homework in Education

RESEARCH

Impact of Group Sandtray Therapy on the Self-Esteem of Young Adolescent Girls

Yu-Pei Shen Tzu Chi University

Stephen A. Armstrong Texas A&M University-Commerce

The effectiveness of group sandtray therapy was examined using a pretest-posttest control group design with young adolescent girls (n ¼ 37) identified as having low self-esteem. A split-plot analysis of variance (SPANOVA) revealed statistically significant differences between participants in the treatment and control groups in self-esteem on five of the six subscales of the Self-Perception Profile for Children. The implications of the findings are discussed.

Keywords: adolescent; group therapy; sandtray therapy; self-esteem

Researchers have identified adolescence as a chaotic and difficult period of time characterized by increased vulnerability, tremendous changes (Brendgen, Vitaro, & Doyle, 2002; Dies, 2000; LeCroy, 2004b), and normative stressors (Hampel & Petermann, 2005). Adolescent girls are particularly at high risk for developing mental health problems (Hampel & Petermann; LeCroy). They seem to be more vulnerable than boys to negative body image, body dissatisfaction, the likelihood of developing an eating disorder (Burrows & Cooper, 2002), and interna- lizing problems such as low self-esteem and depression (Bolognini, Plancherel, & Bettschart, 1996; Martin & Gentry, 1997; Ohannessian, Lerner, & Lerner, 1999). More adolescent girls than boys experience low self-esteem (Birndorf, Ryan, & Auinger, 2005; Carlson, Uppal, & Prosser, 2000; Israel & Ivanova, 2002).

Connor, Doerfler, and Toscano (2004) advocated gender-specific prevention=intervention programs for a variety of issues that ado- lescent girls may encounter. LeCroy (2004a, 2004b) recommended that gender specific programs should be implemented in early adolescence

Yu-Pei Shen is an Assistant Professor in the Department of Child Development and Family Studies at Tzu Chi University in Taiwan. Stephen A. Armstrong is an Assistant Professor in the Department of Counseling at Texas A&M University-Commerce. Correspondence concerning this article should be addressed to Stephen A. Armstrong, Department of Counseling, Education North 202, P.O. Box 3011, Commerce, TX-75429-3011. E-mail: [email protected].

THE JOURNAL FOR SPECIALISTS IN GROUP WORK, Vol. 33 No. 2, June 2008, 118–137

DOI: 10.1080/01933920801977397

# 2008 ASGW

118

prior to the onset of some of the serious problems that many adolescent girls develop such as sexually transmitted diseases and illegal drug use. In addition, the need for prevention programs that strengthen self-esteem (Polce-Lynch, Myers, & Kilmartin, 1998), improve signifi- cant relationships (Moran & DuBois, 2002) and assist the transitional development of adolescent girls has been identified (Rice & Meyer, 1994). A warm, non-confrontational, flexible, creative and activity- based approach can promote the establishment of a therapeutic relationship with adolescents and positively impact their behaviors (Leadbeater, Kuperminc, & Blatt, 1999; Pollock & Kymissis, 2001; Russell & Phillips-Miller, 2002). Some researchers have recommended school-based programs designed to address self-esteem issues in adolescent girls (Block & Robins, 1993; Merwin & Ellis, 2004).

Even though many authors and researchers have recommended pro- grams for the enhancement of self-esteem in early adolescent girls, iden- tifying girls with low self-esteem may be challenging because many girls this age conceal their problems and have difficulty trusting someone to help them (LeCroy, 2004a, 2004b). Therefore, many girls who need help may not seek it. Counselors who work in middle schools and are familiar with the issues and needs of the young adolescent girls whom they serve are in an advantageous position to identify girls with low self-esteem. If a trusting relationship with these girls can be developed, they may be will- ing to participate in programs that enhance self-esteem.

Self-esteem refers to ‘‘the extent to which one prizes, values, approves, or likes oneself’’ (Blascovich & Tomaka, 1991, p. 115). High self-esteem is considered to be a protective factor for mental health problems (Birndorf et al., 2005; DuBois, Burk-Braxton, & Swenson, 2002) and for decreasing overall adjustment difficulties (DuBois et al.). Research findings show that self-esteem typically declines in girls from early to late adolescence (Block & Robins, 1993; Harter, 1999; LeCroy, 2004a, 2004b; Rosenberg, 1985). In addition to the typi- cal declines of self-esteem in girls this age, early adolescent girls from diverse populations may be more at risk for declining self-esteem. Holcomb-McCoy and Moore-Thomas (2001) noted that societal images and stereotypes have negatively affected the self-esteem of young adolescent African American girls. Therefore, it appears that imple- menting preventive and treatment programs for enhancing self- esteem during the middle school years would be strategic.

Activity Group Therapy

Many girls from the age of 11–14 are still in the concrete oper- ational stage of development (Meece, 1997). Children in this stage often have difficulty expressing their feelings adequately through

Shen, Armstrong / GROUP SANDTRAY THERAPY 119

language. Draper, Ritter, and Willingham (2003) noted that children in this age group could ‘‘benefit from a modality that is not completely dependent on verbalization, that is, some experiences may be better accessed through activity such as play, art, or some other form of metaphoric communication’’ (p. 245). Slavson (1943) introduced an activity group approach to therapy that used toys and games. Accord- ing to Slavson, therapists need to convey acceptance and allow parti- cipants to engage in self-directed activities. Activity group therapy emphasizes the expression of feelings and fantasies through activities and play rather than through verbal means (Lomonaco, Scheidlinger, & Aronson, 2000). In addition to the therapist providing a sense of freedom and acceptance, the interactions of children with each other help them to improve communication and learn the rules of peer inter- action, such as taking turns.

Bratton and Ferebee (1999) noted that expressive art activities could be used therapeutically in activity group therapy. Expressive art activities are experiential by nature and provide a unique means of self-expression that works well with children and young adoles- cents. Bratton and Ferebee also stated that structured expressive art activities provide young adolescents with ‘‘opportunities to change perceptions about self, others, and the world as they try out new roles and solutions in the safety of the group’’ (p. 193). One expressive art activity that Bratton and Ferebee used effectively in group work with young adolescents was sandtray therapy.

Group Sandtray Therapy

Group sandtray therapy is an activity-based approach that has been used successfully with children and young adolescents (Draper et al., 2003; Flahive & Ray, 2007; Homeyer & Sweeney, 1998). Group sand- tray therapy is an intervention in which ‘‘group members build small worlds with miniature figures in individual trays of sand and share about their worlds as they are willing’’ (Draper et al., p. 244). The therapeutic effects of sandtray therapy are experienced through the creation of sandtray scenes and the facilitation of a trained therapist to unfold and process the intrapsychic and interpersonal issues of group members (Homeyer & Sweeney).

Because young adolescent girls are developmentally interested in peers, a group therapy approach to sandtray is preferable to individual sandtray (Dies, 2000; Draper et al., 2003; Flahive & Ray, 2007). Group sandtray therapy is physically active and provides children an indirect means to self-disclose material that is painful or uncomfortable. With sandtray, young adolescents can feel a sense of safety and empower- ment (Homeyer, & Sweeney, 1998). Group sandtray therapy offers

120 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

young adolescent girls an opportunity to build significant relation- ships with peers, which is an important developmental task for this age (Moran & DuBois, 2002).

Even though a group approach that uses sandtray is developmen- tally appropriate and has been recommended in the literature, there is little quantitative research to support the effectiveness of this approach, particularly with young adolescent girls. Draper et al. (2003) noted that there was a need for research on group sandtray therapy to determine its effectiveness. Others have stressed the need to provide empirical research results in sandtray therapy and the need to examine the effectiveness of sandtray therapy for different populations and presenting issues (Flahive & Ray, 2007; Kestly, 2001).

There are no studies in the literature that examine the effective- ness of sandtray therapy with diverse populations. However, Cochran (1996) noted that person-centered therapies are beneficial for treating culturally diverse populations of students. Constantine (2001) stated that Hispanics tend to prefer interpersonal relation- ships that are nurturing and loving rather than goal-oriented. Simi- larly, Bailey and Bradbury-Bailey (2007) recommended that group counselors who are not African American should respond to African American adolescents in an accepting and non-judgmental manner. Thus, a therapeutic approach that is warm, supportive and accepting appears to be advantageous when conducting group work with diverse populations.

In the current study, the group therapists used a humanistic theor- etical approach to counseling. A humanistic approach with children is grounded in child-centered theory, which is based on a belief that chil- dren are capable of self-direction given an accepting and supportive environment (Landreth, 2002). In addition, child-centered theory is based on the premise that it is essential for child clients to experience the core conditions in the therapy relationship in order for growth and change to occur. Humanistic therapists believe that participants will experience healing in a safe and contained environment that is full of understanding, acceptance, freedom, and a sense of control (Kalff, 1980). Therefore, a dynamic interpersonal relationship between thera- pist and clients is essential with a humanistic approach. Humanistic therapists also believe that all people possess an inherent actualizing tendency toward growth. In the current study, counselors were instructed to accept all feelings in the group sessions while limiting only those behaviors that were potentially destructive in nature (Landreth, 2002). This approach allowed for cultural differences in communication including ‘‘the emotive and affective quality of inter- personal interactions’’ that are valued by many African Americans (Sue & Sue, 2004, p. 15).

Shen, Armstrong / GROUP SANDTRAY THERAPY 121

Sue and Sue (2004) noted that culturally competent counselors are in the process of becoming aware of their values, biases and precon- ceived notions and how they may affect minority clients. Culturally competent counselors attempt to understand the worldview of cultu- rally different clients and practice culturally sensitive intervention strategies (Sue & Sue). All of the group therapists in the current study had multicultural coursework and previous clinical experience with culturally diverse clients. Through their studies and clinical experience, the therapists developed at least a basic level of cultural competence, including an understanding of how cultural group membership influences the formations of worldviews (Sue, 2001). In supervision, the therapists were challenged to become aware of how their biases might affect their relationships with culturally diverse clients. The therapists also were challenged in supervision to be congruent in their interactions with all clients. Congruence and authenticity is valued highly by African American clients (Sue & Sue).

The researchers limited the group to one grade because Zinck and Littrell (2000) indicated that adolescent girls feel more comfortable expressing themselves with girls of their age. Lomonaco et al. (2000) noted that groups comprised of participants and therapists of the same gender are developmentally beneficial particularly in interpersonal learning. Draper et al. (2003) also recommended same-sex groups with girls in this age group because the flirting that occurs ‘‘when adolescent boys and girls are together can be a distraction to the group work and may inhibit the openness and honest disclosure’’ that group leaders strive to achieve (p. 251).

The purpose of the current study was to examine the effectiveness of group sandtray therapy with seventh grade girls with low self- esteem. Flahive and Ray (2007) recently conducted a group sandtray therapy study with preadolescents with behavioral problems. The cur- rent study extended and modified Flahive and Ray’s approach and examined the effectiveness of this modality with young adolescent girls. Thus, this research resembled Flahive and Ray’s study with a different population, presenting issue, and instrumentation. Flahive and Ray’s study focused on group sandtray therapy with preadoles- cents experiencing behavioral problems. They suggested that their research could be replicated, and that similar research could be conducted on different populations either with group or individual sandtray therapy. Flahive and Ray also recommended limiting group size with children who participate in group sandtray therapy.

In the current study, the effectiveness of group sandtray therapy in enhancing self-esteem in seventh grade adolescent girls was evaluated by comparing scores of the treatment and control groups obtained on

122 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

the Self-Perception Profile for Children (SPPC, Harter, 1985). The following question was addressed: Does group sandtray therapy improve the self-esteem of a sample of seventh grade adolescent girls as evidenced by scores on the SPPC? This research question was tested across all six of the subscales in the SPPC. We hypothesized that the posttest scores on the six subscales of the SPPC of the treatment group would be statistically significantly higher than the posttest scores of the control group.

METHOD

Research Design

This study utilized a quasi-experimental, pretest-posttest design with a treatment group and a control group to determine the effects of sandtray therapy on the self-esteem of young adolescent girls. The treatment group received group sandtray therapy following the pretest, whereas the control group did not receive therapy until participants in the treatment group completed their posttest. Both groups took the pretest and the posttest. The duration of group sandtray therapy was 50 minutes per session, twice a week for nine sessions. Researchers in previous studies recommended that group sandtray therapy sessions last an hour (Flahive & Ray, 2007; Hedges-Goettl & Tannenbaum, 2001); however, in the current study therapists shortened the session time to accommodate the school schedule. The frequency of group meetings (twice a week) has promoted the development of trust in relationships among group members (Tyndall-Lind & Landreth, 2000).

Participants

Participants were 40 seventh-grade adolescent female students from three middle schools in three different cities in the southwest- ern United States. Though 40 participants completed the pretest, three participants were not included in the posttest results due to attrition and unavailability for the posttest. Two of the original part- icipants were not allowed to continue in the project because of ser- ious discipline referrals to the principal. A third original participant was screened out of the project because of a court case involving sexual abuse. The researchers assigned the participants to either the treatment or control group according to the equality of the group means of the total scores on the pretest (M ¼ 97.75, SD ¼ 17.51 for the control group; M ¼ 98.65, SD ¼ 17.13 for the

Shen, Armstrong / GROUP SANDTRAY THERAPY 123

treatment group). In other words, the researchers assigned parti- cipants to the two groups (based on pretest scores) so that the group means would be equal. The 37 remaining female participants included 20 African-Americans (54%), 12 Whites (32%), and 5 His- panic Americans (14%). The treatment group included 9 African Americans, 6 Whites, and 3 Hispanic Americans whereas the control group included 11 African Americans, 6 Whites and 2 Hispanic Americans. Ages ranged from 11 to 13 years, with a median age of 12 years. Of the 37 participants, 26 (70%) were identified as aca- demically at-risk students (students who failed the state mandated criterion referenced test). Twenty-nine participants (78%) qualified for reduced or free lunch, which is an indicator of lower socioeco- nomic status.

School counselors identified seventh grade adolescent girls who pre- sented with low self-esteem and referred them to the therapist assigned to each school. Parents gave written permission for their chil- dren to participate in the study. The therapist in each school screened the referred students individually for approximately 15 minutes. After screening, therapists included all girls who had been referred and obtained written assent from all of them. Participants were informed that the group would meet twice weekly for four and one half weeks for a total of nine sessions.

Therapists

Three female therapists conducted group sandtray therapy in this study. All of the therapists had training in the areas of sandtray ther- apy, group therapy and play therapy. The therapists received additional training in group sandtray therapy from the second author. Two of the therapists were doctoral candidates in counseling and one of these two was a licensed professional counselor. The third therapist was a post-masters licensed professional counselor.

All of the therapists in the current study used a humanistic approach to group sandtray therapy. In addition, all of the therapists had previous experience counseling culturally diverse clients.

Instrumentation

Harter’s Self-Perception Profile for Children (SPPC; Harter, 1985), a self-administered instrument, has been widely used in research to measure feelings of competence in specific domains and overall self- worth (Schumann et al., 1999). The instrument has five subscales that assess perceived domain-specific competence (Scholastic Competence, Social Acceptance, Athletic Competence, Physical Appearance, and

124 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

Behavioral Conduct) and a separate scale that assesses overall self-worth (Global Self-worth). Each subscale on the SPPC consists of:

six items, half of which are reversed with respect to whether the first part of the statement reflects high or low competency. To avoid socially desirable responses, Harter devised a structured alternative format that first asked children to decide which part of a two-part statement describes them best and then asks them to decide if this is really true or only sort of true (Schumann et al., 1999).

Reliability for the SPPC was established with an initial sample of 1543 children in grades 3 through 8, including 754 boys and 789 girls (Harter, 1985). The sample was diverse socioeconomically, but did not include representative proportions of children from different ethnic groups. The SPPC had acceptable reliability with Cronbach’s alpha ranging from .76 to .86 on the six subscales. Schumann et al. (1999) studied the psychometric properties of the SPPC with a socioeconomi- cally diverse sample of 2,267 Black and 2,167 White girls. Cronbach’s alphas were higher for the White girls (.81 to .85) than the Black girls (.71 to .76) on the six subscales.

Procedures

The procedures used in this research were modifications of those used by Flahive and Ray (2007). Participants in the current study were divided into small groups of four. All of the small groups were heterogeneous by race. Given the school setting for the intervention and recommendations in the empirical literature about the number of sessions needed for group therapy to be effective, the researchers decided that nine group sessions were sufficient (Muller, 2000; Young, 1994; Zinck & Littrell, 2000).

The treatment intervention was provided to five groups of four part- icipants. Participants were provided with the opportunity to visually and physically see and touch the sand, miniatures, and sandtray (Draper et al., 2003). After the pretest was administered and parti- cipants were assigned to condition, girls in the treatment group parti- cipated in nine sessions of group sandtray therapy. Participants in the control group received no treatment during this time period. After the collection of posttest data was completed, participants in the control group received the same group sandtray therapy as the treatment group.

Session structure. Structure in group sessions can help reduce anxiety in group members. Structuring group sessions also has been acknowledged as developmentally appropriate for adolescents,

Shen, Armstrong / GROUP SANDTRAY THERAPY 125

because adolescents do not handle ambiguity well (Nichols-Goldstein, 2001). Therefore, a selected topic was structured into each session. Topics were chosen based on the concerns and issues that most ado- lescent girls experience in their development including physical appearance (Keery, Boutelle, & van den Berg, 2005), relationships, and social acceptance and competence (Hedges-Goettl & Tannenbaum, 2001; LeCroy, 2004b; Polce-Lynch et al., 1998). For example, in the friends’ sandtray session, group members were instructed to build a scene of themselves and their friends.

Sandtray room. Therapists set up a sandtray room in each middle school. In each sandtray room, there were four 21-inch round plastic sandtrays that were painted with ocean blue on the inside. Shelves or tables were used to display miniatures. Each group had miniatures in eight categories as recommended by Homeyer and Sweeney (1998) including people, animals, buildings, vehicles, fences and signs, natu- ral items, fantasy, and spiritual-mystical. Duplicates of the miniatures were provided at each school in order to provide participants greater choice in the selection of items.

Sandtray therapy. In the sandtray sessions, participants were instructed on how to create or build their sandtray scenes (Homeyer & Sweeney, 1998). Questions were encouraged if participants were not clear about the instructions. As recommended by Flahive and Ray (2007), therapists used a standard statement for participants to help them build their scenes:

As you can see there are many and various miniatures on the shelves. I would like you to take a look at them. I would like you to build a scene of (the selected topic as described earlier) in your own sandtray by using the sand and miniatures there. You may build your sandtray scene in any way you like and use as many miniatures as you would like. When you are working on building your scene, I will sit here quietly. Let me know when you are finished.

Therapists allowed approximately 15 minutes for each group mem- ber to create her scene in the tray, leaving about 35 minutes available for processing the scenes. However, as the group sessions progressed, participants took longer to build their scenes. During the building phase, the therapist observed how each of the participants approached the miniatures and built her sandtray scene.

After participants had created their scenes in the sandtrays, group lea- ders invited group members to share their sandtray scenes. There were no restrictions regarding the way they shared, nor were they forced to share (Draper et al., 2003). The leader began the processing part of the

126 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

session with a statement such as, ‘‘What would you like to share about your sandtray scene with us? There may be something there that you do not want to share, and that is fine. You may share it in any way you like.’’ Such statements encouraged participants to share at their own pace. The leaders facilitated the processing of sandtray scenes by asking questions such as the following. ‘‘What will you name your sand- tray scene’’? ‘‘Can you tell me more about this (miniature’’? ‘‘What is this (miniature) doing’’? Facilitators used reflective responses as individuals talked about their scenes. The therapists also used group skills such as linking to normalize responses and feelings and to promote universality (Yalom & Leszcz, 2005). More facilitative and elaborative questions such as the questions above may be referenced in Homeyer and Sweeney (1998). The role of the therapist in sandtray therapy is to be a facilitator. The therapist intervened or set limits if negative or harmful comments were delivered in a session (Draper et al., 2003).

Data Analysis

Split-Plot Analysis of Variance (SPANOVA) was utilized in data analyses to determine the effectiveness of group sandtray therapy on self-esteem, and SPSS1 13 computer software was used to run the analysis. SPANOVA is used to analyze data from mixed designs, designs with at least one between-subjects factor and at least one within-subjects factor (Shavelson, 1996).

With SPANOVA, all of the relevant factors in the analysis are ana- lyzed together rather than separately, which comes closer to reflecting the relationship of the factors being analyzed. SPANOVA also reduces the risk of experiment-wise type I error and increases power by partitioning error.

RESULTS

Pre- and post-test scores were obtained on all six of the subscales in the SPPC: Scholastic Competence, Social Acceptance, Athletic Competence, Physical Appearance, Behavioral Conduct and Global Self-worth. The means and standard deviations (SD) for the scores on the six subscales are presented in Table 1.

A SPANOVA was conducted on each subscale; the between-subjects variable was treatment (waiting-list control group and sandtray therapy group), and the within-subjects variable was time (pretest and posttest). Model assumptions of split-plot analysis of variance (SPANOVA), which include independence, normality, homogeneity of variances, and homogeneity of covariances (Shavelson, 1996), were

Shen, Armstrong / GROUP SANDTRAY THERAPY 127

tested and verified. Because numerous statistical tests were conducted in this study, a conservative alpha level of .01 was used in order to reduce the risk of experiment-wise type I error (Armstrong & Henson, 2005). Depending upon the existence of a significant interaction, a fol- low-up analysis of either main effects or simple effects was conducted to further analyze the results. Then effect sizes were calculated to determine the practical significance of the results. Cohen’s d was the effect size used in this study.

SPANOVA results showed that there was a statistically significant interaction in the scores of the Scholastic Competence subscale across treatment and time, F(1, 35) ¼ 13.758, p ¼ .001. The statisti- cally significant interaction is shown in Figure 1. After a statistically significant interaction is found, simple effects rather than main effects are examined. A significant interaction indicates that the results of main effects cannot be generalized. In the current study, main effects would include differences between the treatment and control groups, whereas simple effects would include differences between pretest and posttest scores across the same independent variable. In the treatment group, a statistically significant simple effect was found between pretest and posttest on Scholastic Com- petence, F(1, 17) ¼ 15.132, p ¼ .001, with a medium effect size, Cohen’s d ¼ .68. There was no statistically significant simple effect of the control group (waiting-list group) between pretest and posttest, F(1, 18) ¼ 2.042, p ¼ .17.

Table 1 Means and Standard Deviations for All Six Subscales in the SPPC

Pretest Posttest

Factor Mean SD Mean SD

Control group (n ¼ 19) Scholastic Competence 15.26 4.21 14.11 3.74 Social Acceptance 16.63 5.33 14.42 4.66 Athletic Competence 16.89 4.40 17.05 3.92 Physical Appearance 14.53 5.33 14.00 4.80 Behavioral Conduct 17.95 4.42 15.32 3.74 Global Self-worth 16.68 4.78 16.11 3.63

Treatment group (n ¼ 18) Scholastic Competence 14.56 4.87 17.56 3.93 Social Acceptance 19.28 3.59 20.72 2.67 Athletic Competence 16.00 3.69 18.00 3.14 Physical Appearance 15.89 5.37 18.33 3.82 Behavioral Conduct 16.94 3.65 18.28 2.76 Global Self-worth 17.00 4.35 20.06 2.86

Note. Higher scores reflect greater self-esteem.

128 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

Results showed that there was a statistically significant interaction in the scores of the Social Acceptance subscale across treatment and time, F(1, 35) ¼ 13.607, p ¼ .001. The statistically significant interac- tion is visually depicted in Figure 2. The results of the simple effects of neither the treatment group nor the control group between pretest and posttest were statistically significant.

There was no statistically significant interaction in the scores of the Athletic Competence subscale across treatment and time,

Figure 1 Interaction of scholastic competence across treatment and time.

Figure 2 Interaction of social acceptance across treatment and time.

Shen, Armstrong / GROUP SANDTRAY THERAPY 129

F(1, 35) ¼ 2.769, p ¼ .105. The main effect between the treatment and control groups also was not statistically significant.

There was a statistically significant interaction in the scores of the Physical Appearance subscale across treatment and time, F(1, 35) ¼ 7.53, p ¼ .01. A statistically significant simple effect of the treatment group between pretest and posttest was identified, F(1, 17) ¼ 12.658, p ¼ .002, with a medium effect size, Cohen’s d ¼ .52, but not in the control group across time, F(1, 18) ¼ .404, p ¼ .533, with a negligible effect size, Cohen’s d ¼ .1.

A statistically significant interaction in the scores of the Behavioral Conduct subscale across treatment and time was detected, F(1, 35) ¼ 10.413, p ¼ .003. The results of the simple effect showed that there was not a statistically significant difference in the treatment group between pretest and posttest, F(1, 17) ¼ 2.693, p ¼ .119. There was a statistically significant simple effect of the control group between pretest and posttest, F(1, 18) ¼ 8.269, p ¼ .01, with a medium effect size, Cohen’s d ¼ .64.

Finally, a statistically significant interaction was found in the scores of the Global Self-worth across treatment and time, F(1, 35) ¼ 8.039, p ¼ .008. The results of the simple effect showed that there was a statistically significant difference in the treatment group between pretest and posttest, F(1, 17) ¼ 14.807, p ¼ .001, with a large effect size, Cohen’s d ¼ .83. No statistically significant difference was found in the control group across time.

In summary, there were statistically significant interactions on five of the six subscales of the SPPC. In addition, the effect sizes of the sim- ple effects were notable, indicating practically significant differences in the treatment group over time with no corresponding changes in the control group.

DISCUSSION

Results of the current study indicated that group sandtray therapy was effective in improving the self-esteem of young adolescent girls. There were statistically significant interactions between participants in the treatment group and the control group across time on five of the six subscales of the SPPC: Scholastic Competence, Social Accept- ance, Physical Appearance, Behavioral Conduct and Global Self- worth. As Table 1 indicates, this contrast was the pattern for five of the six subscales: the mean of the treatment group scores increased and the mean of the control group scores decreased. In addition, given that 25 of the 37 participants were not White, it

130 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

appears that group sandtray therapy is effective in improving the self-esteem of young adolescent girls from different cultural backgrounds.

Two of the areas of self-esteem that appear to mean the most to young adolescent girls are perceptions of their social acceptance among peers and their physical appearance. In the control group, the mean of the scores on the Social Acceptance subscale declined by over two points (13%) in five weeks. It appears girls in the con- trol group were experiencing struggles in Social Acceptance while girls in the treatment group were improving. An analysis of the sim- ple effect sizes confirms this trend. The Cohen’s d effect size of the control group declining (.44) is almost identical to the effect size of the improvement in the treatment group (.46). Researchers have identified the link between social support and self-esteem (Merwin & Ellis, 2004; Moran & DuBois, 2002). It appears group sandtray therapy may have prevented or protected girls in the treatment group from experiencing the drop in this aspect of self-esteem that is typical for girls in this age group (Block & Robins, 1993; LeCroy, 2004a).

Block and Robins (1993) noted that positive interpersonal relation- ships are more strongly related to high self-esteem in adolescent girls than boys. The group format in the current study provided girls with an accepting group of peers and opportunities to interact with girls and the female therapist in a safe and supportive environment. In group sandtray therapy, messages that the participants were cared for, respected, unique, and significant were delivered through the nature of sandtray therapy, the dynamics and cohesion of the group, and the therapeutic relationship between the girls and the therapists.

Many young adolescent girls struggle with feelings of inadequacy related to physical appearance. Holcomb-McCoy and Moore-Thomas (2001) noted that African American adolescent females are especially vulnerable to negative feelings related to physical appearance. Burrows and Cooper (2002) noted that many girls are concerned about their weight or shape. In the current study, girls in the treatment group made statistically significant gains in their self-esteem in this area. One of the advantages of same-gender groups with same-gender therapists is the safety that can be experienced by girls when an issue such as physical appearance is the topic of focus.

In the group session that focused on physical appearance, many girls reported that they had negative feelings about their bodies, but many of the participants created scenes that were beautiful. These participants chose miniatures that symbolized beauty, such as princess-like figures, butterflies, flowers, and shining and colorful

Shen, Armstrong / GROUP SANDTRAY THERAPY 131

miniatures. They stated that these beautiful miniatures symbolized their personalities as well as their physical appearance. When the participants chose to focus on the beauty of their personalities rather than their physical attributes, it was different from what the thera- pists expected, but the therapists accepted and reflected the subjec- tive inner worlds of the participants. It appears that girls in the treatment group were able to accept themselves better by combining their views of their physical appearance with their views of their inner beauty. In addition, group members conveyed acceptance of each other’s appearance and personalities. This acceptance appeared to help the girls place less emphasis on negative feelings they had about their appearance and more emphasis on aspects of their personalities they valued.

For example, one girl chose a mirror as a symbol in her sandtray to represent how she felt about her physical appearance. As the therapist focused on her feelings about her body image, other group members encouraged her to focus on her attributes as a person. This feedback appeared to help her accept herself more because the group accepted her physical appearance and valued her strengths as a person. The group’s acceptance appeared to help her improve her self-perception and self-acceptance.

Tuckman and Jensen (1977) proposed five stages of group develop- ment that include forming, storming, norming, performing and adjourning. In the current study, the girls moved quickly through the forming stage into the norming stage. As Tuckman and Jensen noted, the norming phase begins early in the process and overlaps with several other stages. ‘‘Norming includes all of the processes that influence the establishment of rules and boundaries for the group’’ (Armstrong & Berg, 2005). By contrast, none of the groups had an observable storming stage. This dynamic may have been influenced by the style of the therapists and the developmental stage of the group members. Although the physical appearance example above occurred in only one group, the three therapists reported that all of the sand- tray groups had a highly supportive and accepting climate. Given the stages of group development, it was timely to cover the physical appearance topic after the group had moved into the performing stage. Without the deeper level of trust and acceptance that had been estab- lished, group members would have probably been unwilling and unable to self-disclose their feelings about their appearance. As the group moved into the adjourning phase of the group, which was a result of time limitations rather than a professional assessment of readiness for closure, the therapists provided a structured positive feedback exercise in the final session that facilitated an awareness of each girl’s strengths.

132 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

Limitations and Suggestions for Future Research

In the research design, there was a control group but no comparison group. Therefore, it is not clear whether group sandtray therapy was more effective in improving the self-esteem of young adolescent girls than other types of treatment would have been. In addition, the sam- ple used in this study was relatively small. The study would have had greater power if a larger sample had been used. Low power increases the risk of type II error. Statistical significance is heavily influenced by sample size (Armstrong & Henson, 2004). In the absence of fol- low-up data, we are unable to determine whether the observed gains were maintained over time. In addition, the SPPC is a self-report instrument, and the validity of the results depends on whether the participants are honestly answering questions. Even though the therapists in the current study had a basic level of cultural compe- tency, more extensive coursework, training and supervision in work- ing with diverse populations may have resulted in larger gains in self-esteem. Given these limitations, future research is needed to further examine the effectiveness of group sandtray therapy with diverse children and adolescents who have self-esteem issues. Although self-esteem can be difficult to measure, it might be helpful for parents and teachers’ perceptions of change in self-esteem to be measured. Also, larger samples would provide more confidence in the generalizability of the results.

Implications for Group Work

LeCroy (2004a) noted that little attention has been given to finding effective preventive approaches with adolescents. The current study provides a short-term, group intervention that enhanced the self- esteem of young adolescent girls. Given that self-esteem is a protective factor for mental health problems (Birndorf et al., 2005; DuBois et al., 2002) and that low self-esteem is associated with adolescent girls who have histories of suicidal ideation (Hull-Blanks, Kerr, & Robinson Kurpius, 2004), it is critical to implement group programs that enhance self-esteem in early adolescence in order to prevent poten- tially serious mental health issues.

Group sandtray therapy is an economical, developmentally appro- priate modality for young adolescents (Draper et al., 2003). The current study used this modality to address self-esteem in young ado- lescent girls at a time when their self-esteem is vulnerable to intra- psychic and interpersonal factors that tend to diminish it (LeCroy, 2004a, 2004b). Developmentally appropriate and culturally responsive

Shen, Armstrong / GROUP SANDTRAY THERAPY 133

prevention programs and group interventions in schools are needed to address important issues such as self-esteem.

Traditional talk approaches may be effective with certain issues young adolescent girls face, but when dealing with issues such as body image, physical appearance, and peer acceptance, indirect approaches such as sandtray therapy that allow girls to express their perceptions through metaphorical scenes may be less threatening. By increasing the therapeutic distance for young adolescent girls, they may be able to work through sensitive issues that are difficult to discuss openly.

Unfortunately, most university programs only provide training in group work with adults rather than young adolescents (Steen, Bauman, & Smith, 2007), and most university programs do not offer training with expressive art approaches such as sandtray. In fact, many university practica do not include any supervision for working with young adolescents even if the trainee is specializing in counseling this population. University programs need to provide training in developmentally appropriate group counseling modalities that have been shown to be effective with young adolescents. Training in sandtray and other expressive art modalities would give adolescent counselors additional developmentally appropriate tools that would allow them to use indirect and nonverbal interventions with young adolescent clients.

CONCLUSION

The findings of the current study support the effectiveness of group sandtray therapy in improving self-esteem with young adolescent girls. The findings indicate that girls can improve how they feel about their physical appearance, peer relationships and academic achieve- ment as a result of participating in this group intervention. Group sandtray therapy appears to be an effective approach for addressing self-esteem with young adolescent girls, which is a critical issue for girls this age.

REFERENCES

Armstrong, S. A., & Berg, R. C. (2005). Demonstrating group process using 12 Angry Men. Journal for Specialists in Group Work, 30, 135–144.

Armstrong, S. A., & Henson, R. K. (2004). Statistical and practical significance in the IJPT: A research review from 1993–2003. International Journal of Play Therapy, 13(2), 9–30.

Armstrong, S. A., & Henson, R. K. (2005). Statistical practices of IJPT researchers: A review from 1993–2003. International Journal of Play Therapy, 14(1), 7–26.

134 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

Bailey, D. F., & Bradbury-Bailey, M. E. (2007). Promoting achievement for African American males through group work. Journal for Specialists in Group Work, 32, 83–96.

Birndorf, S., Ryan, S., & Auinger, P. (2005). High self-esteem among adolescents: Longi- tudinal trends, sex differences, and protective factors. Journal of Adolescent Health, 37, 194–201.

Blascovich, J., & Tomaka, J. (1991). Measures of self-esteem. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psycho- logical attitudes: Vol. 1. Measures of social psychological attitudes (pp. 115–160). New York: Academic Press.

Block, J., & Robins, R. W. (1993). A longitudinal study of consistency and change in self-esteem from early adolescence to early adulthood. Child Development, 64, 909–923.

Bolognini, M., Plancherel, B., & Bettschart, W. (1996). Self-esteem and mental health in early adolescence: Development and gender differences. Journal of Adolescence, 19, 233–245.

Bratton, S. C., & Ferebee, K. (1999). The use of structured expressive art activities in group activity therapy with preadolescents. In D. S. Sweeney & L. E. Homeyer (Eds.), The handbook of group play therapy: How to do it how it works whom it’s best for (pp. 192–214). San Francisco, CA: Jossey-Bass.

Brendgen, M., Vitaro, F., & Doyle, A. B. (2002). Same-sex peer relations and romantic relationships during early adolescence: Interactive links to emotional, behavioral, and academic adjustment. Merrill-Palmer Quarterly, 48, 77–103.

Burrows, A., & Cooper, M. (2002). Possible risk factors in the development of eating disorders in overweight pre-adolescent girls. International Journal of Obesity, 26, 1268–1273.

Carlson, C., Uppal, S., & Prosser, E. C. (2000). Ethnic differences in processes contri- uting to the self-esteem of early adolescent girls. Journal of Early Adolescence, 20, 44–67.

Cochran, J. L. (1996). Using play and art therapy to help culturally diverse students overcome barriers to school success. School Counselor, 43, 287–299.

Connor, D. F., Doerfler, L. A., & Toscano, P. F. Jr. (2004). Characteristics of children and adolescents admitted to a residential treatment center. Journal of Child and Family Studies, 13, 497–510.

Constantine, M. G. (2001). Multicultural training, theoretical orientation, empathy, and multicultural case conceptualization ability in counselors. Journal of Mental Health Counseling, 23, 357–374.

Dies, K. G. (2000). Adolescent development and a model of group psychotherapy: Effective leadership in the new millennium. Journal of Child & Adolescent Group Therapy, 10, 97–111.

Draper, K., Ritter, K. B., & Willingham, E. U. (2003). Sand tray group therapy with adolescents. Journal for Specialists in Group Work, 28, 244–260.

DuBois, D. L., Burk-Braxton, C., & Swenson, L. P. (2002). Race and gender influences on adjustment in early adolescence: Investigation of an integrative model. Child Development, 73, 1573–1592.

Flahive, M., & Ray, D. (2007). Effect of group sandtray therapy with preadolescents in a school setting. Journal for Specialists in Group Work, 32, 362–382.

Hampel, P., & Petermann, F. (2005). Age and gender effects on coping in children and adolescents. Journal of Youth and Adolescence, 34, 73–83.

Harter, S. (1985). Manual for the self-perception profile for children. Denver: University of Denver.

Harter, S. (1999). The construction of the self. New York: Guilford.

Shen, Armstrong / GROUP SANDTRAY THERAPY 135

Hedges-Goettl, L., & Tannenbaum, M. (2001). Navigating the social landscape: A ration- ale and method for interpersonal group therapy with pre-adolescent boys. Journal of Child & Adolescent Group Therapy, 11, 135–146.

Holcomb-McCoy, C. C., & Moore-Thomas, C. (2001). Empowering African-American adolescent females. Professional School Counseling, 5, 19–26.

Homeyer, L., & Sweeney, D. (1998). Sandtray therapy: A practical manual. Royal Oak, MI: Self-Esteem Shop.

Hull-Blanks, E. E., Kerr, B. A., & Robinson Kurpius, S. E. (2004). Risk factors of suicidal ideations and attempts in talented, at-risk girls. Suicide and Life-Threatening Behavior, 34, 267–276.

Israel, A. C., & Ivanova, M. Y. (2002). Global and dimensional self-esteem in preadoles- cent and early adolescent children who are overweight: Age and gender differences. International Journal of Eating Disorders, 31, 424–429.

Kalff, D. (1980). Sandplay: A psychotherapeutic approach to the psyche. Cloverdale, CA: Temenos Press.

Keery, H., Boutelle, K., & van den Berg, P. (2005). The impact of appearance-related teasing by family members. Journal of Adolescent Health, 37, 120–127.

Kestly, T. (2001). Group sandplay in elementary schools. In A. Drewes, L. Carey & C. Schaefer (Eds.), School-based play therapy (pp. 329–349). New York: John Wiley & Sons.

Landreth, G. L. (2002). Play therapy: The art of the relationship. New York: Brunner- Routledge.

Leadbeater, B. J., Kuperminc, G. P., & Blatt, S. J. (1999). A multivariate model of gender differences in adolescents’ internalizing and externalizing problems. Developmental Psychology, 35, 1268–1282.

LeCroy, C. W. (2004a). Evaluation of an empowerment program for early adolescent girls. Adolescence, 39, 427–441.

LeCroy, C. W. (2004b). Experimental evaluation of ‘‘Go Grrrls’’ preventive intervention for early adolescent girls. Journal of Primary Prevention, 25, 457–473.

Lomonaco, S., Scheidlinger, S., & Aronson, S. (2000). Five decades of children’s group treatment: An overview. Journal of Child & Adolescent Group Therapy, 10, 77–96.

Martin, M. C., & Gentry, J. W. (1997). Stuck in the model trap: The effects of beautiful models in ads on female pre-adolescents and adolescents. Journal of Advertising, 26(2), 19–33.

Meece, J. L. (1997). Child and adolescent development for educators. New York: McGraw-Hill.

Merwin, R. M., & Ellis, J. B. (2004). Children’s reasons for living, self-esteem, and violence. Archives of Suicide Research, 8, 251–261.

Moran, B. L., & DuBois, D. L. (2002). Relation of social support and self-esteem to problem behavior: Investigation of differing models. Journal of Early Adolescence, 22, 407–435.

Muller, L. E. (2000). A 12-session, European-American-led therapy group for African- American females. Professional School Counseling, 3, 264–269.

Nichols-Goldstein, N. (2001). The essence of effective leadership with adolescent groups: Regression in the service of the ego. Journal of Child & Adolescent Group Therapy, 11, 13–17.

Ohannessian, C. M., Lerner, R. M., & Lerner, J. V. (1999). Does self-competence predict gender differences in adolescent depression and anxiety? Journal of Adolescence, 22, 397–411.

Polce-Lynch, M., Myers, B. J., & Kilmartin, C. T. (1998). Gender and age patterns in emotional expression, body image, and self-esteem: A qualitative analysis. Sex Roles, 38, 1025–1048.

136 THE JOURNAL FOR SPECIALISTS IN GROUP WORK / June 2008

Pollock, K. M., & Kymissis, P. (2001). The future of adolescent group therapy. Journal of Child & Adolescent Group Therapy, 11, 3–11.

Rice, K. G., & Meyer, A. L. (1994). Preventing depression among young adolescents: Preliminary process results of a psycho-educational intervention program. Journal of Therapy & Development, 73, 145–152.

Rosenberg, M. (1985). Self-concept and psychological well-being in adolescence. In R. L. Leahy (Ed.), The development of the self (pp. 205–246). New York: Academic Press.

Russell, K. C., & Phillips-Miller, D. (2002). Perspectives on the wilderness therapy process and its relation to outcome. Child & Youth Care Forum, 31, 415–437.

Schumann, B. C., Striegel-Moore, R. H., McMahon, R. P., Waclawiw, M. A., Morrison, J. A., & Schreiber, G. B. (1999). Psychometric properties of the Self-Perception Profile for Children in a bi-racial cohort of adolescent girls: The NHLBI growth and health study. Journal of Personality Assessment, 73, 260–275.

Shavelson, R. (1996). Statistical reasoning for the behavioral sciences (3rd ed.). Needham Heights, MA: Allyn & Bacon.

Slavson, S. R. (1943). An introduction to group therapy. New York: International Universities Press.

Steen, S., Bauman, S., & Smith, J. (2007). Professional school counselors and the practice of group work. Professional School Counseling, 11, 72–80.

Sue, D. W. (2001). Multidimensional facets of cultural competence. The Counseling Psychologist, 29, 790–821.

Sue, D. W., & Sue, D. (2004). Counseling the culturally different: Theory and practice. New York: Wiley & Sons, Inc.

Tuckman, B. W., & Jensen, M. A. C. (1977). Stages of small group development revisited. Group and Organizational Studies, 2, 419–427.

Tyndall-Lind, A., & Landreth, G. L. (2000). Intensive short-term group play therapy. In G. L. Landreth (Ed.), Innovations in play therapy: Issues, process, and special populations (pp. 203–215). New York: Brunner-Routledge.

Yalom, I. D., & Leszcz, M. (2005). The theory and practice of group psychotherapy (5th ed.). New York: Basic Books.

Young, J. L. (1994). Sapphires-in-Transition: Enhancing personal development among Black female adolescents. Journal of Multicultural Counseling and Development, 22, 86–95.

Zinck, K., & Littrell, J. M. (2000). Action research shows group therapy effective with at-risk adolescent girls. Professional School Counseling, 4, 50–59.

Shen, Armstrong / GROUP SANDTRAY THERAPY 137