it is better not to push in this direction, because other factors may be at work in the group learning process. Such adolescents may take advantage of modeling, guidance and advice, or socializing techniques more than of cathartic experiences. This may, of course, require flexibility on the part of the leader, especially if he or she was trained in interpersonal or psychodynamic theories. Indeed, multicultural counseling competencies include awareness of the leader's own culture, values, and biases (Sue et al., 1992). Future research should indeed focus on differences between counseling groups of varying cultures in respect to the importance of therapeutic factors in the group process.

Limitations

Some limitations of the study need to be taken into account. First is the relatively small number of participants in each ethnic group. Although about 50 participants is a respectable size in group research, a larger number of participants and groups would permit us to take group effects into account, because the interdependence of the participant and the group in counseling has been recognized. Second, gains were measured by self-reports only. Additional measures are required to establish the effectiveness of the intervention. Finally, only some of the dimensions on the self-disclosure instrument were applied in this study, which might have affected its psychometric properties.

Conclusions

Overall, all participants in the study, regardless of culture, took advantage of group counseling and gained from the group process. This in itself should encourage the use of group counseling in all ethnic groups. However, sensitivity to cultural differences is required because an approach that is too demanding may cause internal conflicts for participants. Under such circumstances, the knowledge and skills of counselors in enhancing behavior in sensitive areas is of utmost importance. Notwithstanding the aforementioned limitations, this study has revealed interesting and unexpected results.

TABLE 1

Means of Proportions and Standard Deviation for Self-Disclosure
(Total and Levels 1-2) by Group (Moslems, Jews, Druze),
F Values, and Post Hoc Analyses

                         Moslems      Jews      Druze
                           (A;         (B;       (C;
                         n = 54)     n = 52)    n = 48)

Variable                 M     SD    M     SD    M     SD

Self total              .24   .11   .13   .07   .12   .06
  Self-Level 1          .06   .05   .04   .04   .02   .03
  Self-Level 2          .18   .09   .10   .06   .10   .06
Family total            .16   .12   .08   .06   .07   .05
  Family-Level 1        .03   .04   .02   .02   .02   .02
  Family-Level 2        .13   .10   .07   .05   .06   .05
Friends total           .08   .06   .09   .07   .04   .04
  Friends-Level 1       .02   .04   .02   .02   .01   .02
  Friends-Level 2       .06   .05   .07   .06   .02   .04
Feelings total          .12   .08   .09   .08   .06   .04
  Feelings-Level 1      .03   .04   .03   .03   .03   .04
  Feelings-Level 2      .08   .06   .06   .06   .03   .03
Self-disclosure total   .60   .21   .40   .19   .29   .13

Variable                F(2, 148)    Post Hoc

Self total              29.58 ***   A > B = C
  Self-Level 1          12.04 ***   A > B > C
  Self-Level 2          20.72 ***   A > B = C
Family total            16.08 ***   A > B = C
  Family-Level 1         2.01
  Family-Level 2        15.68 ***   A > B = C
Friends total           11.93 ***   A = B > C
  Friends-Level 1        0.96
  Friends-Level 2       13.20 **    A = B > C
Feelings total           8.87 ***   A > B > C
  Feelings-Level 1       0.33
  Feelings-Level 2      13.87 ***   A > B > C
Self-disclosure total   35.14 ***   A > B > C

Note. Level 1 = simple self-disclosure; Level 2 = intimate
self-disclosure.

** p < .01. *** p < .001.

TABLE 2

Means of Proportions and Standard Deviation of Responses
to Self-Disclosure by Group (Moslems, Jews, Druze),
F Values, and Post Hoc Analyses

                                    Moslems      Jews       Druze
                                      (A;         (B;        (C;
                                    n = 54)     n = 52)     n = 48)

Variable                           M     SD    M     SD    M     SD

Positive affiliation              .76   .29   .70   .39   .63   .41
  Affirmation and understanding   .24   .18   .30   .27   .28   .29
  Loving and approaching          .29   .23   .14   .21   .09   .22
  Nurturing and protecting        .23   .20   .25   .25   .26   .26
Negative affiliation              .16   .19   .20   .32   .12   .18
  Ignoring and neglecting         .04   .09   .01   .04   .00   .02
  Attacking and rejecting         .08   .14   .13   .26   .08   .14
  Belittling and blaming          .05   .09   .06   .17   .03   .09

Variable                          F(2, 148)    Post Hoc

Positive affiliation               2.73
  Affirmation and understanding    1.71
  Loving and approaching          11.49 ***   A > B = C
  Nurturing and protecting         0.20
Negative affiliation               0.60
  Ignoring and neglecting          5.69 **    A > B = C
  Attacking and rejecting          0.59
  Belittling and blaming           0.19

** p < .01. *** p < .001.

TABLE 3

Results of t Tests and Multivariate Analysis of Variance on the
Child Behavior Checklist (CBCL) by Group

               Moslems (A; n = 42)       Jews (B; n = 45)

CBCL
Dimension     Pre   Post   t(1, 41)    Pre   Post   t(1, 44)

Withdrawal    .77    .74     0.59      .64    .59     1.54
              .39    .28               .32    .31
Anxiety       .79    .69     2.92 *    .60    .46     4.22 **
              .41    .38               .34    .29
Social        .50    .40     2.73 *    .43    .34     2.45 *
              .30    .24               .31    .25
Cognitive     .77    .62     2.45 *    .50    .43     1.54
              .35    .36               .30    .29
Attention     .83    .80     0.83      .63    .49     3.42 *
              .31    .30               .30    .28
Delinquency   .39    .34     1.63      .31    .27     1.62
              .25    .23               .18    .17
Aggression    .63    .57     1.53      .51    .44     2.27 *
              .28    .30               .32    .26
Total         .62    .56     2.65 **   .50    .41     4.99 ***
              .21    .21               .19    .15

                 Druze (C; n = 48)       Prescore Difference

CBCL
Dimension     Pre   Post   t(1, 47)       F         Post Hoc

Withdrawal    .81    .74     4.01 *     3.50 *     A = C > B
              .27    .33
Anxiety       .82    .71    23.37 ***   4.86 **    A = C > B
              .34    .31
Social        .55    .51     5.16 *     1.63       A = B = C
              .30    .29
Cognitive     .68    .64     2.29       7.39 ***   A = C > B
              .33    .36
Attention     .87    .76    15.86 **    8.34 ***   A = C > B
              .31    .34
Delinquency   .39    .45     0.25       2.46       A = B = C
              .20    .28
Aggression    .64    .60     6.16 *     2.43       A = B = C
              .34    .28
Total         .65    .61     2.10 *     7.62 ***   A = C > B
              .20    .20

                 Gain Difference

CBCL
Dimension       F       Post Hoc

Withdrawal    0.12     A = B = C

Anxiety       0.36     A = B = C

Social        1.11     A = B = C

Cognitive     1.34     A = B = C

Attention     1.60     A = B = C

Delinquency   3.72 *   A = B > C

Aggression    0.22     A = B > C

Total         1.47     A = B = C

* p < .05. ** p < .01. *** p < .001.

REFERENCES

Achenbach, T. M., & Adelbrock, C. (1991). The Child Behavior Checklist (CBCL). Burlington: University of Vermont, Center for Children, Youth & Families.

Al-Sabaie, A. (1989). Psychiatry in Saudi Arabia: Cultural perspectives. Transcultural Psychiatric Research Review, 26, 245-262.

Barakat, H. (1993). The Arab world. Berkeley, CA: University of California Press.

Budman, C. L., Lipson, J. G., & Meleis, A. I. (1992). The cultural consultant in mental health care: The case of an Arab adolescent. American Journal of Orthopsychiatry, 62, 359-370.

Conyne, R. K., Wilson, R. F., & Tang, M. (2000). Evolving lessons from group work involvement in China. The Journal for Specialists in Group Work, 25, 252-268.

Conyne, R. K., Wilson, R. F., & Ward, D. E. (1997). Comprehensive group work: What it means and how to teach it. Alexandria, VA: American Counseling Association.

Corey, M. S., & Corey, G. (1992). Group process and practice (4th ed.). Pacific Grove, CA: Brooks/Cole.

Cramer-Azima, F. (1989). Confrontation, empathy, and interpretation: Issues in adolescent group psychotherapy. In E J. Cramer-Azima & L. H. Richmond (Eds.), Adolescent group psychotherapy (pp. 3-9). Madison, WI: International Universities Press.

Crouch, E. C., Bloch, S., & Wanlass, J. (1994). Therapeutic factors: Interpersonal and intrapersonal mechanisms. In A. Fuhriman & G. M. Burlingame (Eds.), Handbook of group psychotherapy (pp. 269-318). New York: Wiley & Sons.

Dana, N. (1998). Hadruzim [The Druze]. Ramat Gan, Israel: Bar-Ilan University Press.

DeLucia-Waack, J. L., Coleman, V. D., & Jensen-Scott, R. L. (1992). Cultural diversity in group counseling. Journal for Specialists in Group Work, 17, 194-195.

Dies, R. R. (1993). Research on group psychotherapy: Overview and clinical application. In A. Alonso & H. I. Swiller (Eds.), Group therapy in clinical practice (pp. 473-518). Washington, DC: American Psychiatric Press.

Duck, S., & Wright, S. P. (1983). Re-examining gender differences in same-gender friendships: A close look at two kinds of data. Sex Roles, 28, 709-727.

Dwairy, M. A. (1998). Cross-cultural counseling: The Arab-Palestinian case. New York: The Haworth Press.

El-Rufaie, O. E., & Absood, G. H. (1993). Minor psychiatric morbidity in primary health care: Prevalence, nature and severity. The International Journal of Social Psychiatry, 39, 159-166.

Erikson, E. H. (1963). Childhood and society (2nd ed.). New York: Norton.

Gazda, G. M. (1989). Group counseling:A developmental approach. Boston: Allyn & Bacon.

Gazda, G. M., Ginter, E. J., & Home, A. M. (2001). Group counseling and group psychotherapy: Theory and application. Needham Heights, MA: Allyn & Bacon.

Gladding, S. T. (1995). Group work: A counseling specialty (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall.

Griefat, Y., & Katriel, T. (1989). Life demands Musayara: Communication and culture among Arabs in Israel. In S. T. Toomey & F. Korzenny (Eds.), Language, communication, and culture (pp. 121-138). London: Sage.

Hadad, L., Nechas, I., & Shechtman, Z. (1998). Amadot clapay hasifa: Hashvaa bein mitbagrim jehudim varavim [Attitudes toward self-disclosure: A comparison of Arab and Jewish adolescents]. Hayoetz Hachinuchi, 7, 76-87.

Hill, C. E., & O'Brien, K. M. (1999). Helping skills. Washington, DC: American Psychological Association.

Ibrahim, A. S., & Ibrahim, R. M. (1993). Is psychotherapy really needed in non-Western cultures? The case of Arab countries. Psychological Reports, 72, 881-882.

Jackson, M. L. (1997). Counseling Arab Americans. In C. C. Lee (Ed.), Multicultural issues in counseling: New approaches to diversity (2nd ed., pp. 333-353). Alexandria, VA: American Counseling Association.

Lee, C. C. (Ed.). (1997). Multicultural issues in counseling: New approaches to diversity (2nd ed.). Alexandria, VA: American Counseling Association.

Leichtentritt, J., & Shechtman, Z. (1998). Therapist, trainees, and child verbal response modes in child group therapy. Group Dynamics, 2, 36-47.

MacKenzie, K. R. (1990). Introduction to time-limited group psychotherapy. Washington, DC: American Psychiatric Press.

MacLennan, B. W., & Dies, K. R. (1992). Group counseling & psychotherapy with adolescents (2nd ed.). New York: Columbia University Press.

Nydell, F. A. (1987). Understanding Arabs: A guide for Westerners. Yarmouth, ME: Intercultural Press.

Okasha, A. (1993). Psychiatry in Egypt. Psychiatric Bulletin, 17, 548-551.

Pedersen, P. B. (1991). Multiculturalism as a generic approach to counseling. Journal of Counseling & Development, 70, 6-12.

Perrone, K. M., & Sedlacek, W. E. (2000). A comparison of group cohesiveness and client satisfaction in homogeneous and heterogeneous groups. The Journal for Specialists in Group Work, 25, 243-251.

Phalet, K., & Claeys, W. (1993). A comparative study of Turkish and Belgian youth. Journal of Cross-Cultural Psychology, 24, 319-343.

Rose, S. R. (1998). Group work with children and adolescents. Thousand Oaks, CA: Sage.

Seginer, R., & Halabi, H. (1991). Cross-cultural variations of adolescents' future orientation: The case of Israeli Druze versus Israeli Arab and Jewish males. Journal of Cross-Cultural Psychology, 22, 224-237.

Sharabany, R. (1994). Intimate Friendship Scale: Review of conceptual understanding and psychological properties. Journal of Social and Personal Relationships, 11, 449-471.

Shechtman, Z. (1994). The effect of group psychotherapy on close same-sex friendships among preadolescent boys and girls. Sex Roles, 30, 829-834.

Shechtman, Z. (2000). Short-term treatment of childhood aggression: Outcomes and process. Psychology in the School, 37, 157-167.

Sue, D. W., Arredondo, P., & McDavis, R. J. (1992). Multicultural counseling competencies and standards: A call to the profession. Journal of Counseling & Development, 70, 477-486.

Timimi, S. B. (1995). Adolescence in immigrant Arab families. Psychotherapy, 32, 141-149.

Vondracek, F. W., & Vondracek, I. W. (1971). The manipulation and measurement of self-disclosure in preadolescents. Merrill-Palmer Quarterly, 17, 51-58.

Yalom, I. D. (1995). The theory and practice of group psychotherapy (4th ed.). New York: Basic Books.

Zipora Shechtman, Aya Hiradin, and Samahar Zina, Faculty of Education, Haifa University, Israel. Correspondence concerning this article should be addressed to Zipora Shechtman, Faculty of Education, University of Haifa, Haifa 31905, Israel (e-mail: ziporas@construct.haifa.ac.il).