J Autism Dev Disord

DOI 10.1007/s10803-013-1900-1

ORIGINAL PAPER

A Replication and Extension of the PEERS Intervention:

Examining Effects on Social Skills and Social Anxiety
in Adolescents with Autism Spectrum Disorders
Kirsten A. Schohl • Amy V. Van Hecke •
Audrey Meyer Carson • Bridget Dolan •
Jeffrey Karst • Sheryl Stevens

Ó Springer Science+Business Media New York 2013

Abstract This study aimed to evaluate the Program for the evidence that those with Asperger’s Syndrome (AS) or
Education and Enrichment of Relational Skills (PEERS: High Functioning Autism (HFA), terms which are often
Laugeson et al. in J Autism Dev Disord 39(4):596–606, used interchangeably, may be the fastest growing segment
2009). PEERS focuses on improving friendship quality and of the autism population (Rao et al. 2008).
social skills among adolescents with higher-functioning ASD symptoms are pervasive and vary greatly in
ASD. 58 participants aged 11–16 years-old were randomly severity. In general, those with ASD have numerous
assigned to either an immediate treatment or waitlist com- domains affected, including social and behavioral func-
parison group. Results revealed, in comparison to the waitlist tioning and language development. They are also distin-
group, that the experimental treatment group significantly guished by the presence of a variety of circumscribed
improved their knowledge of PEERS concepts and friend- interests and stereotyped, repetitive behaviors. While those
ship skills, increased in their amount of get-togethers, and with AS/HFA usually function within the typical range
decreased in their levels of social anxiety, core autistic with regard to language and intelligence, they display
symptoms, and problem behaviors from pre-to post-PEERS. impairments in social skills, which is the hallmark feature
This study provides the first independent replication and of AS/HFA (Mitchell et al. 2010).
extension of the empirically-supported PEERS social skills These marked social deficits are problematic, especially
intervention for adolescents with ASD. during adolescence, when the demands of peer relation-
ships and social network affiliations become heightened
Keywords Autism
Asperger’s disorder
ASD
(Mitchell et al. 2010). In addition to these challenges of
Adolescence
PEERS
Intervention
Social skills
adolescence, those with AS/HFA are typically self-con-
Social anxiety
Friendships scious of their differences in social functioning, and indi-
cate that they experience stronger feelings of loneliness and
poorer quality friendships than their typically developing
Introduction peers (Bauminger and Kasari 2000). As a result, a signif-
icant number of adolescents with AS/HFA are at an
The number of youth diagnosed with Autism Spectrum increased risk for a variety of secondary psychopathology,
Disorder (ASD) has increased dramatically over the past such as depression and anxiety, in addition to other nega-
decade and currently affects approximately 1 in 88 children tive outcomes both in adolescence and adulthood, includ-
in the U.S. (Centers for Disease Control and Prevention ing isolation, rejection, teasing, bullying, low self-esteem,
2012). It has also been suggested by empirical and clinical school dropout, and unemployment (Mitchell et al. 2010).
Unfortunately, there have been very few interventions
developed that have focused on improving social adaptation
K. A. Schohl (&)
A. V. Van Hecke
A. M. Carson
among adolescents with AS/HFA. In response to this need,
B. Dolan
J. Karst
S. Stevens
the Program for the Education and Enrichment of Relational
Department of Psychology, Marquette University,
PO Box 1881, Milwaukee, WI 53201-1881, USA Skills (PEERS) intervention was recently developed, in
e-mail: kirsten.schohl@mu.edu order to teach adolescents with AS/HFA the skills necessary

123
 J Autism Dev Disord

to make and keep friends (Laugeson et al. 2009). Although understanding irony, jokes, lies, deception, or bullying
the intervention has demonstrated positive gains in skills and (Grynszpan et al. 2011). Individuals with AS/HFA also
social contacts for adolescents with ASD who complete the experience difficulty with the social rules of conversation,
program (Laugeson et al. 2009, 2012), PEERS has not yet such as taking turns, providing enough information to be
been replicated outside of its site of development. clear without being verbose, and selecting information that is
Intervention replication is critical, particularly replication relevant to the topic at hand (Krasny et al. 2003).
by an independent investigatory team. The requirement of These initial core deficits displayed in social situations
replication helps to protect from drawing erroneous con- can be exacerbated during adolescence, which is a time when
clusions based on one aberrant finding. Replication by an identification with a peer group is common. Further, ado-
independent team of investigators also provides some pro- lescence can be a distressing phase of life for many adoles-
tection against investigator bias or reliance on findings that cents with AS/HFA due to their difficulty engaging socially
prove unique to a particular setting, specific characteristics with peers. Because adolescents with AS/HFA typically
of local samples, health care settings, or group of therapists. have normal to high intelligence and thus greater capacity for
Further, replication of randomized clinical trial interven- insight, they are often painfully aware of the difficulties they
tions, at different sites, with different samples, increases the experience when interacting with peers (Grynszpan et al.
validity and generalizability of data as compared to the data 2011). In a research study, youth with AS/HFA rated them-
gathered at a single site. Replication of intervention effects in selves on average more than one standard deviation below
different settings is necessary for an intervention model to be the mean of typically developing children on social skills,
considered as well established (Chambless et al. 1998). such as joining groups, demonstrating social competence,
Moreover, replication of interventions promotes clinical and developing close friendships (Rao et al. 2008). These
utility and helps facilitate the dissemination of evidence- findings suggest that adolescents with AS/HFA are, in fact,
based interventions (Drotar 2006). Only when a treatment cognizant of their social inabilities.
has been found efficacious in at least two studies by inde- In addition to the increased awareness adolescents with
pendent research teams do some researchers consider its AS/HFA may possess, adolescence is a time when ‘‘fitting
efficacy to have been established and label it an efficacious in’’ with one’s classmates is of prime importance. Since the
treatment (Chambless et al. 1998). Although replication is majority of today’s youth with AS/HFA are placed in
crucial, it has not been widely practiced in relation to social regular education classrooms as opposed to special needs
skills for individuals with ASD. According to a recent classrooms (Sofronoff et al. 2010), presenting with social
Cochrane Review, which investigated current social skills incompetence may lead to the opposite of ‘‘fitting in.’’
interventions for people ages 6–21 years with ASD, there Despite the finding that regular education placement leads
were no replicated findings reported (Reichow et al. 2012). to increases in the complexity of interactions and decreases
This paper will first review core deficits in adolescents in nonsocial activity, adolescents with AS/HFA often
with ASD, social anxiety, associated challenges of ado- report feeling lonelier and having poorer quality friend-
lescence, and validated interventions for this develop- ships than their typically developing peers (Bauminger and
mental period. Then, the current study, which examines Kasari 2000). It has been suggested that having one or two
whether social skills and social anxiety in adolescents with best friends is of great importance to later adjustment.
AS/HFA change due to a Randomized Controlled Trial Specifically, having friends buffers the impact of stressful
(RCT) and replication of the PEERS intervention (Lauge- life events, correlates positively with self-esteem, and
son and Frankel 2010), is presented. correlates negatively with anxious and depressive symp-
toms (Buhrmester 1990). Unfortunately, these benefits are
not possible for many adolescents with AS/HFA, as it has
Core Deficits in Adolescents with AS/HFA been found that nearly 50 % of adolescents with ASD do
and Associated Challenges not have a friend (Howlin 2000).
Unfortunately, the idea that those with AS/HFA will
Adolescents with AS/HFA have significant difficulties with simply ‘‘outgrow’’ their social skill deficits after adoles-
their social behavior. These deficits might include inade- cence is not supported by research. Instead, these diffi-
quate use of eye contact, problems initiating social interac- culties persist into adulthood, where they continue to
tions, and difficulty interpreting both verbal and nonverbal negatively impact social and occupational functioning. It
social cues such as tone of voice, facial expression, gesture, has been found that adults with AS/HFA are more likely
gaze, and posture (Weiss and Harris 2001). Those with AS/ than the general population to be unemployed or under-
HFA often have problems with pragmatics, which refers to employed, as well as less likely to have satisfying social
the ability to use language to communicate effectively in relationships and community connections (Rao et al. 2008).
social situations. For example, they display problems in As this research demonstrates, understanding and being

123
J Autism Dev Disord

connected to the social world is essential for those with AS/ Social Skills Interventions for Adolescents
HFA to function properly and gain autonomy. with AS/HFA

Interventions that aim to improve social skills are essential

Social Anxiety in Adolescents with AS/HFA for individuals with AS/HFA. Given the pervasive impact
and long-term nature of social skill deficits in AS/HFA,
In addition to displaying social skill deficits and dealing with social skills training programs aimed at adolescents may
the heightened social demands of adolescence, adolescents prevent or at least lessen subsequent social dysfunction
with AS/HFA may also present with anxiety, especially (Goldstein and McGinnis 2000). Although there have been
during social situations. Anxiety-related concerns are among a minimal number of interventions developed and imple-
the most common presenting problems for school-age chil- mented for adolescents with AS/HFA, there are a few
dren and adolescents with ASD, as 11–84 % experience programs that prove important and should be mentioned.
some degree of impairing anxiety (White et al. 2010). Fur- Ozonoff and Miller (1995) developed a 14-week social
thermore, research indicates that anxiety may be somewhat skills intervention, which focused on teaching adolescents
universally comorbid with ASD. There is evidence to sug- basic interactional and conversational skills and how to infer
gest that anxiety difficulties occur more frequently in ASD the mental states of others, otherwise called Theory of Mind
populations, as compared to children with severe mental (ToM). The participants included five high-functioning
retardation, epilepsy, conduct disorder, and children who adolescent boys with ASD. Post-treatment ratings completed
have a language disorder (Chalfant et al. 2007). by adolescents’ parents and teachers, however, suggested
Regarding social anxiety in particular, those with AS/ that there were no improvements of adolescents’ social
HFA report significantly more social anxiety symptoms competence skills or adolescents’ generalization of skills to
than their typically developing peers, and these symptoms other settings. Tse et al. (2007) conducted a social skills
increase as they get older, in contrast to the decreasing intervention for 46, 13–18-year-old adolescents for
pattern of anxiety symptoms often displayed in typically 12 weekly sessions. Many of the exercises used to teach new
developing adolescents (Sebastian et al. 2009). Moreover, skills were adapted from the book, Skillstreaming the Ado-
one research study found that from a sample of 41 high- lescent (Goldstein and McGinnis 2000). There was no con-
functioning adolescents with AS/HFA, 49 % of the sample trol group utilized, although they did find that parent report
scored above the clinically significant level of social anx- measures showed gains in adolescent social competence and
iety on a self-report measure (Bellini 2004). decreases in problem behaviors following the intervention.
There are different theories as to why social anxiety is so Another study (Mitchell et al. 2010) focused on the gener-
common among the AS/HFA population. Most of the theo- alization effects of a group social skills training program
ries, however, are encompassed in Bellini’s (2006) devel- with parent training for three adolescents with AS/HFA. The
opmental pathway to social anxiety. According to Bellini, social skills curriculum was adapted from ‘‘Navigating the
there is a feedback loop between physiological arousal, Social World’’ (McAfee 2002) and included topics such as
social anxiety, and social interaction. The pathway begins privacy circles, offering and asking for help, giving and
with the notion that individuals with AS/HFA present with a receiving compliments, resolving conflicts, and basic rules
temperament that is marked by a high degree of physiolog- for initiating conversations. This study was limited, as it had
ical arousal. This physiological arousal may make it more a very small sample size, however, the utilization of parent
likely that the individual will become overwhelmed by sessions and focus on generalization of skills are two vital
interactions with others and avoid later social interactions. components that were incorporated (Mitchell et al. 2010).
This social withdrawal then limits the opportunity for the The majority of programs reviewed here reported null
individual to develop and practice effective social skills by findings for generalization and flexible use of the skills in
reducing interactions with peers. The impairment in social the naturalistic setting (White et al. 2007). Fortunately, one
skill functioning then significantly increases the chances for new social skills treatment directly addresses the problems
negative peer interactions and social failure. To complete the with generalization of social skills into naturalistic settings.
pathway, the presence of physiological hyperarousal makes
it more likely that the individual will be adversely condi-
tioned by these negative social experiences, thus leading to PEERS
increased social anxiety. To intensify the problem, the
presence of social anxiety may lead to further social with- The Program for the Education and Enrichment of Relational
drawal, thus beginning the cycle again (Bellini 2006). Thus, Skills (PEERS: Laugeson et al. 2009) incorporates and
it is critical that interventions be implemented in order to builds upon many of the elements integral for social skills
break this cycle of social anxiety. teaching success. PEERS content, as well as the lesson

123
 J Autism Dev Disord

format, was adapted from Children’s Friendship Training Table 1 PEERS sessions and associated content
(CFT), an evidence-based parent-assisted social skills cur- Session Didactic
riculum (Frankel and Myatt 2003). The PEERS intervention
modified the curriculum and methods of instruction, and 1 Introduction and conversational skills I: trading information
added new modules, in order to be more applicable for 2 Conversational skills II: Two-way conversations
adolescents with AS/HFA (Laugeson et al. 2009). 3 Conversational skills III: Electronic communication
The most important aspects of the PEERS intervention 4 Choosing appropriate friends
are that it is empirically supported, is based on a large 5 Appropriate use of humor
sample (compared to prior studies), and is a manualized 6 Peer entry I: entering a conversation
treatment, which promotes replication. There are three 7 Peer entry II: exiting a conversation
other critical features of this intervention that should be 8 Get-togethers
mentioned, as each adds to the distinctiveness of the 9 Good sportsmanship
PEERS program. 10 Rejection I: teasing and embarrassing feedback
First, teaching of social skills is conducted in a small 11 Rejection II: bullying and bad reputations
group format, as this allows for a more personal experience 12 Handling disagreements
for the adolescents. PEERS also utilizes many evidence- 13 Rumors and gossip
based strategies for teaching social skills to adolescents 14 Graduation and termination
with AS/HFA, which include brief didactic instruction,
role-playing, modeling, behavioral rehearsal, coaching
with performance feedback, and weekly socialization
assignments with consistent homework review (Gresham improved their knowledge of social skills, increased fre-
et al. 2001; Laugeson et al. 2009). quency of hosted get-togethers, and improved overall social
Second, PEERS allows the parents of the adolescent skills as reported by parents. Moreover, in two long term
participants to play an integral part in the treatment process, follow-up studies of the PEERS participants, researchers
as parents are required to engage in separate, concomitant found that the improvements made from baseline to post-
sessions. Many previous programs have not incorporated intervention were maintained at 14-weeks post-intervention
parents into the treatment process. Research, however, sug- and between 1 to 5 years after treatment (respectively,
gests that parents can have a profound impact on their child’s Laugeson et al. 2012; Mandelberg et al. 2011). Although
friendships (Frankel and Myatt 2003). This may be through PEERS has shown evidence of success in both the short and
direct instruction, modeling appropriate social behavior, and long term (Laugeson et al. 2009, 2012; Mandelberg et al.
supervision. By supporting their child’s development of an 2011), it has not been replicated outside of its site of
appropriate peer network, learning to act as social coaches, development.
and encouraging them to engage in social situations despite
their struggles, parents can be critical components of their
adolescents’ social development and retention of newly Aims of the Current Study
learned skills once the program has ended (Frankel and
Myatt 2003; Laugeson et al. 2009). Thus, the current study was an independent replication and
Third, PEERS focuses on teaching rules of social etiquette extension of the PEERS intervention in order to evaluate the
through the identification of common social situations using effectiveness of the program for improving social skills and
accompanying concrete rules and steps of appropriate social social anxiety. This PEERS extension was distinctive from
etiquette. This style of learning complements those with AS/ the first PEERS trial (Laugeson et al. 2009) in several ways.
HFA, as they thrive on structure and concrete presentation of First, it addressed and resolved one of the shortcomings of
information (Carnahan et al. 2009). Skills covered in PEERS the first implementation of PEERS, as it used more ‘‘gold-
include conversational skills, peer entry and exiting skills, standard’’ diagnostic screening. Second, another shortcom-
expanding and developing friendship networks, how to ing of the first study was the low teacher measure return rate,
handle teasing, bullying, and arguments with peers, prac- which was improved upon in this study. Third, this study was
ticing good sportsmanship, changing bad reputations, and conducted within a medium-sized Midwestern city, poten-
good host behavior during get-togethers (Laugeson et al. tially resulting in a different demographic than the Los
2009) (see Table 1). Angeles, California, location where the first PEERS study
The PEERS program was empirically supported with 33 was conducted, and which provided an opportunity for
adolescents, ages 13–17 years with AS/HFA (Laugeson independent replication. Lastly, this study investigated the
et al. 2009). Results revealed that in comparison with the effect PEERS may have on social anxiety with the addition of
waitlist control group, the treatment group significantly an adolescent self-report measure. These questions were

123
J Autism Dev Disord

addressed utilizing a randomized controlled trial (RCT) Table 2 Means and SD for demographic variables for experimental
design. treatment and waitlist control groups
First, it was hypothesized that adolescents in PEERS Group (n = 58) p
would show evidence of PEERS’ efficacy, by gaining
Experimental Waitlist control
knowledge of PEERS concepts and friendship skills, (n = 29) (n = 29)
increasing in their amount of get-togethers, and having Pre Pre
better quality friendships. Secondly, it was hypothesized M (SD) M (SD)
that adolescents in PEERS would decrease in their levels of Age (years) 14.00 (1.28) 13.31 (1.65) ns
social anxiety. Thirdly, it was hypothesized that adoles-
Gender (% male) 82.8 79.3 ns
cents in PEERS would significantly decrease in their levels
Race (% Caucasian) 96.3 89.7 ns
of autistic symptoms per parent and teacher report. Lastly,
Income (%) ns
it was hypothesized that adolescents in PEERS would
Under 25 K 3.6 7.4
significantly decrease in their problem behaviors and
25–50 K 14.3 14.8
increase in their social skills per parent and teacher report.
50–75 K 32.1 14.8
75–100 K 14.3 11.1
Over 100 K 35.7 51.9
Methods
Parent education (%)— ns
primary caregiver
Participants
High school 3.4 6.9
Some college 17.2 3.4
There were 58 adolescents between 11 and 16 years of age
B.A./B.S. 51.7 58.6
with ASD who participated in and completed this study
M.A./M.S. 17.2 6.9
with their parents. All participants had a previous and
Ph.D/M.D./J.D. 3.4 10.3
current diagnosis of ASD. 47 participants were male and 11
KBIT-2 Verbal IQ 102.17 (16.16) 98.45 (15.85) ns
were female. The average age of participants was 13.65-
ADOS Total Score 10.90 (3.46) 10.97 (3.25) ns
years-old (SD = 1.50). 52 of the participants identified
School type (% public 82.8 89.7 ns
themselves as Caucasian; 3 as African American; 1 as school)
Asian; and 2 chose not to communicate this information
Medication (% current 62.1 65.5 ns
(see Table 2, Demographics; see Fig. 1, Consort diagram, usage)
for details by group assignment). Vineland-communication 74.86 (11.70) 79.77 (11.67) ns
Vineland-socialization 70.14 (11.70) 73.58 (15.81) ns
Procedure Vineland-composite 72.71 (14.04) 79.42 (11.84) ns
The following measures had different n-values: Experimental Vine-
Recruitment and Eligibility
land-communication (n = 29); Waitlist Vineland-communication
(n = 27); Experimental Vineland-socialization (n = 29); Waitlist
Participants were recruited from local intervention agen- Vineland-socialization (n = 26); Experimental Vineland-composite
cies, autism support groups, and an in-house waiting list for (n = 28); Waitlist Vineland-composite (n = 26); Income experi-
mental (n = 28); Income waitlist (n = 27)
PEERS treatment, over a period of 2 years. Relationships
KBIT-2 Kaufman Brief Intelligence Test-Second Edition, ADOS
were established with local organizations, and permission
Autism Diagnostic Observation Schedule, ns nonsignificant
from the Institutional Review Board (IRB) was gained to
advertise at these sites. Families were provided with an
informational letter, which included a phone number and
email address for the study. Interested families were then
contacted by a graduate research assistant in order to by the parent; (c) English fluency for the adolescent;
conduct a phone screening. Phone screenings consisted of (d) parent or family member was a fluent English speaker and
telling the family about the program, gauging if the ado- was willing to participate in the study; (e) no history of
lescent met the inclusion criteria (see below), and gaining a adolescent major mental illness, such as bipolar disorder,
sense of the adolescent’s interest in participating in the schizophrenia, or psychosis; (f) no history of hearing, visual,
program. If the family passed the phone screening, then the or physical impairments which precluded the adolescent
graduate research assistant scheduled an approximately from participating in PEERS activities; (g) a previous and
2.5 h-long intake with the family. current diagnosis of either HFA, AS, or Pervasive Devel-
Inclusion criteria for adolescents were: (a) chronological opmental Disorder—NOS, with current as assessed via the
age between 11 and 16 years; (b) social problems as reported Autism Diagnostic Observation Schedule (ADOS: Lord

123
 J Autism Dev Disord

Assessed for eligibility treatment group (EXP) or the waitlist control group (WL).
(N = 71) EXP families completed the intake and entered a PEERS
group immediately, after which they completed the outtake
Excluded (N = 8)
appointment. WL families completed the intake, did not
Intake

Did not show up to intake (N = 2)

enter PEERS, and completed the outtake appointment
Did not meet IQ criteria (N = 6) approximately 13 weeks later. WL families then entered
the next available PEERS group, no more than 14 weeks
later. Pre- and post-assessments were compared at week 1
and week 14 for both groups. This design allowed an
examination of differences between these two groups over
Randomized (N = 63)
a 14-week period, in which the EXP group received the
intervention, while the WL group had not yet received the
Allocated to WL Allocated to EXP intervention. PEERS group sizes were maintained at 10 or
(N = 29) (N = 34)
fewer adolescents, and consisted entirely of either EXP or
Allocation

Received allocated Received allocated

intervention (N = 29) intervention (N = 34)
WL families (i.e., EXP families did not participate in
intervention groups containing WL families).
There were two events of the data collection process,
including (1) the intake that occurred before PEERS treat-
ment and (2) the outtake that occurred after PEERS treatment
(EXP) or after the 14-week delay (WL). At the intake visit,
written informed consent and assent were obtained, adoles-
Outtake

Discontinued intervention Discontinued intervention cent interest was confirmed via the Mental Status Checklist
(N = 0) (N = 5) (Laugeson and Frankel 2010), adolescent language skill,
Assessed at Outtake Assessed at Outtake
(N = 29) (N = 29) ASD diagnosis, and IQ were confirmed, and research mea-
sures were completed (see ‘‘Measures’’, below). Adolescents
and parents completed the measures in the presence of the
Analyzed (N = 29) Analyzed (N = 29) research team, while teachers were given the measures by the
Analysis

adolescent or parent in order to complete and mail back to the

research team. Teachers were blind to the condition assigned
to the subject. Once the PEERS intervention was complete,
Fig. 1 CONSORT recruitment diagram
the outtake was scheduled. During the outtake, all of the
same measures, excluding the diagnostic test, cognitive test,
and interest checklist, were conducted.
et al. 2001); and (h) a adolescent verbal IQ of 70 or above
assessed via the Kaufman Brief Intelligence Test-Second Treatment
Edition (Kaufman and Kaufman 2005; Laugeson et al.
2009). In order to gauge the adolescents’ motivation for The PEERS intervention consisted of 90-min sessions,
participating in the intervention, the study only included delivered once a week, over the course of 14-weeks.
adolescents who verbally expressed an interest in learning PEERS was provided in either a fall (August–December) or
how to make and keep friends. Further, adolescents com- spring (January–May) session. Treatment followed the
pleted a screening questionnaire in which they were asked commercially-available PEERS manual (Laugeson and
explicitly if they were interested in the group (Adolescent Frankel 2010). Parents and adolescents attended separate,
Mental Status Checklist: Laugeson and Frankel 2010). concurrent sessions where they learned how to make and
Adolescents who came to the intake visit and did not meet maintain friends and implement the rules taught. Prior to
inclusion criteria were compensated with a $30 gift card. study initiation, one of the study authors (Van Hecke)
Adolescents who met criteria were provided with a $30 attended an official PEERS training workshop in Los
incentive prize at the completion of PEERS. PEERS inter- Angeles, CA, and was certified in providing PEERS. She
vention was provided free of charge to families. then returned to the site of the current study and trained
graduate students in a clinical psychology Ph.D. program
Participant Assignment and Data Collection to assist with and co-lead the PEERS adolescent and
caregiver groups, and undergraduate students to serve as
Prior to the intake, eligible participants were randomly coaches/assistants for the PEERS groups. All graduate
assigned to one of two conditions, either the experimental students had extensive experience in research, diagnostic,

123
J Autism Dev Disord

and clinical practice in ASD, and all adolescent group was outlined in a handout given in the parent group, was
leaders had at least a Master’s degree in clinical psychol- conducted (see Table 1). Parents were given instruction on
ogy and had completed formal coursework in general ways in which they could help their adolescent overcome
aspects of group therapy. There were five graduate students hindrances to weekly socialization homework assignments.
involved in the study as group leaders, with three students At the end of group, either parent or adolescent, home-
leading the adolescent groups included in this analysis. work was assigned for the coming week, allowing time to
Adolescent and parent graduate group leaders received troubleshoot potential obstacles to homework completion.
training via observing the certified author (Van Hecke) Multiple homework assignments were given on a weekly
conducting sessions. The certified author conducted the basis, and typically corresponded to the current didactic
first adolescent group, in order for the first, most senior lesson. The sessions concluded with parents and adolescents
graduate student to train with her. The senior graduate reuniting in the same room, where the adolescents provided a
student then led an adolescent group in the next cohort, brief review of the lesson for parents, and homework
with the next most senior graduate student training with assignments were finalized. In order to minimize parent-
her, and another graduate student co-leading the parent adolescent conflict during the completion of these assign-
group with the certified author. This pattern was repeated, ments, the level of parental involvement as well as adoles-
such that group leaders typically co-led a parent group first. cent refusal to do the homework was individually negotiated
Then, they co-led an adolescent group with a more expe- at the end of the session with the help of group leaders
rienced interventionist or the certified interventionist. In (Laugeson et al. 2009). Homework compliance was strongly
subsequent cohorts, they were then allowed to lead an enforced, and failure to attempt more than two homework
adolescent group independently. When leading teen ses- assignments resulted in dismissal from the group. In addi-
sions independently, leaders and the certified author tion, families were allowed two absences to sessions, and, if
reviewed video of their own sessions with her and received exceeded, families were dismissed from the group.
feedback and supervision weekly. During each semester,
the certified author observed the adolescent group’s first,
Measures
midpoint, and final sessions in order to check treatment
provision accuracy.
Descriptive Measures
Undergraduate research assistants acted as ‘‘coaches’’ in
the adolescent sessions with at least one coach in each
At the intake visit, caregivers were asked to complete a
session. Coaches helped with role-play activities, behav-
demographic questionnaire and a questionnaire concerning
ioral rehearsal, and behavioral management. These coaches
their adolescent’s health and medication status. Diagnoses
were undergraduate students in psychology and health
were confirmed using the Autism Diagnostic Observation
sciences and were trained in all aspects of the PEERS
Schedule Modules 3 and 4 (ADOS-G: Lord et al. 1999),
intervention. Undergraduates also monitored the treatment
given by examiners trained to research-level reliability.
protocol for adherence in the adolescent sessions through
Adolescents’ cognitive abilities were assessed via the
completion of weekly fidelity check sheets covering all
Kaufman Brief Intelligence Test-Second Edition (Kaufman
elements of the intervention. Their role was to view the
and Kaufman 2005).
session outline and follow along with the group leader.
Further, if the group leader missed a main point of the
session, the research assistant would politely interrupt the Kaufman Brief Intelligence Test-Second Edition
leader and remind them to discuss a missed point.
The PEERS adolescent group always began with a Adolescent verbal intellectual functioning was assessed
homework review of the assignment from the previous using the verbal subscale of the Kaufman Brief Intelligence
week. Adolescents were then taught specific social skills Test-Second Edition (KBIT-2; Kaufman and Kaufman
for the week. Regarding the adolescent group’s didactic 2005), which takes approximately 25 min to administer.
lessons, they were enhanced by demonstrations in which Normative data is available and expressed as standard
the group leaders modeled the appropriate social skill being scores with a mean of 100 and a standard deviation of 15.
taught through role-play exercises. The newly learned The KBIT-2 demonstrates good psychometric estimates,
skills and rules for that week were then rehearsed by the including an internal reliability for the IQ composite of
adolescents in the session, while receiving feedback from 0.93, a test–retest reliability range of 0.88–0.89, and a
the group leader and coaches. standard error of the measurement of 4.3 points (Kaufman
In the parent session, time was devoted to trouble- and Kaufman 2005). The KBIT-2 has also been shown to
shooting any problems that may have occurred due to the be comparable to the Wechsler Intelligence Scale for
incompletion of homework. Next, a didactic lesson, which Children-fourth edition (WISC-IV), in terms of acceptable

123
 J Autism Dev Disord

correlations with the WISC-IV for diverse populations higher scores reflecting greater knowledge of the taught
(Walters and Weaver 2003). social skills. According to Laugeson et al. (2009), coeffi-
cient alpha for the TASSK was 0.56. However, they
Autism Diagnostic Observation Schedule-Generic asserted that this was acceptable, given the large domain of
questions on the scale. In the current study, the TASSK
The Autism Diagnostic Observation Schedule-Generic coefficient alpha was similarly very low, as the questions
(ADOS-G; Lord et al. 1999), Modules 3 or 4, is a struc- on the TASSK were not expected to cohere with one
tured, interview-based observational assessment conducted another.
with the adolescent. The adolescent is presented with
activities and questions which aim to elicit communicative Quality of Socialization Questionnaire
and social behaviors that are typically difficult for indi-
viduals with ASD. Algorithm scores for communication In order to assess PEERS efficacy, the Quality of Sociali-
and socialization are calculated to support the likelihood, zation Questionnaire (QSQ; Frankel and Mintz 2008) was
or lack thereof, of ASD diagnosis. The ADOS-G typically used. The QSQ is comprised of 12 items that are admin-
takes 30–45 min to complete and has excellent test–retest istered to parents (QSQ-P-R) and adolescents (QSQ-A-R)
reliability (0.82) and inter-rater reliability (0.92) (Lord independently to assess the frequency of adolescent get-
et al. 2001). All participants enrolled in the study obtained togethers with peers, number of friends involved, and the
combined scores (Communication and Social Interaction) level of conflict during these get-togethers. Two items ask
above the algorithm diagnostic threshold for ASD, thus for an estimate of the number of hosted and invited get-
confirming their previous ASD diagnosis. togethers the adolescent has had over the previous month.
The QSQ was developed through factor analysis of 175
Vineland Adaptive Behavior Scales-Second Edition boys and girls (Laugeson et al. 2009). Given that the total
get-togethers variable consists of only two question items,
The Vineland Adaptive Behavior Scales-Second Edition coefficient alpha was not provided by the developer of the
(Vineland II—Survey Form; Sparrow et al. 2005) is a mea- instrument and was not calculated in the current study.
sure of adaptive behavior skills needed for everyday living
for individuals and provides an assessment of adolescent Friendship Qualities Scale
functioning within the domains of communication, daily
living skills, and socialization. The Vineland-II took parents In order to assess PEERS efficacy, the Friendship Qualities
approximately 30 min to complete. Only the communica- Scale (FQS; Bukowski et al. 1994) was completed by
tion, socialization, and composite scores were reported in adolescents. The FQS assesses the adolescent’s perceptions
this study. Parents rated the degree to which their adolescent of the quality of his/her best friendships. It has 23 items,
exhibited each behavior item as either ‘‘Never,’’ ‘‘Some- each on a scale from 1 to 5, where 1 means not true and 5
times/Partially,’’ or ‘‘Usually.’’ Domain and Adaptive means really true. It takes approximately 10 min to com-
Behavior Composite scores are presented as standard scores plete. Adolescents are instructed to identify their best
with a mean of 100 and a standard deviation of 15. Higher friend and keep this friendship in mind when completing
scores represented better adaptive functioning. Reliability this measure. An example of an item is, ‘‘My friend and I
coefficients for the Adaptive Behavior Composite score are spend all of our free time together.’’ The Total score ranges
in the mid-90’s. Content validity has been established for from 23 to 115, with higher scores reflecting better quality
each domain of the Vineland-II (Sparrow et al. 2005). friendships. Previous research has noted that confirmatory
factor analysis supported the structure of the measure, and
Questionnaire Measures comparisons between ratings by reciprocated versus non-
reciprocated friends supported the discriminant validity of
Test of Adolescent Social Skills Knowledge the measure (Bukowski et al. 1994). In the current study,
the coefficient alpha for the Total score was acceptable at
In order to assess PEERS efficacy, the Test of Adolescent 0.89.
Social Skills Knowledge (TASSK; Laugeson and Frankel
2010) was completed by adolescents. The TASSK consists Social Interaction Anxiety Scale
of 26-items that assess the adolescent’s knowledge about
the specific social skills taught during the intervention. In order to assess social anxiety, the SIAS (Mattick and
Two items are included from each of the 13 didactic les- Clarke 1998) was completed by adolescents. The SIAS was
sons. The TASSK is comprised of sentence stems and two designed to measure feelings of anxiety in social interac-
possible answers. Total scores range from 0 to 26, with tions, with the main concerns relating to ‘‘being

123
J Autism Dev Disord

inarticulate, boring, sounding stupid, not knowing what to adolescents. Social Skills scale coefficient alphas were 0.93
say or how to respond within social interaction, and of for teacher and 0.90 for parent forms and for the Problems
being ignored.’’ The SIAS is comprised of 20 items, and Behavior scale they were 0.86 and 0.81, respectively. Cor-
participants’ rate each item on a 0 (not at all) to 4 (extre- relations between teacher and parent forms were low (Social
mely) scale based on how characteristic they believe each Skills and Problem Behavior scales r’s = 0.36) but statisti-
statement is of them. Total scores are computed, and they cally significant. Both scales were transformed into standard
range from 0 to 80, with higher scores indicating more scores with a mean of 100 and standard deviation of 15.
anxiety. Internal consistency for the items on this measure Higher scores on the Social Skills scale indicated better social
is excellent, with a Cronbach’s alpha of 0.94 in a large functioning and lower scores on the Problem Behavior scale
sample. The test–retest reliability for up to a 12-week indicated better behavioral functioning. In the current
period between tests is excellent (r = 0.90; Mattick and study, the coefficient alphas were acceptable (Social Skills-
Clarke 1998). In the current study, the coefficient alpha Parent = 0.91, Problem Behavior-Parent = 0.91, Social
was acceptable (Total Score = 0.89). Skills-Teacher = 0.88, Problem Behavior-Teacher = 0.81).

Social Responsiveness Scale

Results
In order to assess adolescents’ core autistic symptomatol-
ogy, the Social Responsiveness Scale (SRS; Constantino Table 2 presents the mean demographic variables for each
2005) was used. The parent (SRS-P) and teacher (SRS-T) group. Chi square analyses for gender, race, income, parent
forms of this measure were used in this study. The SRS is a education, school type, and medication use between the
65-item rating scale that measures the severity of autism groups were not significant (p’s [ 0.22). T-tests for group
spectrum symptoms as they occur in natural social settings differences on age, grade, KBIT-2 Verbal IQ, Vineland
and takes approximately 15–20 min to complete. It is Communication and Socialization subscales, Vineland
appropriate for use with children through adolescents from Composite scale, and intake questionnaire scores all failed
4 to 18 years of age. Each item is rated on a scale from ‘‘0’’ to reach significance, (p’s [ 0.05). In addition, all baseline
(never true) to ‘‘3’’ (almost always true). Of interest to this variables were not significantly different amongst groups at
study, the SRS generates a Total raw score that serves as an the pre-PEERS assessment point (p’s [ 0.05). Potential
index of severity of social deficits in the autism spectrum, group leader effects were analyzed, and it was found that
whereupon higher scores indicate higher severity of autism outcome variables did not vary due to differences in group
symptoms. Other subscales of this measure were not used leader assignment. Another preliminary analysis of atten-
in this particular study due to the multitude of variables and dance rates was conducted, and results indicated that out-
analyses. The psychometric properties of the SRS have come variables did not vary due to the number of sessions
been previously tested in studies involving over 1,900 attended, which was never more than two absences.
children ages 4–15 years and have yielded good reliability Examination of distributions, separately by group, time,
and have demonstrated good validity. Specifically, previ- and in total, revealed no significant underlying problems
ous research has found that the test–retest reliability coef- with the assumptions of normality, homogeneity of vari-
ficient was 0.88 for the Total scaled score (Constantino and ance, or outlying values in all measures, except for the
Todd 2000; Constantino and Todd 2003). In the current QSQ-A-R. For this measure, seven significant outliers were
study, the coefficient alpha for the Total score was noted over the pre- and post-test QSQ-A-R adolescent and
acceptable at 0.84. parent scores. More specifically, five adolescent outliers
(2 host pre, 1 invite pre, 1 host post, 1 invite post) and two
Social Skills Rating System parent outliers (1 host pre and 1 invite pre) were found.
These values were replaced with the next most extreme
The Social Skills Rating System (SSRS; Gresham and Elliott value in the distribution (Winsorization: Howell 2012).
1990) consists of 38-items and took approximately 10 min to Due to the large number of parent and adolescent
complete. Questionnaires were completed independently by dependent variables, a Group (EXP vs. WL) 9 Time (pre
the adolescent’s parent (SSRS-P) and teacher (SSRS-T). For vs. post) repeated measures multivariate analysis of vari-
example, items included ‘‘Starts conversations rather than ance (MANOVA) approach was used to analyze the data.
waiting for someone to talk first.’’ The items were rated as In order to appropriately run the MANOVA, multiple
either ‘‘Never,’’ ‘‘Sometimes,’’ or ‘‘Very Often.’’ The Social imputation was conducted for any missing items/questions
Skills and Problem Behaviors scales were derived from factor (Howell 2012). For each scale, the missing data was less
analysis. Gresham and Elliott (1990) reported the psycho- than 10.34 %. Regarding the teacher data, multiple impu-
metric properties of the parent and teacher forms for tation was not conducted, as there was a larger amount of

123
 J Autism Dev Disord

missing data. Therefore, for the teacher data three, separate p \ 0.05, Mpre = 106.28, SDpre = 21.62, Mpost = 98.55,
2 9 2 repeated measures analyses of variance (ANOVAs) SDpost = 22.53. Further, the EXP group showed a signifi-
were conducted for each teacher outcome variable. All cant decrease in problem behaviors on the SSRS-P,
statistical tests were analyzed at p \ 0.05 in SPSS 19.0 t(27) = 2.10, p \ 0.05, Mpre = 154.79, SDpre = 7.49,
(SPSS for Windows 2011). Mpost = 150.59, SDpost = 10.21, while the WL group did
Results of the repeated measures MANOVA revealed that not, t(27) = -0.37, ns, Mpre = 153.21, SDpre = 10.57,
the main effect of Group was significant for combined ado- Mpost = 153.72, SDpost = 7.69.
lescent and parent outcome variables, Wilks’ Lambda = To examine the teacher data, three separate 2 9 2 mixed
0.41; F(1, 56) = 4.39, p \ 0.001. The main effect for Time model repeated measures ANOVAs were conducted, for
was also significant, Wilks Lambda = 0.17, F(1, 56) = the SRS Total score, SSRS-T social skills, and SSRS-T
16.68, p \ 0.001. However, both of these main effects were problem behaviors scales. There were no significant find-
qualified by a significant Group by Time interaction, ings for the SRS and SSRS social skills scales, however, it
Wilks’ Lambda = 0.19; F(1, 56) = 13.54, p \ 0.001 (see was found that the teacher-rated SSRS problem behaviors
Table 3). Moreover, the Group X Time interaction reached scale yielded significant results. There was a significant
significance for four adolescent outcome measures at the main effect for time, Wilks Lambda = 0.84, F(1, 39) =
univariate level: TASSK, F(1, 56) = 146.45, p \ 0.001, 7.41, p \ 0.01. However, there was a significant Group X
partial g2 = 0.72; QSA-A-R (hosted get-togethers), F(1, Time interaction, Wilks Lambda = 0.91, F(1, 39) = 3.93,
56) = 10.02, p \ 0.005, partial g2 = 0.15; QSA-A-R p \ 0.05. A post hoc paired t test with a Bonferroni cor-
(invited get-togethers), F(1, 56) = 7.50, p \ 0.01, partial rected alpha level was performed to further examine this
g2 = 0.12; SIAS, F(1, 56) = 6.78, p \ 0.01, partial g2 = interaction. The analysis confirmed that the EXP group
0.12; and two parent outcome measures: SRS (Total), F(1, significantly decreased in problem behaviors on the SSRS-T,
56) = 9.38, p \ 0.01, partial g2 = 0.14; SSRS Problem t(19) = 2.82, p \ 0.01, Mpre = 135.19, SDpre = 8.64,
Behaviors, F(1, 56) = 3.75, p \ 0.05, partial g2 = 0.06. Mpost = 130.43, SDpost = 7.99, while the WL group did
Post hoc paired t-tests with a Bonferroni corrected alpha not, t(18) = 0.70, ns, Mpre = 135.45, SDpre = 6.92,
level were performed on the adolescent and parent outcome Mpost = 134.70, SDpost = 7.45.
variables to further examine the univariate interactions.
Regarding the adolescent measures, analyses confirmed that
the EXP group significantly improved in knowledge of Discussion
PEERS concepts and friendship skills on the TASSK,
t(27) = -17.91, p \ 0.001, Mpre = 13.34, SDpre = 2.72, The current study presented the results of a randomized
Mpost = 21.90, SDpost = 3.05, while the WL group did not, controlled replication of PEERS, a manualized, parent-
t(27) = -1.47, ns, Mpre = 13.38, SDpre = 2.98, Mpost = assisted intervention to improve friendships for 58 ado-
14.03, SDpost = 2.77. The EXP group showed a significant lescents with ASD, the second largest number of partici-
increase in hosted get-togethers on the QSQ-A-R, t(27) = pants reported in the ASD treatment outcome literature for
-3.60, p \ 0.001, Mpre = 1.41, SDpre = 3.32, Mpost = individuals 6–21 years-old (Reichow et al. 2012). The
3.69, SDpost = 3.24, while the WL group did not, results of this replication and extension of the PEERS
t(27) = 1.36, ns, Mpre = 2.77, SDpre = 4.76, Mpost = 1.52, intervention were encouraging, as improvement was dem-
SDpost = 3.10. The EXP group showed a significant onstrated on 7 of 14 outcome measures.
increase in invited get-togethers on the QSQ-A-R, Most hypotheses were supported in this study and were
t(27) = -3.44, p \ 0.005, Mpre = 0.41, SDpre = 0.87, also a replication of similar results found in the original
Mpost = 1.39, SDpost = 1.73, while the WL group did not, PEERS study (Laugeson et al. 2009). In the current study,
t(27) = 0.58, ns, Mpre = 1.08, SDpre = 2.10, Mpost = 0.90, the experimental treatment group showed evidence of
SDpost = 1.42. The EXP group showed a significant PEERS efficacy, by gaining knowledge of PEERS concepts
decrease in social anxiety on the SIAS, t(27) = 3.19, and friendship skills. Although it is not completely unex-
p \ 0.005, Mpre = 32.28, SDpre = 14.39, Mpost = 24.72, pected that adolescents displayed retention of learned
SDpost = 9.67, while the WL group did not, t(27) = -0.04, information, this finding does point to the effectiveness of
ns, Mpre = 26.83, SDpre = 13.44, Mpost = 26.90, SDpost = PEERS in teaching the targeted social skills. Further, the
16.03. Regarding the parent measures, analyses revealed experimental treatment group showed an increase in hosted
that the EXP group significantly decreased in core autistic and invited get-togethers. In the original PEERS study, a
symptoms on the SRS (Total score), t(27) = 6.24, significant increase in hosted get-togethers was found as
p \ 0.001, Mpre = 101.17, SDpre = 23.08, Mpost = 79.12, well, however, they did not find a significant increase in
SDpost = 20.21, in addition to the WL group also showing a invited get-togethers. It seems that adolescents are culti-
smaller decrease in core autistic symptoms, t(27) = 2.52, vating reciprocal relationships during the intervention, as

123
J Autism Dev Disord

Table 3 Means and SD for outcome variables for experimental treatment and waitlist control groups
Group (n = 58) p
Experimental (n = 29) Waitlist control (n = 29)
Pre Post Pre Post
M (SD) M (SD) M (SD) M (SD)

Adolescent
TASSK 13.34 (2.72) 21.90 (3.05) 13.38 (2.98) 14.03 (2.77) 0.001
QSQ-A-R (host) 1.41 (3.32) 3.69 (3.24) 2.77 (4.76) 1.52 (3.10) 0.005
QSQ-A-R (invite) 0.41 (0.87) 1.39 (1.73) 1.08 (2.10) 0.90 (1.42) 0.01
SIAS 32.28 (14.39) 24.72 (9.67) 26.83 (13.44) 26.90 (16.03) 0.01
QSQ-A-R (conflict) 1.62 (2.97) 2.79 (3.29) 3.83 (5.28) 3.03 (6.65) ns
FQS 83.71 (14.88) 82.45 (15.41) 86.64 (15.04) 82.65 (19.42) ns
Parent
SRS 101.17 (23.08) 79.12 (20.21) 106.28 (21.62) 98.55 (22.53) 0.005
SSRS problem behaviors 154.79 (7.49) 150.59 (10.21) 153.21 (10.57) 153.72 (7.69) 0.05
QSQ-P-R (host) 0.57 (1.20) 2.11 (2.06) 1.05 (1.46) 1.51 (1.69) ns
QSQ-P-R (invite) 0.91 (2.02) 1.50 (2.09) 0.86 (1.94) 0.63 (0.82) ns
QSQ-P-R (conflict) 2.52 (5.00) 2.44 (3.40) 5.52 (7.82) 1.97 (3.94) ns
SSRS social skills 112.79 (10.34) 119.76 (9.23) 110.41 (13.96) 114.28 (14.60) ns
Teachera
SSRS problem behaviors 135.19 (8.64) 130.43 (7.99) 135.45 (6.92) 134.70 (7.45) 0.05
SSRS social skills 124.19 (8.45) 127.14 (6.30) 124.30 (8.09) 123.75 (11.77) ns
SRS 77.90 (29.94) 67.95 (27.46) 90.35 (17.95) 82.75 (27.38) ns
TASSK = Test of Adolescent Social Skills Knowledge; QSQ-P-R = Quality of Socialization Questionnaire—Parent; QSQ-A-R = Quality of
Socialization Questionnaire—Adolescent; SIAS = Social Interaction Anxiety Scale; FQS = Friendship Qualities Scale; SRS = Social
Responsiveness Scale; SSRS = Social Skills Rating Scale; p = probability, p = interaction p value
a
n’s are 21 for experimental and 20 for waitlist control groups

they are also being invited to get-togethers. This overall control group, although these changes also failed to reach
finding is important, as get-togethers provide an opportu- traditional levels of significance (see Table 3). This dem-
nity for adolescents to practice their social skills and onstrates that with both parent and teacher ratings of ado-
develop meaningful friendships. lescents’ social skills, scores moved in the predicted
In contrast to the original PEERS study, the current direction, although a larger sample might be needed to
study did not find that friendship quality significantly reach statistical significance. It will also be illuminating to
improved in the experimental treatment group. Upon closer examine parent and teacher ratings of adolescent’ social
examination, in the original PEERS study, the significant skills at a longer-term follow-up appointment. This trend of
friendship quality finding was due to the waitlist control continued or later improvement was previously found by
group demonstrating worse friendship quality over time the UCLA group at 14-weeks post-PEERS (Laugeson et al.
(Laugeson et al. 2009). Therefore, it not yet known if 2012).
PEERS positively affects friendship quality, or this may be This study aimed to extend current findings relating to
a domain that requires more than 14 weeks to develop. PEERS in several ways. With the addition of a new mea-
The original PEERS study found a significant increase sure, it was found that the experimental treatment group
in adolescent social skills per parent report (specifically, on significantly decreased in their social anxiety symptoms as
the SSRS social skills scale). In the current study, parents’ compared to the waitlist control group from pre- to post-
ratings of adolescent social skills on the SSRS increased PEERS. Social anxiety reduction is not targeted in the
from pre- to post-PEERS for the experimental treatment PEERS intervention, which makes this finding even more
group, although this change failed to reach traditional significant. Further, this finding is of great importance as it
levels of statistical significance (see Table 3). It was also suggests that by teaching adolescents with ASD social
found that teachers’ scores of adolescents’ social skills on skills and thus increasing the likelihood of more positive
the SSRS increased from pre- to post-PEERS for the peer interactions, the common trajectory of heightened
experimental treatment group and decreased for the waitlist social anxiety in ASD (Bellini 2006) is altered. Instead,

123
 J Autism Dev Disord

learning and practicing social skills may create a sense of Additionally, more attention should be paid to capturing
confidence and comfort for adolescents with ASD in social teacher report, as teachers not only provide another infor-
situations, perhaps counteracting their social anxiety. mant, but would also not be subject to the same biases in
Another extension of previous findings was that the reporting outcome as parents, since they are not directly
experimental treatment group significantly decreased in involved in treatment. Although this study had a better
their levels of autistic symptoms per parent report (spe- return rate of teacher data as compared to the original
cifically, on the SRS scales), as compared to the waitlist PEERS study, there was still a large amount of missing
control group, from pre- to post-PEERS. It should be noted teacher data, which may have decreased power in analyses.
that, for the waitlist control group, autistic symptoms
changed for the better from pre- to post-PEERS (an 8 point
mean difference), however, the decrease from pre- to post- Future Directions and Conclusions
PEERS was not as large as the experimental treatment
group’s difference (22 point mean difference). This sug- One future direction of the current study includes gather-
gests that PEERS had a more profound effect on parents’ ing data, especially on social anxiety and friendship
rating of adolescent core autistic symptoms. This decrease development, at a long-term follow-up. This would yield
in core autistic symptoms might allow adolescents who useful information toward determining the durability of
participated in PEERS to better function in day to day life the findings as well as assess for any changes that occur
in addition to being more successful in social interactions. during the months following PEERS. For example, per-
The fact that PEERS led to a drop in autistic symptom- haps once adolescents have had some time outside of
atology from the ‘‘severe’’ level to the ‘‘moderate’’ severity PEERS, their friendship quality may improve as they gain
level gives additional support to utilization of the PEERS confidence and practice the skills they have learned in
intervention with adolescents with ASD (Aldridge et al. meaningful relationships. Recent report by the PEERS
2012). developer indicate that 14 weeks after PEERS, there was
A new finding relating to PEERS was that the experi- maintenance of social skills knowledge, social respon-
mental treatment group significantly decreased in their siveness, and overall improvements in social skills
problem behaviors per parent and teacher report. These (Laugeson et al. 2012), and that some of these improve-
problem behaviors included items relating to aggressive ments continued to be evident one to 5 years later (Man-
acts, poor temper control, sadness, anxiety, fidgeting and delberg et al. 2011). Lastly, it might also be helpful to gain
impulsive acts. This finding is extremely robust as both a physiological measure of anxiety that is not dependent
parents and teachers similarly agreed upon ratings. This on self-report, and a behavioral measure of social skills, as
suggests that teaching social skills may positively affect discussed above.
other domains of behavior. Further, it may be that ado- Social anxiety and social skills are likely related to one
lescents substitute problematic behavior with more positive another (Bellini 2004). In addition, those with AS/HFA
social behavior. Overall, these findings indicate that have been found to significantly report more social anxiety
PEERS is effective at other sites, in addition to its site of symptoms than their typically developing peers (Sebastian
development, and that a reduction in social anxiety is et al. 2009). Thus, it is highly important to focus on social
another potential outcome for adolescents with ASD who anxiety in treatment with individuals with AS/HFA. Future
complete the program. social skills interventions, including PEERS, should aim to
teach adolescents with AS/HFA how to handle social
anxiety in addition to providing social skills training.
Limitations of the Present Study The present study was a replication and extension of the
PEERS intervention and greatly adds to the minimal lit-
There were some limitations to the present study. The erature regarding social skill interventions for adolescents
sample included mostly males who were Caucasian. This with AS/HFA. This study provides the first independent
lack of diversity in the sample causes the findings to be less replication of a social skills treatment for adolescents with
generalizable to a larger, more diverse population. Another ASD, and thus greatly augments knowledge on treatment
limitation was that the parent ratings may have been biased efficacy. In addition, the current study supported previously
due to the parent involvement in the intervention. The noted positive outcomes of participation in PEERS, as well
absence of behavioral ratings of social skills is another as found that involvement in PEERS decreased social
limitation of this study, as paper-and-pencil rating scales anxiety, core autistic symptoms, and problematic behav-
were used. Future studies should include in vivo measures iors. These findings suggest that PEERS is an appropriate
of adolescents’ social skills in naturalistic interactions. intervention for widespread national use.

123
J Autism Dev Disord

Acknowledgments This research was supported by a Research Frankel, F., & Myatt, R. (2003). Children’s friendship training. New
Development Grant and a Regular Research Grant from Marquette York: Brunner-Routledge.
University, and a Grant from the Autism Society of Southeastern Goldstein, P., & McGinnis, E. (2000). Skillstreaming the adolescent:
Wisconsin. We would like to acknowledge undergraduate assistants New strategies and perspectives for teaching prosocial skills.
Rheanna Remmel, Chelsea Gasaway, Grand McDonald, Noelle Fritz, Champaign: Research Press.
Alexandra Reveles, Jenna Kahne, Janel Wasisco, Justin Abraham, Gresham, F. M., & Elliott, S. (1990). The social skills rating system.
Jennifer Hilger, Nina Linneman, Rachel Olinger, Meghan Gwinn, and MN: American Guidance Service.
Benjamin Gemkow. We also wish to thank Elizabeth Laugeson, Gresham, F. M., Sugai, G., & Horner, R. H. (2001). Interpreting
Psy.D., UCLA, for her assistance in starting the PEERS program in outcomes of social skills training for students with high
Wisconsin, and the families that participated for their time and ded- incidence disabilities. Exceptional Children, 67(3), 331–345.
ication to clinical research. Grynszpan, O., Nadel, J., Constant, J., Le Barillier, F., Carbonell, N.,
Simonin, J., et al. (2011). A new virtual environment paradigm
for high-functioning autism intended to help attentional disen-
gagement in a social context. Journal of Physical Therapy
References Education, 25(1), 42–47.
Howell, D. (2012). Statistical methods for psychology. Canada:
Aldridge, F., Gibbs, V., Schmidhofer, K., & Williams, M. (2012). Cengage Learning.
Investigating the Clinical Usefulness of the Social Responsive- Howlin, P. (2000). Autism and developmental receptive language
ness Scale (SRS) in a Tertiary Level, Autism Spectrum Disorder disorder-a follow-up comparison in early adult life II: Social
Specific Assessment Clinic. Journal of Autism and Develop- behavioral and psychiatric outcomes. Journal of Psychology and
mental Disorders, 42, 294–300. Psychiatry and Allied Disciplines, 41, 561–578.
Bauminger, N., & Kasari, C. (2000). Loneliness and friendship in Kaufman, A. S., & Kaufman, N. L. (2005). Kaufman Brief
high-functioning children with autism. Child Development, Intelligence Test (K-BIT). Circle Pines, MN: American Guidance
71(2), 447–456. Service.
Bellini, S. (2004). Social skill deficits and anxiety in high-functioning Krasny, L., Williams, B. J., Provencal, S., & Ozonoff, S. (2003).
adolescents with Autism Spectrum Disorders. Focus on Autism Social skills interventions for the autism spectrum: Essential
and other Developmental Disabilities, 19(2), 78–85. ingredients and a model curriculum. Child and Adolescent
Bellini, S. (2006). The development of social anxiety in adolescents Psychiatric Clinics of North America, 12(1), 107–122.
with Autism Spectrum Disorders. Focus on Autism and other Laugeson, E. A. & Frankel, F. (Eds.). (2010). Test of Adolescent
Developmental Disabilities, 21(3), 138–145. Social Skills Knowledge, Mental Status Checklist. In The
Buhrmester, D. (1990). Intimacy of friendship, interpersonal compe- PEERS treatment manual. New York: Routledge.
tence, and adjustment during preadolescence and adolescence. Laugeson, E. A., Frankel, F., Gantman, A., Dillon, A. R., & Mogil, C.
Child Development, 61(4), 1101–1111. (2012). Evidence-based social skills training for adolescents with
Bukowski, W. M., Hoza, B., & Boivin, M. (1994). Measuring autism spectrum disorders: The UCLA PEERS Program. Journal
friendship quality during pre- and early adolescence: The of Autism and Developmental Disorders, 42(6), 1025–1036.
development and psychometric properties of the Friendship Laugeson, E. A., Frankel, F., Mogil, C., & Dillon, A. R. (2009).
Qualities Scale. Journal of Social and Personal Relationships, Parent-assisted social skills training to improve friendships in
11(3), 471–484. adolescents with autism spectrum disorders. Journal of Autism
Carnahan, C., Hume, K., Clarke, L., & Borders, C. (2009). Using and Developmental Disorders, 39(4), 596–606.
structured work systems to promote independence and engage- Lord, C., Rutter, M., DiLavore, P. C., & Risi, S. (1999). Autism
ment for students with Autism Spectrum Disorders. Teaching Diagnostic Observation Schedule (ADOS). Los Angeles, CA:
Exceptional Children, 41(4), 6–14. Western Psychological Services.
Centers for Disease Control and Prevention. (2012). Autism Spectrum Lord, C., Rutter, M., DiLavore, P. C., & Risi, S. (2001). Autism
Disorders. Atlanta, GA: Department of Health and Human Diagnostic Observation Schedule-WPS (ADOS-WPS). Los
Services. Angeles, CA: Western Psychological Services.
Chalfant, A., Rapee, R., & Carroll, L. (2007). Treating anxiety Mandelberg, J., Laugeson, E., Frankel, F., Gantman, A. & Bates, S.
disorders in children with high functioning Autism Spectrum (2011). Long term outcomes of a parent-assisted social skills
Disorders: A controlled trial. Journal of Autism and Develop- intervention for adolescents with autism: The UCLA PEERS
mental Disorders, 37(10), 1842–1857. program. Poster presented at the International Meeting for
Chambless, D. L., Baker, M. J., Baucom, D. H., Beutler, L. E., Autism Research (IMFAR) annual conference, San Diego, CA.
Calhoun, K. S., Crits-Christoph, P., et al. (1998). Update on Mattick, R. P., & Clarke, J. C. (1998). Development and validation of
empirically validated therapies, II. The Clinical Psychologist, measures of social phobia scrutiny fear and social interaction
51(1), 3–16. anxiety. Behaviour Research and Therapy, 36, 455–470.
Constantino, J. N. (2005). Social responsiveness scale. Los Angeles: McAfee, J. (2002). Navigating the social world: A curriculum for
Western Psychological Services. individuals with Asperger syndrome, high functioning autism
Constantino, J. N., & Todd, R. D. (2000). Genetic structure of and related disorders. Arlington, TX: Future Horizons, Inc.
reciprocal social behavior. American Journal of Psychiatry, 157, Mitchell, K., Regehr, K., Reaume, J., & Feldman, M. (2010). Group
2043–2045. social skills training for adolescents with asperger syndrome or
Constantino, J. N., & Todd, R. D. (2003). Autistic traits in the general high functioning autism. Journal on Developmental Disabilities,
population: A twin study. Archives of General Psychiatry, 60(5), 16(2), 52–63.
524–530. Ozonoff, S., & Miller, J. N. (1995). Teaching theory of mind: A new
Drotar, D. (2006). Psychological interventions in childhood chronic approach to social skills training for individuals with autism.
illness. Washington, DC: American Psychological Association. Journal of Autism and Developmental Disorders, 25(4), 415–
Frankel, F. & Mintz, J. (2008). Measuring the quality of play dates. 433.
Available from UCLA Parenting and Children’s Friendship Rao, P. A., Beidel, D. C., & Murray, M. J. (2008). Social
Program, 300 Medical Plaza, Los Angeles. skills interventions for children with Asperger’s syndrome or

123
 J Autism Dev Disord

high-functioning autism: A review and recommendations. Jour- syndrome and high functioning autism. Journal of Autism and
nal of Autism and Developmental Disorders, 38(2), 353–361. Developmental Disorders, 37(10), 1960–1968.
Reichow, B., Steiner, A., & Volkmar, F. (2012). Social skills groups Walters, S., & Weaver, K. (2003). Relationships between the
for people aged 6 to 21 with autism spectrum disorders (ASD). Kaufman Brief Intelligence Test and the Wechsler Adult
Cochrane Database of Systematic Reviews, 7, 1–48. Intelligence Scale-Third Edition. Psychological Reports, 92(3),
Sebastian, C., Blakemore, S., & Charman, T. (2009). Reactions to 1111–1115.
ostracism in adolescents with autism spectrum conditions. Jour- Weiss, M. J., & Harris, L. S. (2001). Teaching social skills to people
nal of Autism and Developmental Disorders, 39(8), 1122–1130. with autism. Behavior Modification, 25(5), 785–802.
Sofronoff, K., Dark, E., & Stone, V. (2010). Social vulnerability and White, S., Albano, A., Johnson, C., Kasari, C., Ollendick, T., Klin, A.,
bullying in children with Asperger syndrome. The National et al. (2010). Development of a cognitive-behavioral interven-
Autistic Society, 15(3), 1–18. tion program to treat anxiety and social deficits in adolescents
Sparrow, S., Balla, D., & Cicchetti, D. V. (2005). The Vineland with high-functioning autism. Clinical Child and Family
adaptive behavior scales (2nd ed.). Circle Pines, Minnesota: Psychology Review, 13(1), 77–90.
American Guidance Service. White, S., Leonig, K., & Scahill, L. (2007). Social skills development
SPSS for Windows. (2011). Release 20.0. Armonk, NY: IBM Corp. in children with autism spectrum disorder: A review of the
Tse, J., Strulovitch, J., Tagalakis, V., Meng, L., & Fombonne, E. intervention research. Journal of Autism and Developmental
(2007). Social skills training for adolescents with Asperger’s Disorders, 37, 1858–1868.

123