Efficacy of a Group Treatment for Children with Significant Social Skills Deficits

Molly M. Parsons
Stacey S. Park
Lee A. Rosén*

Colorado State University

Abstract

This study examined the effectiveness of an eight session, outpatient social skills group therapy treatment for children with significant social skills deficits.  The curriculum was designed to teach nine specific social skills including: Eye Contact, Personal Space, Self-Emotions, Other-Emotions, Pedantic Speech, Greetings, Conversations, Friendliness, and Play Skills.  Seven elementary aged children with an Autism Spectrum Disorder and/or Attention Deficit Hyperactivity Disorder participated.  Parents were asked to fill out the Child Social Skills Inventory, which measures these nine social skills domains at two time points: the first day of social skills group (SSG) and the last day of SSG.  Results indicate that the treatment was partially effective in improving social skills - pre to post changes in the domain of Self-Emotion were significant.  In addition, the domains of Play Skills and Other-Emotion showed improvement near statistical significance.  Improvement in these important social skill domains helps substantiate the efficacy of social skills group therapy treatment for children with severe social skill deficits.  

Efficacy of a Group Treatment for Children with Significant Social Skills Deficits

There are many psychological and developmental challenges that affect today's youth.  However, two of the most prevalent include Attention Deficit Hyperactivity Disorder (ADHD) and Autism Spectrum Disorders (ASD).   In the United States, approximately 1 in 68 children are affected by an ASD, and 1 in 12 children have ADHD (Center for Disease Control, 2014).  Furthermore, these disorders are commonly comorbid, with estimates of as many as 30% children diagnosed with an ASD also having ADHD (Center for Disease Control, 2013). Researchers have started to examine children who present with symptoms of both disorders (Cervantes et al., 2013).  Though the developmental course of ASD and ADHD, the symptoms for each, and associated causes and factors and different, a primary problem in both children with ASD or ADHD is significant social skills deficits (Cervantes et al., 2013) - causing difficulty in both home and school settings and lifelong challenges.    

A social skills deficit can be defined as significant impairment in social interactions, an inability to understand emotions, and failure to develop age-appropriate relationships (Cotungo, 2009).  Children with ASD develop social skills deficits as a result of atypical interactions, unusual interests that are not shared by peers, or difficulty understanding social cues (Barry et al., 2003).  Research has shown that children with ASD show deficits in adaptive or appropriate social communication behaviors, for example, the ability to share laughter with others, or smile to engage others (Cervantes et al., 2013).  Subsequently, children with ASD typically struggle to make friends, or maintain friendships.  However, children with ADHD develop socialization deficits due to marked impulsivity or hyperactivity in which they may interrupt a conversation or intrude on peer relationships (see Nijmeijer et al., 2008).  Cervantes and colleagues (2013) demonstrated that children with ADHD exhibit more hostile or inappropriate assertiveness in their interactions.  As a result of these particular social skills deficits, children with ADHD are more likely to be rejected by their peers, and are generally less socially preferred (Nijmijer et al., 2008).  Though children with ASD and children with ADHD demonstrate different types of social skills deficits, individuals from either diagnostic category are more likely to experience poor peer relationships and social functioning.  Research consistently shows that social skills training and intervention is crucial for these specific groups.   

One intervention aimed at remediating these deficits in children is social skills group therapy.  Social skills groups (SSGs) allow children to learn and to practice social skills in an outpatient setting while receiving guidance and feedback from group leaders and other peers in the group.  Children are exposed to adult and peer models of behaviors, and promptly reinforced to encourage motivation (Matson, Matson, & Rivet, 2007).  Social skills interventions target populations with ASD by providing a social script to improve the understanding of social rules (Matson, et al., 2007).  For children with ADHD, interventions should target social skills such as listening, relaxation, assertiveness, and accepting consequences (Cervantes, et al., 2013).  Typically, SSGs include four to 10 children similar in developmental age and one or more group leaders or facilitators (e.g., Barry, Klinger & Lee, 2003; Cotungo, 2009; Gol, 2005).  One hour is the standard amount of time for sessions and one to two times per week is the average frequency.  Group sessions focus on improving social behavior such as conversations, play skills, and emotions.  These groups can run from four weeks to nine months. 

Despite the number of SSGs for ASD and other disorders, there is little research examining the efficacy of SSGs as a treatment option.  Most studies focusing on the impact of SSGs on children with ASD see some improvement in social competency, but it is often limited and is rarely observed as generalized into the school or home setting (Barry, Klinger & Lee, 2003; Epp, 2009).  For example, in the Barry, et al. (2003) study measuring improvement within a group of four elementary aged students with high-functioning Autism (HFA), significant progress were demonstrated in greeting and play skills.  However, there was no clear improvement in conversation skills.  Also, the only skills showing generalization to naturalistic settings in this study was greeting skills.  Another study including sixty-six child participants with ASD conducted by Epp (2008) also demonstrated success in attainment of social skills.  Of the seven domains in this study, assertion, internalizing and hyperactivity all showed significant improvement.  A longitudinal study that found statistically significant improvement in all domains measured was conducted by Cotungo (2009) and focused on children with HFA.  The domains measured included anxiety management, joint attention and flexibility/transitions.  Both parents and teachers assessed these improvements over the course of nine months in a pretest-posttest design, and the results showed significant improvement from both perspectives.   

Research on SSGs for children with ADHD show limited benefits from SSGs and those benefits are rarely generalized into naturalistic settings.  Frankel, Myatt and Cantwell (1995) failed to find significant improvement in problem behavior following a SSG intervention for ADHD boys with socialization issues.  Although parent-rated understanding and demonstration of social skills improved significantly, aggression levels and participant likability showed no improvement.  Such results question the effectiveness of SSGs for the ADHD population.  However, a study conducted by Pfiffner and McBurnett (1997) showed more success than their predecessors.  In this study, SSG related to statistically and clinically significant improvements in social interactions and at-home behavior problems demonstrating that the possibility for generalization exists.  Pfiffner and McBurnett (1997) theorize that the benefits of large magnitude measured in this study may have been augmented by the realistic role-plays in the curriculum that included family and school examples as well as their behavior modification plan.  This plan utilized potent positive reinforcement and "prudent" negative consequences (Pfiffner & McBurnett, 1997). 

Current Study

The present research study examines the efficacy of a social skills group treatment for children with atypical development.  In group, children were given didactic instruction, practice opportunities where trained aides and leaders provided corrective feedback, and homework assignments to be completed with a parent (to help generalize the skills outside of group).  Due to the exploratory nature of this study, a control group was not included.  We hypothesized that children (ages five to 10) who have been diagnosed with ADHD and/or ASD would show improvement in social skills after completing the eight-session social skills group therapy program.  As an exploratory study for future research to build on, the present research holds important applications for treatment of social skills in atypically developing children.   

Method

Participants

Children with ASD or ADHD and their families were primarily recruited from a department clinic mailing list.  Following an application, participants were screened to determine if they would be an appropriate fit for a social skills group.  Severely non-verbal children and children who are extremely disruptive were excluded from attending the group on the basis that they were less likely to receive a therapeutic benefit from their participation.  In order to qualify for this study, parents needed to complete both the Child Social Skills Inventory (CSSI) administered on the first day of group and the CSSI administered on the last day of group. Eighteen children participated in three separate social skills groups over the course of eight months in 2013.  Of the 18 children who participated in group, seven qualified for inclusion in this study because their parents completed the questionnaires on both the first and last day of group.   The mean age of the participants was 7.86 years (SD = 2.12); the range was five to 10 years old.  There was one female in this study.  The demographic questionnaire revealed that six of the participants identified as white/Caucasian and one as Latino or Hispanic.  Parents were asked to report each child’s psychological diagnoses: one child was diagnosed with an ASD, two had Pervasive Developmental Disorder – Not Otherwise Specified (PDD-NOS), and four children in the group were reported to have ADHD as their primary diagnoses. The most common co-morbid diagnoses were Anxiety (four participants), Oppositional Defiant Disorder (three participants), Mood Disorders (two participants) and Learning Disorders (two participants). Only one participant did not list any co-morbid disorders.  

Measures

Measures used for this study include a demographic survey and the Child Social Skills Inventory (CSSI).  The demographic survey asked for information regarding the age of the child, grade in school, sex, ethnicity, and primary diagnoses.  The CSSI, which was created by the researchers for the purpose of this study, was also used (See Appendix A).  The CSSI measures a parent’s rating of their child’s social skills in nine domains: Eye Contact, Personal Space, Self-Emotions, Other-Emotions, Pedantic Speech, Greetings, Conversations, Friendliness, and Play Skills.  Domains were selected and items were written by a team consisting of a licensed clinical psychologist and two advanced doctoral students.  Each individual specialized in working with children with ASD and ADHD and both doctoral students have led over 12 SSGs with children.  Domains and items were based on the curriculum of the group; each session targets specific skills and follows the format recommended by Painter (2008).    

Containing 46 items, the CSSI used a five-point (from never to always) Likert-item scale which prompted parents to rate how often in the last week they observed their child engaging in certain behaviors in the nine domain categories.  Each domain contained five items (with the exception of Friendliness which included six).  Sample items include: Looks others in the eye when spoken to for Eye Contact, Sits or stands in others personal space for Personal Space, Identifies their own emotions for Self-Emotions, and Greets familiar individuals by name for Greetings.  See Appendix A for a copy of the inventory.  Precisely half (23 out of 46) of these items are reverse-scored.  Domain scores can range from 0 to 20 (with the exception of friendliness which can range from 0 to 24).  The total score on the CSSI scale for each participant can range from 0 to 184, with higher scores reflecting higher frequency in observed social skills.  Reliability for the total CSSI was good at Time 1 (Cronbach’s α = .82) and excellent at Time 2 (α = 0.92; George & Mallery, 2003).  All domains individually showed acceptable reliability with a Cronbach's alpha above 0.70 with the exception of Friendliness, Other-Emotions, and Play Skills in Time 1 and Pedantic Speech in Time 2.  See Table 3 and Table 4 for Cronbach's alpha for individual domains and total scores.  The scale had relatively good internal consistency for the sample used in this study.   

Table 3 Correlations for the Total and Domain scores of the CSSI at Time 1 (N = 7)

Note.   * indicates p < .05, ** indicates p < .001.   Internal consistency values are in italics. 

Table 4 Correlations for the Total and Domain scores of the CSSI at Time 2 (N = 7)

Note.   * indicates p < .05, ** indicates p < .001.   Internal consistency values are in italics. 

Procedure

Participants were provided with an informed consent form that detailed the study, any potential risks and benefits from participating in the study, and an assurance of anonymity and confidentiality.  Each family filled out a demographic questionnaire requesting information including diagnoses and demographics.  Parents were asked to complete the CSSI at the first session and at the last session of group.  Each participant was assigned a random number, so as to match participant data from Time 1 and Time 2 and keep individual information confidential and anonymous.   Upon completion of the study parents were provided with a debriefing form, and thanked for their participation.  All aspects of the study were reviewed and approved by the Institutional Review Board for human research participants.     

Treatment consisted of eight one-hour sessions over the course of eight weeks for spring participants and four weeks (with two sessions a week) for summer participants. The curriculum written by Painter (2006) targeted specific domains of social functioning.  Each session was composed of the same basic structure, including: free play with opportunities for feedback, a review of the previous week’s topic and the homework, didactic instruction of a new topic, practice opportunities for group members to practice the skills learned, and a snack/free time.  During free time, the first 10 minutes of the session, games and puzzles were provided and placed around the room.  Practice opportunities consisted of pairing up with a partner and being provided with feedback from a leader or volunteer, role-plays by the leaders, group games, and having the group members act out certain skills.  Didactic instruction and activities occupied approximately 40 minutes of the session.  The 10 minutes of group consisted of social snack time where the games were brought back out for participants to play and eat snack together.  During this time one or both of the group leaders would meet with the parents to summarize the day's lesson, provide techniques for using the skills at home, and hand out a homework assignment to reinforce generalization of skills.  Each session pinpointed a different set of social skills, and topics included making introductions, conversations, friendship skills, emotions, anger and anxiety, peer conflict, sportsmanship, and manners.  Didactic topics, group activities, and social skills are detailed in Table 1.

Table 1 Group Curriculum by Session and Targeted Skills

Results

Scores were computed for each CSSI domain as well as a total score for all nine domains of the CSSI for Time 1 (at the first group) and Time 2 (at the last group).  Less than 0.05% of the data was missing, and values were replaced using mean substitution.  The variables appeared to be normally distributed.  Thirteen of 18 domains showed moderate Platykurtic distribution, or flatter peak than normal.  However, this is somewhat mediated by the very small sample size.  A positive correlation was found between Time 1 and Time 2 CSSI scores (r(7) = 0.75, p < .05).  Results indicate that at Time 1, Total CSSI scores were positively related to Friendliness (r(7) = 0.90, p < .05), Play Skills (r(7) = 0.88, p < .05), and Greetings (r(7) = 0.75, p < .05).  At Time 2 the same domains were positively correlated to Total CSSI scores in addition to Other-Emotions (r(7) = 0.80, p < .05) and Personal Space (r(7) = 0.89, p < .05).  Means and standard deviations for domain scores at Time 1 and Time 2 are reported in Table 5.   

Table 5 Time 1, Time 2, and Time 1–Time 2 Gains for CSSI Total and Domain Scores (N = 7)

A series of repeated measures t-tests were conducted to examine differences in CSSI scores from first to last group treatment session.  There was no significant difference between Time 1 and Time 2 scores for the CSSI total score (t(6) = -1.07, p = 0.33).  However, the change from Time 1 to Time 2 score for the domain of Self-Emotion was found to be significant (t(6) = -2.97, p < 0.05).  In addition, Other-Emotion and Play Skills were close to significant (t(6) = -2.32, p = 0.06t(6) = -2.25, p = 0.07, respectively).   

Analysis of individual participant growth in social skills revealed that only three of the seven participants showed an increase in overall social skills between Time 1 and Time 2 as measured by total scores on the CSSI.  However, all seven participants showed improvement in at least one domain.  Self-Emotions, Other-Emotions and Play Skills showed improvement for the most participants, each only having one participant score lower at Time 2 than Time 1.  The domains that the fewest participants improved on were Personal Space (one participant improved), Conversation Skills, and Eye Contact (See Table 2).   

Table 2 Percentage of Participants that Improved between Time 1 and Time 2 by Domain and Diagnoses

Note.   * indicates p < .05, ** indicates p < .001.

Mean gains in total CSSI scores were also calculated for each diagnostic group.  The three children who were diagnosed with ASD had an average Time 1 CSSI score of 110.67 and a Time 2 score of 114.33.  The four children diagnosed with ADHD had an average score of 91.75 at Time 1 and 97.00 at Time 2 as demonstrated in Figure 1.  Average scores for the domains were also calculated.  Both the ASD group and ADHD group showed increases in scores in six domains from Time 1 to Time 2.  Interestingly, there was little overlap in domains of improvement and no overlap in domains lacking improvement.  See Table 1 for differences in improvement based on domain.  The domains in which ASD children had a lower score at Time 2 than Time 1 (representing a lack of improvement) were Friendliness, Pedantic Speech, and Play Skills.  For the children in the ADHD group, however, Eye Contact, Personal Space and Greetings were the domains that lacked improvement.   

Figure 1.  A figure that looks at differences in CSSI scores by diagnosis.  

Discussion

 The goal of this study was to better understand the efficacy of a social skills group treatment for children with atypical development.  Given the increasing prevalence of ADHD and ASD (CDC, 2013), and the harmful effect of the disorders on socialization (Cervantes et al., 2013), it will be important to develop brief, targeted interventions that can be used by therapists, as well as, school educators and parents.   The result of this study provides preliminary support for the hypothesis that social skills group therapy can improve the social skills of children with ADHD or ASD.  The treatment was successful in improving recognition and understanding of self-emotions in all seven of the participants.  Play skills and other-emotions showed improvement trends, but were just short of the threshold of significance. 

Interestingly, improvements were seen among children with ADHD and children with ASD across different domains.  Research has shown that children with either disorder struggles with social skills for different reasons (Cervantes et al., 2013; Nijmeijer et al., 2008).  Within this study, even with the limited sample size, at least two of the three children with ASD improved in eye contact, identifying their own emotions, identifying others’ emotions, friendliness towards others, play skills with peers, offering greetings to others, and holding reciprocal back-and-forth conversations.  Research suggests that the primary social deficit for children with ASD is in social-emotional reciprocity – skills such as using nonverbal forms of communication like eye contact and facial expressions, as well as, reciprocal conversations, initiating interactions, and theory of mind (Cervantes et al., 2013).  This preliminary study suggests that didactic instruction, opportunities to practice, and prompt and delayed reinforcement (via homework) can improve children’s skills in these specific areas. 

Similarly, at least two of the four children with ADHD improved on eye contact, identifying one’s own emotions, identifying others’ emotions, friendliness (including not teasing others), play skills like sharing or turn-taking, greeting others appropriately, speaking to others appropriately, and not monopolizing conversations.  Given the research suggesting that children with ADHD struggle with social relationships because of impulsive actions or inattentiveness that could lead to peer rejection (Nijmeijer et al., 2008), the study here demonstrates that group therapy could be one effective modality for teaching children with ADHD to modify their peer skills so as to make friends and maintain relationships.  Also notable, children with ASD or ADHD could help each other in learning different aspects of a skill.  For example, while one child with ASD learned to initiate and maintain conversations, another child with ADHD learned not to monopolize the conversation.  Both children learned a new conversation skill while simultaneously helping the other.  School educators may have to work with groups of children with atypical development with diverse ranges of disorders.  This study offers some support for pairing children with ASD and children with ADHD together when learning social skills.

The primary limitations of this study were small sample size and lack of a control group.  With only seven children in the study and no control group, interpretation of the results is limited.  A larger sample size may have resulted in higher statistical power to detect smaller effects, like the trends seen in Play Skills and Other-Emotion.  Alternative methods to improve participant data collection and create a control group should be considered in the future. For example, creating incentives for completing both sets of CSSI measures may assist in retaining a larger number of participants that meet the inclusion criteria and therefore a higher sample size. Additionally, replicating the study with a control group or adding an additional rater (i.e. a second parent or teacher) whom would also complete the CSSI measures would validate the findings in the data and help eliminate bias. However, despite its limitations this pilot study provided enough evidence of improvement to justify proceeding with steps mentioned above.

Furthermore, although not reflected in the group summary data, every child in the study showed improvement in at least one domain between the first day of group and the last day of group.  Also, at least one child showed improvement in every domain measured by the CSSI.  Finding significance in any domains with such a limited sample size is encouraging - it appears that this treatment approach can indeed improve social skills in children with significant deficits. 

References

Appendix A

The Childhood Social Skills Inventory (CSSI)

In the last week, rate how often you’ve observed your child doing the following behaviors.