BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 1

Published in: Research in Developmental Disabilities, 26, 359-383 (2005)

A comparison of intensive behavior analytic and

eclectic treatments for young children with autism
Jane S. Howard
California State University, Stanislaus and The Kendall School

Coleen R. Sparkman
The Kendall School

Howard G. Cohen
Valley Mountain Regional Center

Gina Green
University of North Texas and San Diego State University

Harold Stanislaw
California State University, Stanislaus

The authors are grateful to Valley Mountain Regional Center and California State University,
Stanislaus for supporting this research; to Shannon Brackett, Beth LeBrun, Schelley McDonald,
Marie Overmyer, and Yasman Dianat for assistance with data collection; and to the children and
families who participated in the study. Preliminary reports of this research were presented at the
international conference of the Association for Behavior Analysis, Venice, Italy, November 2001
and the annual meeting of the California Association for Behavior Analysis, San Francisco,
February 2003. Requests for reprints should be addressed to Jane S. Howard, PhD, BCBA,
Psychology Department, California State University, Stanislaus, 801 W. Monte Vista Avenue,
Turlock, CA 95382.

Abstract
We compared the effects of 3 treatment approaches on preschool-age children with
autism spectrum disorders. Twenty-nine children received intensive behavior analytic
intervention (IBT; 1:1 adult:child ratio, 25-40 hours per week). A comparison group (n =
16) received intensive “eclectic” intervention (a combination of methods, 1:1 or 1:2 ratio,
30 hours per week) in public special education classrooms (designated the AP group). A
second comparison group (GP) comprised 16 children in nonintensive public early
intervention programs (a combination of methods, small groups, 15 hours per week).
Independent examiners administered standardized tests of cognitive, language, and
adaptive skills to children in all 3 groups at intake and about 14 months after treatment
began. The groups were similar on key variables at intake. At followup, the IBT group
had higher mean standard scores in all skill domains than the AP and GP groups. The
differences were statistically significant for all domains except motor skills. There were
no statistically significant differences between the mean scores of the AP and GP groups.
Learning rates at followup were also substantially higher for children in the IBT group
than for either of the other two groups. These findings are consistent with other research
showing that IBT is considerably more efficacious than “eclectic” intervention.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 2

A comparison of intensive behavior analytic

and eclectic treatments for young children with autism

Early intervention targets differences between the skills of children who have or
are at risk for developmental delays and the skills of their typically developing peers.
Those discrepancies may be small initially but are generally acknowledged to increase
with the passage of time (e.g., Guralnick, 1998; Ramey & Ramey, 1998). Developmental
trajectories are not fixed, however, even for children with known risk factors or
disabilities. Instead, each child’s progress can be influenced by many factors, such as
experience. As Ramey and Ramey (1998) noted, “…a widespread hope for early
intervention (is) . . . that children could be placed on a normative developmental
trajectory and thus continue to show optimal development after early intervention ends”
(p. 113). Accordingly, they posited a “zone of modifiability,” a period of time during
which the precise developmental trajectory for children at risk is likely determined by the
timing, intensity, and appropriateness of treatment. Convergent evidence supporting this
hypothesis has come from a variety of sources. Longitudinal studies (e.g., the North
Carolina Abecedarian Project, Infant Health and Development Program) demonstrated
that the effects of early intervention on children at risk for developmental delay and
mental retardation were evident when the children were 3 years old, and some gains were
maintained into adolescence and adulthood (Campbell, Pungello, Miller-Johnson,
Burchinal, & Ramey, 2001; Campbell, Ramey, Pungello, Sparling, & Miller-Johnson,
2002; for a review, see Ramey & Ramey, 1999).The likelihood that effective early
intervention can produce lasting neurobiological as well as behavioral changes has been
suggested by research showing that early experiences play a critical role in shaping brain
architecture as well as brain function (Dawson & Fischer, 1994; Shore, 1997).
Additionally, studies have shown that specific types of interactions with the physical and
social environment can remediate some types of damage to the central nervous system
(e.g., Hannigan & Berman, 2000). In a series of studies using mouse models of some
mental retardation syndromes and neurological disorders, Schroeder, Tessel, and their
colleagues demonstrated that behavior analytic discrimination training reversed
abnormalities in brain structures and neurotransmitter levels as well as learning and
behavior. Training was most effective when it began early in development (Loupe,
Schroeder, & Tessel,1995; Stodgell, Schroeder,& Tessel, 1996; Tessel, Schroeder,
Loupe, & Stodgell, 1995; VanKeuren, Stodgell, Schroeder, & Tessel, 1998).
Findings from early intervention research indicate that treatment that is intensive,
long in duration, and delivered directly to children (rather than just to their caregivers)
produces better outcomes than treatment that lacks those elements (Ramey & Ramey,
1998, 1999). Few of those variables have been isolated and investigated in controlled
studies, however. For example, despite the apparent relationship between the intensity of
early intervention and outcome (e.g., Guralnick, 1998), there has been little experimental
research on the effects of treatment intensity or duration. Nor has there been much
research on the relation between type of early intervention and outcomes. Guralnick
(1998) argued that the next generation of research in early intervention must progress
beyond basic demonstrations of its effectiveness. There is a need for studies that delineate
which aspects of early intervention are most efficacious, and for which populations. A
better understanding of the optimal timing, intensity, duration, and type of intervention
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 3

could benefit all children who have or are at risk for developmental delays. Given the
reported recent increase in the number of children diagnosed with autism spectrum
disorders, such issues may be particularly germane to this population (e.g., California
Department of Developmental Services 2003a; Yeargin-Allsop et al., 2003; but see
Fombonne, 2001, 2003 for critiques of such reports). In addition, the cost of lifespan
services for people with autism may be disproportionately higher than the cost of serving
individuals with other disabilities (e.g., California Department of Developmental
Services, 2002, 2003b). Effective early intervention can substantially reduce those costs
(Jacobson, Mulick, & Green, 1998). Therefore, there are several compelling reasons to
examine outcomes produced by various types of early intervention for children with
autism.
There is considerable empirical evidence that early intensive behavior analytic
intervention produces large and lasting functional improvements in many children with
autism. Although a number of behavior analysts have been documenting the effectiveness
of behavior analytic intervention for individuals with autism since the early 1960s (e.g.,
Ferster & DeMyer, 1961; Wolf, Risley, & Mees, 1964; see also Matson, Benavidez,
Compton, Paclawskyj, & Baglio, 1996), a study by Lovaas (1987) was singular for
documenting substantially improved functioning in a sizeable proportion of children who
received comprehensive, intensive, long-duration behavior analytic intervention starting
before they reached 4 years of age. Nine of 19 children in that study who received early
intensive behavior analytic treatment for at least two years had cognitive and language
test scores in the normal range by the age of 6-7 years and completed first grade without
special instruction. In contrast, few gains were made by children with autism in two
control groups who received either 10 hours of behavior analytic treatment per week or
typically available community services over the same time period. A follow-up study
found that the “best outcome” children from the Lovaas (1987) study continued to
function normally into adolescence (McEachin, Smith, & Lovaas, 1993).
Several studies of comprehensive, intensive behavior analytic treatment for young
children with autism spectrum disorders have been published prior to and since the
Lovaas (1987) study. Collectively, these studies have documented the efficacy of
intensive behavior analytic intervention, both center-based (e.g., Eikeseth, Smith, Jahr, &
Eldevik, 2002; Fenske, Zalenski, Krantz, & McClannahan, 1985; Harris, Handleman,
Gordon, Kristoff, & Fuentes, 1991) and home-based (e.g., Anderson, Avery, DiPietro,
Edwards, & Christian, 1987; Birnbrauer & Leach, 1993; Smith, Groen, & Wynne, 2000;
Weiss, 1999). In several studies, standardized test data indicated that cognitive
functioning, language skills, and academic performance approached or exceeded normal
levels in many children who received at least two years of early intensive behavior
analytic treatment (for a review, see Green, 1996 and Smith, 1999). Instruments such as
the Vineland Adaptive Behavior Scales also detected substantial improvements in
adaptive functioning (Anderson et al., 1987; Birnbrauer & Leach, 1993; Smith et al.,
2000; Weiss, 1999). Similar outcomes have been documented in systematic case studies
in which independent evaluators used objective measurement instruments to track
children’s progress (Green, Brennan, & Fein, 2002; Perry, Cohen, & De Carlo, 1995).
Finally, parents whose children received intensive behavior analytic intervention showed
high satisfaction and reduced stress over the course of treatment in comparison to parents
whose children did not receive intensive behavior analytic intervention (Anderson et al.,
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 4

1987; Birnbrauer & Leach, 1993; Smith et al., 2000).

Although all published studies of early intensive behavior analytic treatment
demonstrated that many children made substantial gains, outcomes varied within and
across studies. The proportions of intensively treated children who achieved normal or
near-normal functioning, more modest improvements, and relatively small improvements
varied from study to study (Green, 1996; Smith, 1999). For instance, a smaller percentage
of children in the Smith et al. (2000) study were able to function independently in regular
classrooms post-treatment than was reported by Lovaas (1987), and no children were
reported to be enrolled in general education settings without supports in the Anderson et
al. (1987) and Birnbrauer and Leach (1993) studies. Those studies differed in several
important ways from the Lovaas (1987) study, however. None involved the 40 hours of
intensive treatment per week that was provided to the experimental group in the Lovaas
(1987) study. Additionally, participants in those studies had lower pre-treatment language
and IQ scores and received intervention for a shorter period of time than their
counterparts in the Lovaas (1987) study. There were also methodological differences
across studies: some were quasi-experimental while others used true experimental
designs, and few assigned participants to groups randomly (see Green, 1996; Kasari,
2002; Rogers, 1998; Smith, 1999). Indeed, although some partial and systematic
replications of the Lovaas (1987) study have been published, so far no full replications
(40 hours of treatment per week for a minimum of two years; multiple outcome
measures; at least one control group) have appeared in the literature. Nevertheless, as an
aggregate, the published studies offer compelling evidence that many children with
autism who received early intensive behavior analytic treatment made substantial gains.
In contrast, there is little objective empirical evidence regarding the efficacy of
non-behavior analytic intervention models such as Treatment and Education of Autistic
and Related Communication Handicapped Children (TEACCH; e.g., Schopler, 1997) or
developmental approaches, such as the Colorado Health Sciences Program (Rogers &
DiLalla, 1991; Rogers, Herbison, Lewis, Pantone, & Reis, 1986). Of the total of 15 early
autism intervention outcome studies evaluated in three separate reviews, only five were
evaluations of what the authors characterized as non-behavior analytic treatments.
Reported treatment effects consisted of small mean gains in standardized test scores (e.g.,
IQ, language) or changes in developmental levels on measures not widely employed to
assess functioning in children; all had serious methodological limitations. Further, no
studies comparing early intensive behavior analytic treatment directly with TEACCH,
Colorado Health Sciences, or any other comprehensive treatment model have been
published to date (Kasari, 2002; Rogers, 1998; Smith, 1999). Several studies, however,
have compared outcomes of intensive behavior analytic treatment with those resulting
from standard interventions that are typically provided to children with autism through
public early intervention and special education programs. In the Lovaas (1987) study, the
41 participants in control groups 1 and 2 were described as receiving treatments
consisting of “resources in the community such as those provided by small education
classes.” Control group 1 also received behavior analytic treatment for 10 hours per
week. Few gains were documented for children in those groups over the course of two or
more years of treatment. Similarly, a comparison group of children in the study by Smith
et al. (2000) who were enrolled in public schools for 10 to 15 hours per week made little
improvement.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 5

Recently Eikeseth and colleagues (2002) compared the effects of intensive

behavior analytic treatment with equally intensive and relatively well-specified “eclectic”
treatment that is similar to the type of intervention that many children with autism receive
in public schools and some private programs. These investigators studied the effects of
intervention provided for 30 hours per week for one year on children with autism who
were 4-7 years of age when they entered treatment. Eleven children received behavior
analytic intervention, while 11 other children received intensive treatment using a
combination of methods including discrete trial training, TEACCH-based procedures,
and sensory integration therapy. All children received 1:1 treatment from therapists who
all had similar educational backgrounds and training. Each therapist received weekly
consultation from behavior analysts. Additional training was provided to parents and
therapists of children in the intensive behavior analytic treatment group. After one year
the children in the behavior analytic treatment group performed significantly better on
standardized measures of cognitive, language, and adaptive functioning than the children
in the intensive “eclectic” treatment group. For example, children in the behavior analytic
treatment group gained an average of 17 points on standardized measures of cognitive
functioning. At followup, seven children in the behavior analytic treatment group
achieved scores in the normal range of functioning, while only two children in the
“eclectic” treatment group produced scores in the normal range. These results suggested
that the type, rather than the intensity, of treatment accounted for the outcomes produced
by intensive behavior analytic treatment.
“Eclectic” intervention like that provided to children in the comparison group in
the Eikeseth et al. (2002) study is widely available to children with autism enrolled in
public early intervention and special education programs. Yet little evidence about the
efficacy of that approach has appeared in the research literature to date. The study
described here was a prospective analysis of the effects of three different early
intervention approaches on young children with autism spectrum disorders. Interim (14-
month) outcomes for children who participated in an intensive behavior analytic
treatment program were compared with those of children who received intensive
“eclectic” intervention in classrooms designed exclusively for children with autism and
children in non-intensive, generic early intervention programs.

Method

Participants
Referral and selection. The participants were 61 children diagnosed with autistic
disorder or pervasive developmental disorder - not otherwise specified (PDD-NOS).
Potential participants were referred by non-profit agencies (“regional centers”) under contract
with the State of California Department of Developmental Services to provide case
management for individuals with developmental disabilities. Referred children were screened
for the following eligibility criteria: (a) diagnosis of autistic disorder or PDD-NOS according
to DSM-IV criteria by qualified independent examiners before the child was 48 months of
age; (b) entry into an intervention program before 48 months of age; (c) English as the
primary language spoken in the child’s home; (d) no significant medical condition other than
autistic disorder or PDD-NOS; and (e) no prior treatment of more than 100 hours.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 6

Under an existing collaborative funding agreement between public schools and

regional centers in the region where the study was conducted, individual education plan (IEP)
and individual family service plan (IFSP) teams for young children with autism spectrum
disorders routinely consider a range of educational options. These include but are not limited
to: early intensive behavior analytic treatment (IBT) from non-public agencies; autism
educational programming (AP) delivered in special education classrooms designed
specifically for children with autism spectrum disorders; and generic educational
programming (GP) for children with various diagnoses. Auxiliary services, such as
occupational therapy and speech and language therapy, can also be considered and
recommended by the IEP or IFSP teams. Although educational placement decisions
regarding participants in this study were made by IEP or IFSP teams, parental preferences
weighed heavily.
Eligibility criteria were met by 37 children who received IBT intervention from a
nonpublic agency and 41 children who were enrolled in AP or GP programs operated by
local school districts and counties from 1996 through 2003. Four children who began in the
IBT group were excluded from analysis because they did not complete 7 months of
intervention. Two of those children were just 2 years old when intervention began. They
acquired some nonverbal skills, but their receptive and expressive language skills did not
improve, and behavioral difficulties increased when the full number of intervention hours
was attempted. This led their IFSP teams to recommend transition to less intensive school
programs.The third child left the IBT group because the child’s parents were not able to
accommodate an intensive intervention program at home, and the fourth child moved out of
the state. Four children who were placed in either AP or GP were excluded because their
parents could not be contacted to arrange followup testing despite repeated attempts (3
children), or because the parent did not allow the child to be tested at followup (1 child).
Nine other children (4 in the IBT group and 5 in the AP and GP groups) were excluded
because more than 18 months elapsed between intake and opportunity for followup. Because
the followup testing did not occur, it was not possible to confirm the treatment group
placement (AP or GP) for those 5 participants. Remaining for analysis were intake and
followup data for 29 children who received IBT, 16 children in AP, and 16 children in GP.
Characteristics. Table 1 summarizes the gender, ethnicity, diagnosis, and parents’
marital status of the participants. The three groups of children were very similar on all of
those characteristics at intake. Although the percentage of children with a given characteristic
varied somewhat from one group to another, none of the differences between group means
was statistically significant.
Table 2 summarizes the mean severity of autism (determined by the number of DSM-
IV criteria for autistic disorder met) and chronological age of the participants in each group,
and the mean educational levels of the participants’ parents. Children in the IBT group were
diagnosed at a younger age than children in the autism program, who in turn were diagnosed
at a younger age than children in the generic program. Children in the IBT group also began
treatment earlier, and had earlier followup testing, than children in the AP and GP groups.
Parents of children in the IBT group averaged 1-2 more years of education than parents of
children in the other two groups. All of those differences were statistically significant, and
were controlled for in subsequent analyses.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 7

Interventions
Participants’ files, including IFSP or IEP documents for the year following
diagnosis, were reviewed to determine services received, educational placement, and
number of hours of intervention per week for each child in the AP and GP groups. For
those groups, classroom and intervention descriptions were obtained through direct
observation of the programs, interviews with classroom and administrative staff of those
programs, and interviews with regional center staff familiar with the programs. The first
two authors, who directed the IBT program, provided information about that intervention.
Intensive behavior analytic treatment (IBT). Children in the IBT group received
intervention in multiple settings including home, school, and the community. Intensive
treatment was defined as 25-30 hours per week of 1:1 intervention for children under 3
years of age and 35-40 hours of 1:1 intervention for children over 3 years of age.
Children had 50-100 learning opportunities per hour presented via discrete trial,
incidental teaching, and other behavior analytic procedures (see Anderson & Romanczyk,
1999; Green, 1996; Hall, 1997). Instruction occurred during formal, structured sessions
as well as less structured situations, such as supervised play dates with typically
developing peers.
Each child’s program comprised individualized goals and objectives derived from
ongoing evaluations employing both standardized tests and direct observational
measurement. Programs similar to those described in several treatment manuals (e.g.,
Maurice, Green, & Luce, 1996; Maurice, Green, & Foxx, 2001) were delivered using a
combination of behavior analytic techniques, including general case programming to
maximize skill generalization and most-to-least prompt and prompt-fading procedures to
minimize errors during skill acquisition. Children were taught to select their own
reinforcers, record their own performances, and sequence their learning activities as
appropriate for each child. Direct observational data on each child’s progress were
reviewed by program supervisors several times each week, and intervention procedures
(e.g., reinforcers, instructions, prompts, pacing of learning opportunities, etc.) were
modified as needed.
Each child’s programming was delivered by a team of 4-5 instructional assistants,
each of whom worked 6-9 hours per week with the child. Instructional assistants were
employed part-time while they attended college. They were trained and supervised by
staff with master’s degrees in psychology or special education and coursework as well as
supervised practical experience in applied behavior analysis with children with autism.
Some supervisors were assisted by staff with Bachelor-level degrees and (typically)
graduate coursework in behavior analysis. Each supervisor was responsible for
programming for 5-9 children and worked under the direction of a Board Certified
Behavior Analyst who was also a licensed psychologist (the first author) and a licensed
speech and language pathologist (the second author). Parents received training in basic
behavior analytic strategies, assisted in the collection of maintenance and generalization
data, implemented programs with their children outside of regularly scheduled
intervention hours, and met with agency staff 1-2 times a month. No additional services,
such as occupational therapy or individual or small group speech therapy, were provided
to the children in the IBT group. Although efforts were made to ensure treatment
integrity (e.g., through frequent direct observation and videotaping of staff implementing
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 8

procedures with children, and frequent feedback from supervisors), no formal

measurement of treatment integrity was undertaken.
Autism educational programming (AP). Children in the AP comparison group
were enrolled in public school classrooms designed for children with autism. The
staff:child ratio was 1:1 or 1:2, depending on individual needs and the structure of the
particular program in which each child was enrolled. A credentialed special education
teacher supervised the work of 4-8 paraprofessional aides in each classroom. Staff
provided 25-30 hours of intervention each week, utilizing a variety of methods designed
primarily for children with autism spectrum disorders. They included discrete trial
training, Picture Exchange Communication System (PECS; Bondy & Frost, 1994),
sensory integration therapy, and activities drawn from the TEACCH model. In addition,
other activities common to preschool programs for typically developing children (e.g.,
“circle time” and music activities) were incorporated into daily routines. Classroom
teachers received consultation from staff with 1-2 years of graduate-level coursework in
behavior analysis but who had not yet completed masters’ degrees. Seven of the 16
children in the autism programs also received individual or small group speech therapy
sessions 1-2 times weekly from a certified speech and language pathologist. No measures
of the integrity of this treatment were available.
Generic educational programming (GP). Children in the GP comparison group
were enrolled in local community special education classrooms identified as early
intervention or communicatively handicapped preschool programs. Those programs
served children with a variety of disabilities, and provided an average of 15 hours of
intervention per week, with a 1:6 adult: child ratio. Each classroom was staffed by
credentialed special education teachers or certified speech and language pathologists who
supervised 1-2 paraprofessional aides. Educational activities were described as
“developmentally appropriate,” with an emphasis on exposure to language, play
activities, and a variety of sensory experiences. Thirteen of the 16 children in this group
also received individual or small group speech and language therapy sessions 1-2 times
weekly from a certified speech and language pathologist. No operational definitions of
this intervention were available, nor were measures of treatment integrity.

Dependent Measures
Assessments were conducted by experienced psychologists and speech and
language pathologists who were independent contractors with the local regional center
and who were not involved in delivering treatment to any of the children in the study. A
test battery, developed by regional center staff to measure intellectual, nonverbal problem
solving, language, and adaptive skills was administered annually to all children with
autism spectrum disorders below 6 years of age in the region. Assessments were
conducted in the child’s home, in a clinician’s office, or at the regional center as agreed
to by the assessors and the parents. Intake testing of participants in this study was
conducted within 2 months of treatment entry. Followup testing occurred an average of
14 months after treatment entry. The previously described educational placement data
were gathered concurrently with followup testing. Some children did not complete the
entire test battery at intake or followup. Table 3 summarizes the numbers of children in
each group for whom scores were available at intake and followup for each dependent
measure.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 9

Cognitive skills. The standard administration of the Bayley Scales of Infant

Development-Revised (BSID-R; Bayley, 1993) provided intake measures of intellectual
functioning for 42 participants. The BSID-R is widely used with both typical children and
children with autism in the age group encompassed by this study (standard scores are
available for ages 2-42 months). The BSID-R yields a mental development index (MDI),
which was used as the standard score for intellectual functioning in our analyses. Other
tests of cognitive skills administered at intake were the Wechsler Primary Preschool
Scales of Intelligence-Revised (WPPSI-R; Wechsler, 1989; 10 children), Developmental
Profile-II (DP-II; Alpern, Boll, & Shearer, 1986; 3 children), and the Stanford-Binet
Intelligence Scale, Fourth Edition (S-B; Thorndike, Hagen, & Sattler, 1986; 2 children).
In addition, the Differential Abilities Scale (DAS; Elliott, 1990), Developmental
Assessment of Young Children (DAYC; Voress & Maddox, 1998) and the
Psychoeducational Profile Revised (PEP-R; Schopler, Reichler, Bashford, Lansing, &
Marcus, 1990) were administered to one child each. One child did not receive a test of
intellectual functioning at intake.
The test used at followup varied with the chronological age of the child. Most
children received the WPPSI-R (47 children). For those children the full-scale IQ score
represented the standard score for cognitive functioning in our analyses. Other tests
administered at followup were the BSID-II (4 children), Stanford-Binet (3 children), and
the DAS (2 children). Three children did not receive tests of intellectual functioning at
followup, and 2 others (1 in the AP group and 1 in the IBT group) were deemed
“untestable” by the evaluators when the WPPSI-R was attempted.
Nonverbal skills. The Merrill-Palmer Scale of Mental Tests (Stutsman, 1948) was
administered to 48 children at intake and 54 children at followup. It assesses visual-
spatial skills and has norms available for ages 18-78 months. The instrument is widely
used due to its appealing materials, “hands-on” nature, and minimal attention demand
characteristics. There is also evidence that it has predictive validity with nonverbal young
children (Lord & Schopler, 1989). Test scores are expressed as standard scores and age
equivalents. Nonverbal skills for one child were assessed by the Stanford-Binet
Performance Test. One child received the Leiter International Performance Scale Revised
(Leiter-R; Roid & Miller, 1997) at followup.
Receptive and expressive language. The Reynell Developmental Language Scales
(Reynell & Gruber, 1990) were used to assess receptive and expressive language
development for 46 children at intake and 47 children at followup. This instrument expresses
scores in developmental ages, standard scores, and percentiles relative to a normative group.
It is also widely used to test young children with autism due to its colorful materials, reliance
on motor responses, and minimal attention demand characteristics. Other tests of language
functioning administered at intake were the Rossetti Infant-Toddler Language Scale
(Rossetti, 1990; 5 children), the Receptive - Expressive Emergent Language Scales-Revised
(REEL-2; Bzoch & League, 1991; 3 children) and the Preschool Language Scale-3 (PLS-3;
Zimmerman, Steiner, & Pond, 1992; 3 children). The Infant -Toddler Developmental
Assessment (Provence, Eriksen, Vater, & Palmeri, 1985), the Peabody Picture Vocabulary
Test –3rd edition (PPVT-III; Dunn & Dunn, 1997) in conjunction with the Expressive
Vocabulary Test (EVT; Williams, 1997), and the language scale of the DP-II were also used
to assess language development at intake (1 child each). Other tests administered at followup
were the Sequenced Inventory of Communication Development-Revised Edition (SICD-R;
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 10

Hedrick, Prather, & Tobin, 1984; 3 children), the PLS, and the PPVT-III in conjunction with
the EVT (2 children each). One child was assessed at followup with both the Expressive
One-Word PictureVocabulary Test (EOWPVT; Brownell, 2000a) and the Receptive One-
Word PictureVocabulary Test (ROWPVT; Brownell, 2000b). One child did not receive a
language functioning test at intake, and 6 children did not receive followup language tests.
Adaptive skills. The Vineland Adaptive Behavior Scales: Interview Edition (VABS;
Sparrow, Balla & Cicchetti, 1984) was administered both at intake (54 children) and
followup (56 children) to the parents or primary caregivers of all participants in the study.
The VABS is the most widely used assessment of adaptive skills and is viewed as a valid
measure of overall adjustment in children with autism spectrum disorders (Klin, Carter, &
Sparrow, 1997; Newsom & Hovanitz, 1997). The VABS yields a composite score expressed
as a standard score and four domain scores (communication, daily living, socialization, and
motor skills), expressed either as standard scores or age equivalents. All were used in our
analyses. Other intake tests of adaptive skills were the personal adjustment or self-help
subscales of the Denver Developmental Screening Test II (Frankenburg, Dodds, Archer,
Shapiro, & Bresnick, 1992; 3 children), the DP-II (Alpern et al., 1986; 1 child), and the
Rockford Infant Development Evaluation Scales (RIDES; Project RHISE, 1979; 1 child
each). Two children did not receive tests of adaptive skills at intake, and 6 children did not
receive followup tests of adaptive skills.

Data Analyses
In our statistical analyses we were primarily interested in comparing the test scores of
children in the IBT group with those of children in the AP and GP groups, to determine the
efficacy of IBT relative to the other forms of treatment. A secondary comparison of interest
(statistically orthogonal to the comparison of primary interest) was between the test scores of
children in the AP group and those of children in the GP group, to determine if the effects of
those two forms of treatment differed from each other. Several statistical approaches are
available to make these comparisons, including t-tests and planned contrasts. We sought to
avoid approaches (such as t-tests) that evaluate data at the group level, because they cannot
readily accommodate individual differences. This was a concern in our study, because the
average age at diagnosis differed between the three groups of children, and because parents
of children in the IBT group were more educated, on average, than parents of the children in
the other two groups. Accordingly, we used multiple regression to compare the three groups
of children while controlling for individual differences in age at diagnosis and parental
education.
For the multiple regression analyses, we created a variable that was used to compare
the children in the IBT group with the children in the AP and GP groups by assigning a
numeric code of 1 to children in the IBT group and a numeric code of –1 to children in the
other groups. Similarly, we created a variable that was used to compare the children in the
AP group with the children in the GP group by assigning a numeric code of 0 to children in
the IBT group, a numeric code of –1 to children in the AP group, and a numeric code of 1 to
children in the GP group. All analyses included both of these variables. All analyses also
included age at diagnosis and parents’ mean level of education, to control for the potential
influence of those two variables. The parents’ mean level of education was used instead of
entering maternal and paternal education levels as separate variables, because the maternal
and paternal years of education were highly correlated (r = .52).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 11

The children in the IBT group were younger, on average, than the children in the
other two groups at both intake and followup testing. No specific correction was made for
age at testing, however. Such a correction could only have affected analyses of age
equivalents; standard scores and learning rates already correct for age at testing. Furthermore,
by controlling for age at diagnosis we essentially controlled for age at testing as well,
because age at diagnosis was highly correlated with age at intake testing (r = .78) and age at
follow-up testing (r = .79).
Learning rates prior to intake were calculated for nonverbal, receptive language,
expressive language, communication, daily living, social, and motor skills by dividing the
age equivalent at intake by the child’s chronological age in months.Nonverbal learning
rates were based on the age equivalent scores derived from the Merrill-Palmer. Receptive
and expressive language learning rates were calculated using age equivalents from
standardized language assessments. Communication, daily living, social, and motor
learning rates were derived from age equivalent scores on the VABS. Learning rates
during the intervention period were calculated by subtracting the intake age equivalent
score on the measure in question from the age equivalent score at followup, and then
dividing by the interval between intake and followup testing.

Results

Intake
At intake there was clear evidence of developmental delay in all three groups of
children. For most skill domains, the mean standard scores for all three groups were
substantially below 100, and the mean learning rates were well below the normal rate of one
year of development per year of age (see Table 4). As might be expected, delays were most
prominent in receptive and expressive language skills, with mean standard scores in all three
groups close to 50, and mean learning rates of about 0.5 age equivalents per year (i.e., half
the normal learning rate).
The mean scores of all three groups of children on all measures were similar at intake.
The only difference that reached statistical significance was in the nonverbal skills domain,
where the GP group had a significantly higher mean age equivalent score than the AP group.

Followup
At followup, there were no statistically significant differences between the mean
scores of children in the AP and GP groups (see Table 5). In contrast, the IBT group had
higher mean scores in all domains than the other two groups combined. Those differences
were statistically significant. The only exception to this general finding was in the motor
skills domain, which yielded no statistically significant group differences when results were
expressed as learning rates. The IBT group had mean standard scores in the normal range on
cognitive, nonverbal, communication, and motor skills, whereas the only mean score in the
normal range for the AP and GP groups was in motor skills (which were not substantially
delayed at intake). Differential treatment effects were also reflected in changes exhibited by
individual children within the three groups. For example, the cognitive (IQ) scores of 13
children in the IBT group increased from one standard deviation or more below average (i.e.,
IQ of 85 or lower) at intake to within one standard deviation of average or above (i.e., IQ of
86 or higher) at followup. Three children in that group had IQ scores in or near the normal
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 12

range at intake (84, 89, and 97); at followup their IQ scores had increased to 122, 114, and
102, respectively. In the AP group, no children had IQ scores in the normal range at intake; at
followup, the IQ scores of two children had moved into the normal range. Three children in
the GP group had IQ scores that moved from one or more standard deviations below average
at intake to within the normal range at followup; however, the two children in that group
whose IQ scores were in the normal range at intake actually had lower IQ scores at followup
(from 91 to 77 and 89 to 85).
Table 5 also shows that IBT produced normal or above-normal mean learning rates in
all skill domains, although the learning rate for motor skills was near normal for this group as
well as the other two groups of children before intervention. In contrast, only nonverbal skills
were acquired at close to normal rates by children in the other two treatment groups during
the intervention period (means = 0.87 and 0.90, respectively). Differential treatment effects
were most evident when rates of acquisition of language skills were compared. Inspection of
Figures 1 and 2 reveals that those differences were not restricted to just a few children. Prior
to intake, children in all three groups exhibited similar, below-normal rates of learning
receptive language skills, although two children in the IBT group were acquiring receptive
language skills at a normal rate prior to intervention (Figure 1). At followup, all but 8
children in the IBT group were acquiring receptive language skills at a normal rate, with
several achieving at above-normal rates and two others at near-normal rates. In contrast,
learning rates at followup remained below normal for the large majority of children in the AP
and GP groups. A small number of children in all three groups, however, appeared to have
lower learning rates in this domain at followup than at intake.
Figure 2 shows similar patterns for expressive language skills, At intake, all children
in the IBT group had expressive language learning rates that were below normal; at followup,
all but 9 of those children were acquiring those skills at normal rates, with rates accelerated
to substantially above normal for several children. Two additional children in this group had
near-normal learning rates at followup. All children in the AP and GP groups also had below-
normal rates of acquisition of expressive language skills at intake. At followup, although 1-2
children in each group exhibited normal or above-normal learning rates, the rate of
acquisition of expressive language skills actually declined over the course of intervention for
several children in both groups. Some of the factors that contributed to these between-group
and individual differences will be explored in a subsequent paper.
Since the mean scores for all three groups of children on all dependent measures were
similar at intake, the analysis of change scores yielded results that were similar to those that
emerged from analyzing the followup scores (see Table 6). Some interesting additional
information was revealed by this analysis. Children in the IBT group exhibited statistically
significantly larger mean treatment gains in all domains than children in the AP and GP
groups combined, with the possible exception of the motor skills domain (which was
significant only when standard scores were used). Indeed, the IBT group had mean gains in
standard scores in all skill domains, ranging from 1.38 points in motor skills (which were
already near-normal at intake) to 29.72 points in cognitive skills. The AP group’s change
scores ranged from –5.13 points in the motor skills domain to 8.44 points in cognitive skills.
Mean change scores for this group actually revealed losses in social and motor skills as well
as the VABS composite score, and negligible-to-small gains in the other domains. For the GP
group, mean change scores ranged from –7.43 in daily living skills to 8.94 in cognitive skills,
with losses in receptive language, expressive language, daily living, social, and motor skills
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 13

as well as the VABS composite score. Similar patterns emerged when age equivalents were
used in change score analyses: the IBT group made gains in all domains (range = 13.44 –
20.81 months), gaining more than 14 months developmentally, on average, in nonverbal,
receptive language, expressive language, overall communication, social, and motor skills
over the 14-month intake-to-followup period. Mean age equivalent gains for the AP and GP
groups were much smaller and were less than 14 months in all domains (ranges = 7.53 –
12.63 months and 4.5 – 13.17 months, respectively).

Discussion

Young children with autism or PDD-NOS who received intensive behavior analytic
treatment (IBT) for about 14 months outperformed comparable children who received
“eclectic” intervention services for the same period of time on virtually every followup
measure. In most cases the differences in mean scores were substantial and statistically
significant. Our analyses corrected for the parents’ level of education and for the children’s
ages at diagnosis. No direct correction was made for the age at testing, but children in the
IBT group had the highest mean age equivalents at followup (see Table 5), despite being
younger than the children in the other groups. Thus, our findings cannot be attributed to
differences in age at testing; if anything, they underestimate the effect of IBT on age
equivalents. These results are consistent with those reported by other investigators who found
that providing at least 30 hours of competently delivered, intensive behavior analytic
intervention to preschool-age children with autism produced large improvements in
intellectual functioning, communication skills, and adaptive behavior. We reported gains
measured just 14 months into treatment, so it was not surprising that they were generally
smaller than gains that have been documented after 2-3 years of IBT (e.g., Green et al., 2002;
Lovaas, 1987; Perry et al., 1995; Weiss, 1999). The gains we observed, however, were
generally larger than gains reported by Anderson et al. (1987) for preschool children with
autism who received only 15-25 hours of behavior analytic treatment for one year.
Analyses of learning rates (Table 5, Figures 1 and 2) provided further evidence of the
efficacy of IBT for accelerating rates of skill acquisition. During 14 months of treatment,
children in the IBT group acquired skills in most domains at a rate that matched or exceeded
the normal rate of one year of development per year of age. That was not the case for the
children in the AP and GP groups; with very few exceptions, their learning rates remained
well below normal. If children with autism are to have any chance to close the gap between
their skills and those of their typically developing peers, their developmental trajectories
must be increased sharply while they are young, before the gap widens even further. That is,
their learning rates need to exceed the normal rate for an extended period of time. Of the
early intervention approaches investigated in this study, only IBT had that effect, producing
above-normal mean learning rates in the nonverbal, receptive language, expressive language,
overall communication, and social skill domains. It is important to note, however, that 14
months of accelerated development was not enough for the children in the IBT group to
make up all of the differences between their skills and those of typically developing
preschoolers. Previous research suggests that at least 1-2 additional years of IBT will be
required before some of those children will have the repertoires required to learn effectively
in typical classrooms without ongoing specialized intervention; some will require more than
that, and some will likely not reach that point even with additional IBT (see Green, 1996;
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 14

Smith, 1999). Projections based on the developmental trajectories produced by IBT in our
study suggest that most children will continue to make progress toward catching up with their
typically developing peers if they continue receiving competently delivered IBT.
Our findings also shed some empirical light on the relation between the type and
intensity of early intervention and benefits for children with autism. “Eclectic” treatment (a
combination of TEACCH, sensory integration therapy, and some applied behavior analysis
methods) did not prove very effective for our AP comparison group, even though it was
provided intensively (i.e., for 30 hours per week with adult:child ratios of 1:1 or 1:2) in
classrooms specifically designed for children with autism by staff with considerable training
and experience with the population. Mean change scores in all skill domains were
substantially lower for the AP group than for their counterparts who received IBT, in fact
reflecting losses rather than gains in some areas over 14 months of treatment (Table 6). These
findings are consistent with those reported by Eikeseth et al. (2002) for a group of children
with autism aged 4-7 years who received similarly intensive “eclectic” treatment in special
education classrooms for one year. Thus the popular notion that virtually any intervention
can produce meaningful benefits for children with autism if it is provided intensively has not
been confirmed by two controlled studies that addressed that hypothesis. Instead, IBT
produced substantially larger improvements than intensive “eclectic” treatment in both
studies. The nonintensive “eclectic” treatment experienced by our GP group (15 hours per
week of “developmentally appropriate” activities and sensory experiences provided in a 1:6
adult:child ratio) was not just ineffective; it produced negative mean change scores in
multiple skill domains. In short, the effect of “eclectic” treatment on both the AP and GP
groups was to flatten or decrease rather than increase the slopes of the developmental
trajectories of most children. Based on these findings, we would project that those children
will lose more ground to their typically developing peers the longer they remain in such
intervention programs.
The ineffectiveness of the “eclectic” early intervention provided to children in the AP
and GP groups in this study should not be surprising. “Eclectic” intervention necessarily
involves multiple transitions per day from one activity or “therapy” to another, and a good
deal of variability in the way intervention is provided by the various adults involved.
Children with autism often do not respond well to changes in routines, have substantial
attentional difficulties, and learn best when instruction is consistent. It does not stand to
reason that typical “eclectic” programming provided in a group format is likely to produce
meaningful benefits for children with those characteristics. Nor does it follow logically that
combining several “therapies” or methods for which there is limited scientific evidence of
effectiveness (such as TEACCH, developmental models, and sensory integration therapy; see
Arendt, MacLean, & Baumeister, 1988; Dawson & Watling, 2000; Smith, 1999) is likely to
be beneficial for young children with autism. What is surprising is how few scientific studies
heretofore have evaluated the “eclectic” approach, and how many prominent individuals and
organizations in the autism community and the education establishment endorse and promote
it.
One interesting observation that was common to all three treatment groups in this
study was a change in the distributions of language learning rates from pre-treatment to
followup (see Figures 1 and 2). For all three groups, the spreads of the distributions were
considerably greater at followup than before treatment, and some modes shifted as well.
Those changes may have been due in part to sampling errors, which could have been
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 15

magnified in the followup data because the followup learning rates were based on a shorter
time period than were the pre-treatment learning rates (14 months versus 34 months, on
average). It is likely that sampling errors affected all data sets equally, however, so the
relative between-group differences in learning rate distributions likely reflect differential
treatment effectiveness. Here again, the effects of IBT appeared to differ substantially from
the effects of the other two interventions. Figure 1 shows that for the IBT group, the mode of
the distribution of learning rates for receptive language skills moved from well below normal
before treatment to above normal after 14 months of treatment, with many more children
achieving normal rates at followup than prior to intervention. The mode of the distribution of
receptive language learning rates for the intensive “eclectic” intervention (AP) group was
slightly higher but still well below normal at followup, with three children acquiring
receptive language skills at normal rates after 14 months. For the GP group, the mode of the
distribution of receptive language learning rates was lower at followup than pre-treatment,
although two children in that group were acquiring receptive language skills at normal rates
at followup. With regard to rates of acquiring expressive language skills (Figure 2), the
distribution spread markedly with IBT, with a number of children in that group exhibiting
learning rates that were well above normal at followup. Prior to intervention, the modal
learning rate for the IBT group was well below normal; at followup a bimodal distribution
was observed, with one mode substantially above normal and the other just below normal.
Intensive “eclectic” treatment also appeared to produce greater spread in the distribution of
expressive language learning rates for the AP group, but only a slight upward shift in the
mode. The nonintensive “eclectic” intervention (GP) group showed a slightly increased
spread in the distribution of learning rates for expressive communication skills at followup,
but the mode shifted down rather than up.
Several limitations to this study constrain the interpretation of our results. First,
assignment to treatment groups was parent-determined rather than random; however, the
three groups were very similar on key dependent measures before treatment began, which is
the main purpose of random assignment (cf. Baer, 1993; Kasari, 2002). Thus, differences in
outcomes across the three groups were likely due to the treatments rather than to any
selection bias or pre-treatment differences among the groups. Second, the examiners who
conducted the assessments were not blind as to the children’s group assignments at followup
testing. They were, however, independent of the investigators as well as all three intervention
programs. It could be argued that some of the examiners were biased toward IBT, which led
them to overestimate the followup status of children in that group. Since there were a large
number of examiners, however, it is just as likely that some of them were biased against IBT
and toward the other interventions. Third, results were analyzed only in terms of
performances on standardized, norm-referenced assessments conducted in formal testing
situations, rather than the repeated direct observational measurement of behavior in situ that
characterizes applied behavior analysis. Additionally, the analyses compared group mean
scores statistically. Group mean scores may not accurately represent the actual performance
of any individual in the group, and between-groups statistical comparisons of mean scores
cannot reveal clinically significant changes in individual behavior over time (Johnston &
Pennypacker, 1993). Nonetheless, standardized instruments like IQ tests and adaptive
behavior scales are widely used in autism research, and scores on such tests have been shown
to correlate reasonably well with overall adjustment for individuals with autism (e.g., Klin et
al, 1997). Further, between-groups comparisons are helpful for answering actuarial questions,
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 16

such as the relative efficacy of interventions for groups of children with autism. Finally,
treatment integrity was not measured in this study. The behavior analytic treatment was
directed by individuals with documented training and credentials in applied behavior
analysis, and incorporated techniques that have been operationally defined and tested in
many previous studies (see Green, 1996, 2001; Matson et al., 1996). Staff in that program
were trained and supervised closely, but it cannot be assumed that they implemented
treatment procedures with fidelity and consistency throughout the study. Even fewer
assumptions can be made about the other interventions. Indeed, measuring the integrity of
those interventions would likely prove challenging, because many of the techniques
employed have not been operationally defined or evaluated, and the skills required to
implement them have not been well-specified.
As noted previously, we plan to conduct further analyses of child, family, and
treatment variables that were correlated with the differential outcomes reported here.
Additional research on the importance of such variables is needed to inform decision-making
by families and policymakers, and to aid in the development of new or modified
interventions for children with autism spectrum disorders who do not respond to IBT. Studies
that further investigate the short- and long-term effects of “eclectic” intervention are also
needed, given the widespread popularity and availability of that approach for children with
autism spectrum disorders. The same can be said of early intervention that is based primarily
or exclusively on models that have not yet been subjected to thorough scientific evaluations,
such as TEACCH, “developmentally appropriate” programming, “floor time,” Relationship
Development Intervention, and sensorimotor techniques.

References

Alpern, G., Boll, T., & Shearer, M. (1986). Developmental Profile II (DP-II). Los
Angeles: Western Psychological Services.
Anderson, S. R., Avery, D. L., DiPietro, E. K., Edwards, G. L., & Christian, W. P.
(1987). Intensive home-based early intervention with autistic children. Education
and Treatment of Children, 10, 352-366.
Anderson, S. R., & Romanczyk, R. G. (1999). Early intervention for young children with
autism: Continuum-based behavioral models. Journal of the Association for
Persons with Severe Handicaps, 24, 162-173.
Arendt, R. E., MacLean, W. E., Jr., & Baumeister, A. A. (1988). Critique of sensory
integration therapy and its application in mental retardation. American Journal on
Mental Retardation, 92, 401-411.
Baer, D. M. (1993). Quasi-random assignment can be as convincing as random
assignment. American Journal on Mental Retardation, 97 (4), 373-375.
Bayley, N. (1993). Bayley Scales of Infant Development, Second Edition. San Antonio,
TX: The Psychological Corporation.
Birnbrauer, J. S., & Leach, D. J. (1993). The Murdoch Early Intervention Program after 2
years. Behaviour Change, 10 (2), 63-74.
Bondy, A. S., & Frost, L. A. (1994). The Picture Exchange Communication System:
Training Manual. Cherry Hill, NJ: Pyramid.
Brownell, R. (Ed). (2000a). Expressive One-Word Picture Vocabulary Test-2000.
(EOWPVT-2000). Novato, CA: Academic Therapy Publications.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 17

Brownell, R. (Ed). (2000b). Receptive One-Word Picture Vocabulary Test-2000

(ROWPVT-2000). Novato, CA: Academic Therapy Publications.
Bzoch, K., & League, R. (1991) Receptive Expressive Emergent Language Scales-
Revised. (REEL-2) Second Edition. Austin, TX: PRO-ED.
California Department of Developmental Services. (2002). Department of Developmental
Services Fact Book – 5th edition. Sacramento, CA: California Department of
Developmental Services. http://www.dds.ca.gov/factsstats/factbook.cfm#pdf
California Department of Developmental Services. (2003a). 2003 DDS Autism Report.
Sacramento, CA: California Department of Developmental Services.
http://www.dds.ca.gov/autism/autism_main.cfm
California Department of Developmental Services. (2003b). Department of
Developmental Services Fact Book – 6th edition. Sacramento, CA: California
Department of Developmental Services.
http://www.dds.ca.gov/factsstats/factbook.cfm#pdf
Campbell, F. A., Pungello, E. P., Miller-Johnson, S., Burchinal, M. & Ramey, C. T.
(2001). The development of cognitive and academic abilities: Growth curves from
an early childhood educational experiment. Developmental Psychology, 37 (2),
231-242.
Campbell, F. A., Ramey, C. T., Pungello, E. P., Sparling, J., & Miller-Johnson, S. (2002).
Early childhood education: Young adult outcomes from the Abecedarian Project. Applied
Developmental Science, 6, 42-57.
Dawson, G., & Fischer, K. W. (Eds.). (1994). Human behavior and the developing brain.
New York: Guilford.
Dawson, G., & Watling, R. (2000). Interventions to facilitate auditory, visual, and motor
integration in autism: A review of the evidence. Journal of Autism and
Developmental Disorders, 30, 415-425.
Dunn, L. M., & Dunn, L. M. (1997). Peabody Picture Vocabulary Test-Third Edition.
Circle Pines, MN: American Guidance Service Publishing.
Eikeseth, S., Smith, T., Jahr, E., & Eldevik, S. (2002). Intensive behavioral treatment at
school for 4-7-year-old children with autism: A 1-year comparison controlled
study. Behavior Modification, 2002, 49-68.
Elliott, C. (1990). Differential Ability Scales. San Antonio, TX: The Psychological
Corporation.
Fenske, E. C., Zalenski, S., Krantz, P. J., & McClannahan, L. E. (1985). Age at
intervention and treatment outcome for autistic children in a comprehensive
intervention program. Analysis and Intervention in Developmental Disabilities, 5,
49-58.
Ferster, C. B. & DeMyer, M. K. (1961). The development of performances in autistic
children in an automatically controlled environment. Journal of Chronic
Diseases, 13, 312- 345.
Fombonne, E. (2001). Is there an epidemic of autism? Pediatrics, 107, 411-413.
Fombonne, E. (2003). The prevalence of autism. Journal of the American Medical
Association, 289, 87-89.
Frankenburg, W. K., Dodds, J., Archer, P., Shapiro, H., & Bresnick, B. (1992). The
Denver II: A major revision and restandardization of the Denver Developmental
Screening Test. Pediatrics, 89, 91-97.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 18

Green, G. (1996). Early behavioral intervention for autism: What does research tell us? In
C. Maurice (Ed.), G. Green, & S. Luce (Co-Eds.), Behavioral intervention for
young children with autism: A manual for parents and professionals (pp. 29-44).
Austin, TX: PRO-ED.
Green, G., Brennan, L. C., & Fein, D. (2002). Intensive behavioral treatment for a toddler
at high risk for autism. Behavior Modification, 26, 69-102.
Green, G. (2001). Behavior analytic instruction for learners with autism: Advances in
stimulus control technology. Focus on Autism and Other Developmental
Disabilities, 16, 72-85.
Guralnick, M. J. (1998). Effectiveness of early intervention for vulnerable children: A
developmental perspective. American Journal on Mental Retardation, 102(4),
319-345.
Hall, L. J. (1997). Effective behavioural strategies for the defining characteristics of
autism. Behaviour Change, 14, 139-154.
Hannigan, J. H., & Berman, R. F. (2000). Amelioration of fetal alcohol-related
neurodevelopmental disorders in rats: exploring pharmacological and
environmental treatments. Neurotoxicology and Teratology, 22, 103-111.
Harris, S. L., Handleman, J. S., Gordon, R., Kristoff, B., & Fuentes, F. (1991). Changes
in cognitive and language functioning of preschool children with autism. Journal
of Autism and Developmental Disorders, 21(3), 281-290.
Hedrick, D., Prather, E., & Tobin, A. (1984). Sequenced Inventory of Communication
Development-Revised Edition. Seattle, WA: University of Washington Press.
Jacobson, J. W., Mulick, J. A., & Green, G. (1998). Cost-benefit estimates for early
intensive behavioral intervention for young children with autism: General models
and single state case. Behavioral Interventions, 13, 201-226.
Johnston, J. M., & Pennypacker, H. S. (1993). Strategies and tactics of behavioral
research (2nd Ed.). Hillsdale, NJ: Erlbaum.
Kasari, C. (2002). Assessing change in early intervention programs for children with
autism. Journal of Autism and Developmental Disorders, 32 (5), 447-461.
Klin, A., Carter, A., & Sparrow, S.S.(1997). Psychological assessment. In D. J. Cohen &
F. R. Volkmar (Eds.), Handbook of autism and pervasive developmental disorders
(2nd ed., pp. 418-427). New York: Wiley.
Lord, C., & Schopler, E. (1989). The role of age at assessment, developmental level, and
test in the stability of intelligence scores in young autistic children. Journal of
Autism and Developmental Disorders, 19, 483-499.
Loupe, P. S., Schroeder, S. R., & Tessel, R. (1995). FR discrimination training effects in
SHR and microencephalic rats. Pharmacology, Biochemistry and Behavior, 51,
869-876.
Lovaas, O. I. (1987). Behavioral treatment and normal educational and intellectual
functioning in young autistic children. Journal of Consulting and Clinical
Psychology, 55, 3-9.
Matson, J., Benavidez, D., Compton, L., Paclawskyj, J., & Baglio, C. (1996). Behavioral
treatment of autistic persons: A review of research from 1980 to the present.
Research in Developmental Disabilities, 17, 433-465.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 19

Maurice, C., (Ed.), Green, G., & Luce, S. (Co-Eds.) (1996). Behavioral intervention for
young children with autism: A manual for parents and professionals. Austin, TX:
PRO-ED.
Maurice, C., Green, G., & Foxx, R. M. (Eds.) (2001). Making a difference: Behavioral
intervention for autism. Austin, TX: PRO-ED.
McEachin, J. J., Smith, T., & Lovaas, O. I. (1993). Long-term outcome for children with
autism who received early intensive behavioral treatment. American Journal on
Mental Retardation, 97, 359-372.
Newsom, C., & Hovanitz, C. A. (1997). Autistic disorder. In E. J. Mash & L. G. Terdal
(Eds.), Assessment of childhood disorders (3rd ed.); Assessment of childhood
disorders (3rd ed.). (pp. 408-452). New York, NY, US: Guilford Press.
Perry, R., Cohen, I., & DeCarlo, R. (1995). Case study: Deterioration, autism, and
recovery in two siblings. Journal of the American Academy of Child and
Adolescent Psychiatry. 34, 232 - 237.
Provence, S., Eriksen, J., Vater, S., & Palmeri, S. (1985). Infant Toddler Developmental
Assessment. Chicago, IL: The Riverside Publishing Co.
Project RHISE (1979). Rockford Infant Development Evaluation Scales. (1979).
Bensenville, IL: Scholastic Testing Services, Inc.
Ramey, C. T. & Ramey, S. L. (1998). Early intervention and early experience. American
Psychologist, February, 109-120.
Ramey, S. L., & Ramey, C. T. (1999). Early experience and early intervention for
children at risk for developmental delay and mental retardation. Mental
Retardation and Developmental Disabilities Research Reviews, 5, 1-10. New
York: John Wiley and Sons.
Reynell, J. K. & Gruber, G. P. (1990). Reynell Developmental Language Scales. Los
Angeles, California: Western Psychological Services.
Rogers, S. J. (1998). Empirically supported comprehensive treatments for young children
with autism. Journal of Clinical Child Psychology, 27 (2), 167-178.
Rogers, S. J., Herbison, J., Lewis, H., Pantone, J., & Reiss, K. (1986). An approach for
enhancing the symbolic, communicative, and interpersonal functioniong of young
children with autism and severe emotional handicaps. Journal of the Division for
Early Childhood, 10, 135-148.
Rogers, S. J. & DiLalla, D.L. (1991). A comparative study of the effects of a
developmentally based instructional model on young children with autism an
dyoung children with other disorders of behavior and development. Topics in
Early Childhood Special Education, 11 (2), 29-47.
Roid, G. & Miller, L. (1997). Leiter International Performance Scale-
Revised. Wood Dale, Ill.: Stoelting.
Rossetti, L. (1990). The Rossetti Infant-Toddler Language Scale: a measure of
communication and interaction. East Moline, IL: Linguisystems.
Schopler, E. (1997). Implemetation of TEACCH philosophy. In D. J. Cohen & F. R.
Volkmar (Eds.), Handbook of autism and pervasive developmental disorders (2nd
ed., pp. 767-795). New York: Wiley & Sons.
Schopler, E., Reichler, R.J., Bashford, A., Lansing, M. D., & Marcus, L. M. (1990). The
Psychoeducational Profile Revised (PEP-R). Austin, Texas: PRO-ED.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 20

Shore, R. (1997). Rethinking the brain: New insights into early development. New York:
Families and Work Institute.
Smith, T. (1999). Outcome of early intervention for children with autism. Clinical
Psychology: Science and Practice, 6, 33-49.
Smith, T., Groen, A. D., & Wynne, J. W. (2000). Randomized trial of intensive early
intervention for children with pervasive developmental disorder. American
Journal on Mental Retardation, 105 (4), 269-285.
Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (1984). The Vineland Adaptive Behavior
Scales (Interview Edition), Circle Pines, MN: American Guidance Service.
Stodgell, C. J., Schroeder, S. R., & Tessel, R. E. (1996). FR discrimination training
reverses 6-hydrosxydopamine-induced striatal dopamine depletion in a rat model
of Lesch-Nyhan syndrome. Brain Research, 713, 246-252.
Stutsman, R. (1948). Merrill Palmer Scale of Mental Tests. Wood Dale, Ill.: Stoelting.
Tessel, R. E., Schroeder, S. R., Loupe, P. S., &Stodgell, C. J. (1995). Reversal of 6HD-
induced neonatal brain catecholamine depletion after operant training.
Pharmacology, Biochemistry & Behavior, 51, 861-867.
Thorndike, R. L., Hagen, E. P., & Sattler, J. M. (1986). The Stanford-Binet Intelligence
Scale: Fourth Edition. Chicago: Riverside.
VanKeuren, K. R., Stodgell, C. J., Schroeder, S. R., & Tessel R. E. (1998). Fixed-ratio
discrimination training as replacement therapy in Parkinson’s disease: Studies in a
6-hydroxydopamine-treated rat model. Brain Research, 780, 56-66.
Voress, J. K., & Maddox, T. (1998). Developmental Assessment of Young Children.
Austin: TX: PRO-ED.
Wechsler, D. (1989). Wechsler Preschool and Primary Scale of Intelligence - Revised.
San Antonio, TX: The Psychological Corporation.
Weiss, M. (1999). Differential rates of skill acquisition and outcomes of early intensive
behavioral intervention for autism. Behavioral Interventions, 14, 3-22.
Williams, K. (1997). Expressive Vocabulary Test. Circle Pines, MN: AGS Publishing.
Wolf, M. M., Risley, T., & Mees, H. (1964). Application of operant conditioning
procedures to the behaviour problems of an autistic child. Behaviour Research
and Therapy, 1, 305–312.
Yeargin-Allsop, M., Rice, C., Karapurkar, T., Doernber, N., Boyle, C., & Murphy, C.
(2003). Prevalence of autism in a US metropolitan area. Journal of the American
Medical Association, 289, 49-55.
Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (1992). The Preschool Language Scale-
3 San Antonio, TX: The Psychological Corp.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 21

Table 1
Number of Participants with Each Characteristic
Treatment group
–––––––––––––––––––––––––––––––––––––––––––––––
Characteristic IBT AP GP

Gender

Male 25 (86%) 13 (81%) 16 (100%)

Female 4 (14%) 3 (19%) 0 (0%)

Ethnicity

Both parents Caucasian 21 (72%) 6 (50%) 8 (57%)

One or both parents Hispanic 4 (14%) 3 (25%) 4 (29%)
Other 4 (14%) 3 (25%) 2 (14%)
Unknown 0 4 2

Diagnosis

Autism 24 (83%) 12 (75%) 9 (56%)

PDD-NOS 5 (17%) 4 (25%) 7 (44%)

Parents’ marital status

Married 23 (79%) 12 (80%) 9 (56%)

Not married, divorced, or separated 6 (21%) 3 (20%) 7 (44%)
Unknown 0 1 0
Note. Percentages are within each treatment group, excluding participants with unknown
characteristics.
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 22

Table 2
Mean Severity of Autism, Age (in Months), and Parents’ Education Level
IBT AP GP IBT mean AP mean
–––––––––––– –––––––––––– –––––––––––– minus minus
Measure M SD M SD M SD AP/GP mean GP mean

Severity (no. of DSM-IV criteria) 7.55 1.39 7.27 1.56 7.33 2.02 0.25 -0.06

Age at diagnosis 30.48 5.96 39.31 5.52 34.94 5.18 -6.65** 4.37*
Age at intake
30.86 5.16 37.44 5.68 34.56 6.53 -5.16** 2.84
Age at followup 45.66 6.24 50.69 5.64 49.25 6.81 -4.31* 1.44
Months between intake and followup 14.21 2.24 13.25 2.84 14.75 1.88 0.21 1.50

Mother's years of education 14.10 2.34 13.00 1.83 13.00 1.41 1.10* 0.00
Father's years of education 14.62 2.77 13.13 2.56 13.00 1.81 1.56* 0.13
Parents’ mean years of education 14.36 2.22 13.06 1.82 12.97 1.36 1.35** 0.09

Note. For the IBT group n = 29, except for severity (n = 20). For the AP group n = 16, except for severity (n =
11) and father’s years of education (n = 15). For the GP group n = 16, except for severity (n = 12) and father’s
years of education (n = 15).
* Difference between means is statistically significant (p < .05).
** Difference between means is statistically significant (p < .01).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 23

Table 3
Number of Children for Whom Dependent Measures
were Available at Intake and Followup
Intake / Followup
Measure IBT AP GP

Standard scores

Cognitive 28 / 26 16 / 16 16 / 16
Nonverbal 21 / 24 16 / 16 13 / 15
Receptive 25 / 26 16 / 15 13 / 14
Expressive 25 / 26 16 / 15 13 / 14
Communication 28 / 25 16 / 16 15 / 16
Self-help 28 / 25 16 / 16 14 / 16
Social 28 / 25 16 / 16 14 / 16
Motor 28 / 25 16 / 16 13 / 16
Composite 26 / 25 16 / 16 13 / 16

Age equivalents

Cognitive 25 / 0 11 / 0 10 / 0
Nonverbal 21 / 24 16 / 16 12 / 15
Receptive 29 / 26 16 / 15 15 / 13
Expressive 29 / 26 16 / 15 15 / 13
Communication 29 / 25 16 / 16 15 / 16
Self-help 29 / 25 16 / 16 15 / 16
Social 28 / 25 16 / 16 15 / 16
Motor 28 / 25 16 / 16 14 / 16

Learning rate

Nonverbal 21 / 21 16 / 16 12 / 12
Receptive 29 / 26 16 / 15 15 / 12
Expressive 29 / 26 16 / 15 15 / 12
Communication 29 / 25 16 / 16 15 / 15
Self-help 29 / 25 16 / 16 15 / 15
Social 28 / 24 16 / 16 15 / 15
Motor 28 / 24 16 / 16 14 / 14
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 24

Table 4

Test Scores and Learning Rates at Intake

IBT AP GP IBT mean AP mean
–––––––––––––– –––––––––––––– –––––––––––––– minus minus
Measure M SD M SD M SD AP/GP mean GP mean

Standard scores

Cognitive 58.54 18.15 53.69 13.50 59.88 14.85 1.76 -6.19

Nonverbal 80.14 11.86 67.44 16.69 77.69 12.33 8.11 -10.25
Receptive 52.16 18.44 45.38 14.97 49.00 13.61 5.16 -3.62
Expressive 51.88 12.91 43.88 6.69 48.77 11.61 5.81 -4.89
Communication 66.18 10.02 63.69 9.68 66.20 8.70 1.28 -2.51
Self-help 70.71 10.14 68.06 11.61 73.43 10.39 0.14 -5.37
Social 72.79 11.26 75.50 14.25 75.07 12.09 -2.51 0.43
Motor 95.11 11.70 93.19 10.10 92.08 13.84 2.42 1.11
Composite1 70.46 11.85 69.81 10.48 71.62 10.47 -0.16 -1.81

Age equivalents (months)

Cognitive1 17.04 6.07 17.27 4.71 17.10 3.93 -0.15 0.17

Nonverbal1 24.43 4.37 24.75 6.01 26.83 6.95 -1.21 -2.08*
Receptive 14.57 5.82 16.81 5.36 16.60 5.34 -2.14 0.21
Expressive1 14.76 4.72 16.38 2.99 17.87 5.45 -2.34 -1.49
Communication1 14.90 4.32 16.19 6.44 16.53 5.25 -1.45 -0.34
Self-help1 18.24 3.83 21.44 7.78 21.20 6.67 -3.08 0.24
Social1 16.39 4.89 22.06 10.62 19.60 5.68 -4.48 2.46
Motor1 28.86 5.86 33.56 7.20 32.00 6.25 -3.97 1.56
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 25

Learning rates prior to intake (age equivalents per year)

Nonverbal 0.79 0.14 0.67 0.17 0.78 0.12 0.08 -0.11

Receptive 0.48 0.21 0.45 0.15 0.48 0.12 0.02 -0.03
Expressive 0.49 0.16 0.44 0.06 0.53 0.17 0.01 -0.09
Communication 0.49 0.15 0.43 0.15 0.49 0.15 0.04 -0.06
Self-help 0.61 0.17 0.57 0.16 0.62 0.18 0.01 -0.06
Social 0.54 0.18 0.58 0.23 0.58 0.19 -0.04 0.00
Motor 0.95 0.18 0.90 0.13 0.93 0.18 0.03 -0.04

1
Age at diagnosis is a significant covariate (p < .05).
* Difference is statistically significant, after controlling for age at diagnosis and parents’ level of
education (p < .05).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 26

Table 5
Test Scores and Learning Rates at Followup
IBT AP GP IBT mean AP mean
–––––––––––––– –––––––––––––– –––––––––––––– minus minus
Measure M SD M SD M SD AP/GP mean GP mean

Standard scores

Cognitive 89.88 20.87 62.13 19.63 68.81 15.32 24.42** -6.68

Nonverbal1 101.67 19.14 73.56 24.94 82.53 16.76 23.77** -8.97
Receptive 71.31 22.72 49.93 19.62 49.21 16.08 21.73* 0.72
Expressive1 70.46 22.88 47.67 23.39 46.79 12.81 23.21* 0.88
Communication 85.44 14.73 64.13 14.18 68.69 14.18 19.03** -4.56
Self-help 76.56 11.59 70.00 11.92 65.19 8.84 8.97** 4.81
Social 82.08 11.73 75.00 18.01 70.56 11.77 9.30** 4.44
Motor 98.16 12.01 88.06 13.43 89.50 10.06 9.38* -1.44
Composite 81.32 11.14 69.25 12.91 68.25 9.86 12.57** 1.00

Age equivalents (months)

Nonverbal2 44.54 8.76 37.38 13.14 40.80 9.97 5.51* -3.42

Receptive 32.23 10.04 26.27 11.56 25.38 10.00 6.37* 0.89
Expressive 31.96 12.00 24.00 12.02 23.31 7.36 8.28* 0.69
Communication 36.60 12.23 23.88 11.82 26.13 8.74 11.60** -2.25
Self-help2 31.88 8.74 31.75 9.75 27.81 5.75 2.10* 3.94
Social 32.04 10.23 30.06 16.10 24.81 7.23 4.61* 5.25
Motor2 44.16 8.22 43.00 7.28 42.25 6.58 1.54* 0.75
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 27

Learning rates between intake and followup (age equivalents per year)

Non-verbal 1.44 0.52 0.87 0.74 0.90 0.39 0.56** -0.03

Receptive 1.23 0.56 0.65 0.47 0.48 0.43 0.66** 0.16
Expressive 1.22 0.73 0.49 0.78 0.33 0.45 0.80** 0.16
Communication 1.43 0.72 0.56 0.76 0.69 0.70 0.81** -0.13
Self-help 0.91 0.58 0.74 0.80 0.48 0.49 0.30* 0.26
Social 1.04 0.74 0.60 0.94 0.40 0.67 0.54* 0.20
Motor 0.99 0.45 0.69 0.49 0.83 0.59 0.24 -0.14

1
Parents’ level of education is a significant covariate (p < .05).
2
Age at diagnosis is a significant covariate (p < .05).
* Difference is statistically significant, after controlling for age at diagnosis and parents’ level of
education (p < .05).
** Difference is statistically significant, after controlling for age at diagnosis and parents’ level of
education (p < .01).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 28

Table 6
Changes in Test Scores and Learning Rates
IBT AP GP IBT mean AP mean
–––––––––––––– –––––––––––––– –––––––––––––– minus minus
Measure M SD M SD M SD AP/GP mean GP mean

Standard scores at followup minus standard scores at intake

Cognitive 29.72 16.29 8.44 15.04 8.94 17.95 21.03** -0.50

Non-verbal1 20.57 16.20 6.13 18.70 2.31 11.61 16.16** 3.82
Receptive 20.17 19.46 3.87 12.09 -4.82 14.81 19.97** 8.68
Expressive1 20.08 22.42 3.80 20.66 -4.45 17.25 19.78* 8.25
Communication 17.17 13.94 0.44 12.47 2.20 14.08 15.88** -1.76
Self-help 5.92 13.60 1.94 15.29 -7.43 11.03 8.35* 9.37
Social 8.00 13.54 -0.50 14.41 -4.64 15.16 10.43* 4.14
Motor 1.38 13.90 -5.13 14.47 -1.23 19.13 4.75 -3.89
Composite 10.52 14.73 -0.56 12.04 -2.77 14.01 12.07** 2.21

Age equivalents at followup minus age equivalents at intake (in months)

Non-verbal 20.81 7.20 12.63 11.20 13.17 5.54 7.95** -0.54

Receptive 17.15 7.88 9.13 8.16 6.83 5.92 9.04** 2.30
Expressive 16.85 10.30 7.53 11.90 4.50 6.05 10.66** 3.03
Communication 21.00 10.88 7.69 9.73 9.53 8.87 12.42** -1.85
Self-help 13.44 8.13 10.31 9.90 6.93 6.84 4.76* 3.38
Social 15.46 9.89 8.00 11.99 5.27 9.11 8.78** 2.73
Motor 14.33 6.20 9.44 6.83 11.43 7.85 3.97 -1.99
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 29

Learning rates between intake and followup minus learning rates prior to intake (age equivalents per
year)

Non-verbal 0.65 0.53 0.20 0.73 0.12 0.38 0.49* 0.08

Receptive 0.73 0.61 0.19 0.42 -0.02 0.41 0.64** 0.21
Expressive 0.72 0.76 0.05 0.77 -0.23 0.54 0.79** 0.27
Communication 0.92 0.75 0.13 0.78 0.21 0.75 0.76** -0.08
Self-help 0.30 0.68 0.18 0.90 -0.14 0.64 0.28* 0.32
Social 0.48 0.83 0.02 0.97 -0.18 0.78 0.56* 0.20
Motor 0.01 0.53 -0.21 0.55 -0.11 0.75 0.17 -0.10
1
Parents’ level of education is a significant covariate (p < .05).
* Difference is statistically significant, after controlling for age at diagnosis and parents’ level of
education (p < .05).
** Difference is statistically significant, after controlling for age at diagnosis and parents’ level of
education (p < .01).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 30

Figure 1. Receptive language learning rates prior to intake (unfilled circles) and at
followup, after about 14 months of intervention (filled circles). The dashed line indicates
the normal learning rate (1 year of development for each year of age).
 BEHAVIOR ANALYTIC AND ECLECTIC TREATMENT OF AUTISM 31

Figure 2. Expressive language learning rates prior to intake (unfilled circles) and at
followup, after about 14 months of intervention (filled circles). The dashed line indicates
the normal learning rate (1 year of development for each year of age).