james e. carr - irfid · conversation-basic yes/no joint attention assertiveness sharing ......
TRANSCRIPT
James E. Carr Behavior Analyst Certification Board
Today’s Presentation
• Describe Early and Intensive Behavioral
Intervention (EIBI) of Autism
• Present Key Findings from the EIBI Literature
• Present 8 research studies on commonly used but
understudied practices
• Illustrate how practitioners can maintain access to
the research literature
Definition of EIBI
Early Intensive Behavioral Intervention or
Treatment (EIBI or EIBT) consists of 20-40 hours
per week of individualized instruction for children
with autism who begin treatment at the age of
four years or younger and who usually continue
for 2-3 years.
www.asatonline.org
Purpose of EIBI
• To build productive skills and minimize problem behavior (learning focused)
• To produce improvements large enough to change the developmental trajectory
• To prepare children for general education
Key Characteristics of EIBI
• Behavior-analytic instructional procedures
– Reinforcement, prompting, fading, etc.
• Hierarchically organized curriculum
– Designed to produce the repertoire of a typically developing 4-5-year old
• Intensive delivery
– 2-3 years; 25-40 hours/week
Characteristics of Quality EIBI
• Supervised by a qualified behavior analyst
• Multiple functioning areas are targeted
• Systems are in place for effective training and treatment fidelity
Key Studies
0
25
50
75
100
ABA (40 hrs/wk) ABA (10 hrs/wk)
% w
ith
Best
Outc
om
es
After 24+ mos.
Treatment Control
General education 9/19 0/19
Language
classroom
8/19 8/19
Special education 2/19 11/19
Lovaas (1987):
Educational Placements
Treatment gains maintained into late childhood
8 children in treatment group were indistinguishable from
typical peers on IQ and adaptive-behavior tests
0
25
50
75
100
ABA Autism
SPED
Generic
SPED
Me
an
IQ
Pre
Post
n = 29 n = 16 n = 16
IQ
309
IQ
105
IQ
39
Adaptiv
e
Behavio
r 309
105
Adaptiv
e
Behavio
r
39
Adaptiv
e
Behavio
r
Early and Intensive
Behavioral Intervention
2-3 years
Early and Intensive
Behavioral Intervention
2-3 years
Compliance
Mand Training
Block Imitation
Motor Imitation
Matching
Play
Songs
Independent work and play
Play Scripts
Receptive Instructions
Receptive Labels
Functional communication
Vocal imitation
Tacts
Conversation-Basic
Yes/No
Joint Attention
Assertiveness
Sharing
Social Initiations
Early and Intensive
Behavioral Intervention
Early and Intensive
Behavioral Intervention
Early and Intensive
Behavioral Intervention
Tacts
Early and Intensive
Behavioral Intervention
Tacts
10-Trial Session
10-Trial Session
10-Trial Session
10-Trial Session
10-Trial Session
10-Trial Session Mastery
Early and Intensive
Behavioral Intervention
10-Trial Session
10-Trial Session
10-Trial Session
10-Trial Session
10-Trial Session
Early and Intensive
Behavioral Intervention
2-3 years
Compliance
Mand Training
Block Imitation
Motor Imitation
Matching
Play
Songs
Independent work and play
Play Scripts
Receptive Instructions
Receptive Labels
Functional communication
Vocal imitation
Tacts
Conversation-Basic
Yes/No
Joint Attention
Assertiveness
Sharing
Social Initiations
Specific EIBI Practices
• There are substantial between-program differences in:
– Prompting
– Error correction
– Reinforcer delivery
– Antecedent stimulus presentation
– Mastery criteria
– Program sequence
– Measurement
– Trial arrangements
– Etc …
Trial Arrangements: Tacts
M
M
mand
mand
tact
tact
listener
listener
5.
2.
4.
1.
3.
6.
Trials
• The enormity of the EIBI task
– Thousands of hours of treatment
– Dozens of program areas
– Thousands of skills
• Published treatment manuals
– e.g., Lovaas (2003), Sundberg & Partington (1998)
• Approaches have been “branded” in the marketplace
– e.g., Lovaas/UCLA, Verbal Behavior, Precision Teaching
• The workshop circuit and Internet permit rapid dissemination
• Demand for services exceeds quality supply
Sources of Procedural Variability
8 EXPERIMENTS ON
UNDERSTUDIED PRACTICES
Study 1
Praise
• Types of Praise
– Descriptive praise
• Identifies to the learner the behavior for which he is
being praised
• e.g., “Good job raising your hand”
– General praise
• Does not specify the behavior
• e.g., “Good job”
Common Recommendations
“Specify the particulars of the accomplishment” and
“information about the value of their accomplishment”
(Brophy, 1981)
“Effects of praise may be bolstered when the praise is
specific”
(Simonsen et al., 2008)
Comparative Research?
• 2 unpublished dissertations – no difference
– Scheer (1977), Zahler (1975)
• Sellers and Higbee (unpublished)
– Children with ASD
– “receptive actions” program
– General & descriptive praise equally ineffective
• Stevens, Sidener, Reeve, & Sidener (2011)
– Children with ASD
– Tact program
– General & descriptive praise (with tokens) equally
effective
The Current Study
• General & Descriptive Praise Comparison
• Intraverbal training for 4 children with ASD
• Initial Reinforcer Evaluation
• Adapted Alternating Treatments Design
– 3 stimuli taught under each condition
• Praise + High-Preference Items + Errorless Prompts
0
1
2
3
4
5
6
7
1 2 3 4 5 6 7 8 9
Descriptive
General
No praise
Bre
ak P
oin
t
Sessions
Brad
Reinforcer Evaluation (PR-1)
0
20
40
60
80
100
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Set 1
Set 2
0
20
40
60
80
100
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Set 3
Set 4
Brad
Pe
rce
nta
ge
Co
rre
ct
9-Trial Blocks
Baseline Prompting + Praise
Descriptive
General
Brad
“What” Questions
0
1
2
3
4
5
6
7
8
9
10
1 2 3 4 5 6 7 8 9
Descriptive
General
No praise
Bre
ak P
oin
t
Sessions
Tina
Reinforcer Evaluation (PR-1)
0
20
40
60
80
100
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Set 1
Set 2
Tina
0
20
40
60
80
100
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Set 3
Set 4Tina
Pe
rce
nta
ge
Co
rre
ct
9-Trial Blocks
Baseline
Prompting +
Praise + HP Item + Errorless
Fill-Ins
Descriptive
General
Brad Tina Brett Shaun
Eval. 1 = Descriptive* = =
Eval. 2 Descriptive General = Descriptive
Acquisition Summary
Brad Tina Brett Shaun
Eval. 1 = = Descriptive
Eval. 2 General* General =
Maintenance Summary
Conclusions
• Findings
– Praise was rarely effective by itself (25% of
evaluations)
– No reliable within- or between-participant benefit of
descriptive praise
– Consistent with existing research
• Prerequisites?
– Listener/receptive repertoires
– Rule-governed behavior
Conclusions
• No evidentiary support for the differential effects of
descriptive praise
– Claims and fervor should be accordingly adjusted?
– Consider staff training issues
• Possible indirect effects?
Study 2
• Definition: delivering higher quality/magnitude reinforcers for independent responding compared to prompted responding
• Lovaas, Freitas, Nelson, and Whalen (1967)
– “This step is a rather important one in training, since continual reinforcement for prompted behavior probably would prevent a shift into imitative responding” (p. 174).
• It appears in many contemporary procedural manuals
– Maurice, Green, & Luce (1996)
– MacDuff, Krantz, & McClannahan (2001)
– Sundberg & Partington (1998)
Differential Reinforcement of
Unprompted Responses
Unprompted Response
SD: What’s this?
Delay
“Ball”
SR+: “That’s right!”
+
SD: “What’s this?”
Delay
Prompt: “Ball”
“Ball”
Sr+: “That’s right!”
Prompted Response
13%
87%
No
Yes
Love et al. (2009)
• Only one study conducted to date
– Olenick and Pear (1980)
• Evaluated differential SR+ via schedule
density
• Teaching procedures not necessarily
representative of contemporary EIBI
Previous Research
• Extend the Olenick and Pear (1980) study by:
– Evaluating the procedure with clinically relevant programs for children with autism
– Manipulating reinforcer type, not frequency
Purpose of the Current Study
• Reinforcer Assessment
– Praise vs. Food + Praise
• Treatment Evaluation
– Adapted ATD
– Non-differential vs. Differential SR
Method
Reinforcer Assessment
Treatment Evaluation
0
20
40
60
80
100
5 10 15 20
Sessions
Perc
en
tag
e U
np
rom
pte
d
Eric
0
25
50
75
100
5 10 15 20
Sessions
Pe
rce
nta
ge
Co
rre
ct
Diff. Sr+
Non-diff Sr+
Storm
Spider
Spider
Picture
Sequencing
Treatment Evaluation
0
10
20
30
40
50
60
70
80
90
100
5 10 15 20 25 30 35 40 45 50 55 60
Sessions
Pe
rce
nta
ge
Un
pro
mp
ted
Lonely v. Bored Crying v. Running Packing v. Watching Helping v. Cutting
0
25
50
75
100
5 10 15 20
Sessions
Pe
rce
nta
ge
Co
rre
ct
Diff. Sr+
Non-diff Sr+
Steve
Tacts
• Replications in other curricular areas
• Evaluation of mechanisms of action
– Adventitious reinforcement of errors in non-differential reinforcement
– Role of reinforcer delay following prompted responses in differential reinforcement
• Negative reinforcement?
Future Research
Study 3
• EIBI: 1000s of trials per week
• Debate re: the amount of data that need to be collected to evaluate performance
– Collecting data on every trial (continuous measurement)
– Collecting data discontinuously (e.g., only on the first trial)
Introduction
• Proponents of intermittent measurement argue that continuous data collection results in:
– increased session lengths
– delays of reinforcer delivery
– disruptions to the flow of teaching
• Until recently, there was no published empirical evidence to directly inform this debate
Introduction
28%
55%
18%
No Data
First Trial Only
Subset of Trials
Every Trial
Love et al. (2009)
Current Study
• Purpose:
– to experimentally compare continuous and discontinuous measurement
– across a number of curriculum areas
– in discrete-trial teaching programs for children with autism spectrum disorders
0
4
8
12
16
20
1 2 3 4 5 6 7 8 Mean
Training Sets
Nu
mb
er o
f S
ess
ion
s to
Ma
stery
All Trials
1st Trial
Erin
Receptive Discriminations
Mastery
0
25
50
75
100
1 2 3 4 5 6 7 8 Mean
Training Sets
Percen
tag
e o
f C
orrect
Resp
on
ses
All Trials
1st Trial
Erin
Receptive Discriminations
Maintenance
0
4
8
12
1 2 3 4 5 6 7 8 Mean
Training Sets
Nu
mb
er o
f S
ess
ion
s to
Ma
stery
Erin
Receptive Instruction Following
Mastery
0
25
50
75
100
1 2 3 4 5 6 7 8 Mean
Training Sets
Percen
tag
e o
f C
orrect
Resp
on
ses
Erin
Receptive Instruction Following
Maintenance
0
2
4
6
8
1 2 3 4 5 6 Mean
Training Sets
Nu
mb
er o
f S
ess
ion
s to
Ma
stery
All Trials
1st Trial
Allison
Receptive Discriminations
Mastery
0
25
50
75
100
1 2 3 4 5 6 Mean
Training Sets
Percen
tag
e o
f C
orrect
Resp
on
ses
All Trials
1st Trial
Allison
Receptive Discriminations
Maintenance
0
2
4
6
8
10
12
1 2 3 4 5 6 Mean
Training Sets
Nu
mb
er o
f S
ess
ion
s to
Ma
stery
Allison
Echoics
Mastery
0
25
50
75
100
1 2 3 4 5 6 Mean
Training Sets
Percen
tag
e o
f C
orrect
Resp
on
ses
Allison
Echoics
Maintenance
16
49
35
0
25
50
75
All Trials 1st Trial No Difference
Nu
mb
er o
f S
kills
.
MASTERY DATA
33
1
66
0
25
50
75
All Trials 1st Trial No Difference
Nu
mb
er o
f S
kil
ls
.MAINTENANCE DATA
• Discontinuous measurement resulted in slightly more efficient mastery
• Continuous measurement resulted in slightly better maintenance of skills
Findings
• Evaluate continuous and discontinuous measurement during Natural Environment Teaching
• Evaluate continuous and discontinuous measurement with maintenance programming during follow-up period
• Evaluate the same experimental question with a less experienced therapist
Future Research
Study 4
The Tact
“You’re right,
that is a guitar!”
Antecedent Behavior Consequence
Guitar
• Tact training recommendations differ
– Lovaas (2003); Leaf & McEachin (1999) –
“expressive labels”
• “What is this?”
– Sundberg & Partington (1998)
• “What is this?” question fading
Early Intervention Curricula
• Introducing “What is this?” at the
beginning of training:
– Prompt dependence?
– Question intraverbally controlling one
response?
– Learner repeats “what is this” as a part of
the answer?
Potential Stimulus Control
Problems
Research on Tact Training
• Prior studies have shown that the question
“What is this?” can:
– Hinder tact acquisition (Sundberg, Endicott, & Eigenheer, 2000)
– Block the acquisition of new tacts (Partington, Sundberg, Newhouse, & Spengler, 1994)
The Current Comparison
What is
this?
• Tact Training
– Object+Question condition produced more
efficient acquisition with 2/4 participants
– Object-Only condition produced more efficient
acquisition with 2/4 participants
– Only 1 participant acquired all of the tacts in
each condition (Michael)
• Maintenance Evaluation
– Tacts maintained at end-of-training levels in
both conditions
Results
• Stimulus control problems occurred but were
not pervasive
• Alternating treatments design might have
impacted the results
• Practitioners should evaluate/correct stimulus
control problems at the individual level
• Future research: when to introduce questions?
Discussion
Study 5
The Mand
Antecedent Behavior Consequence
Guitar Guitar
deprivation.
• Alternating treatments design might have
impacted the results
• Object presence might have acquired stimulus
control over the behavior (blocking the Q)
• Practitioners should evaluate/correct stimulus
control problems at the individual level
• Future research: when to introduce questions?
Discussion
Study 6
Discrete-Trial Teaching
• The discrete trial
– SD (instruction)
– Prompt
– Learner’s response
– Consequence
– Intertrial interval
• Characteristics
– Fast paced
– Highly structured
– Teacher led
Criticisms of DTT
• Too structured
– Not how typical children learn
– Skills might not generalize
– Rote responding
• Too “demanding”
– Might evoke problem behavior
Embedded DTT
• DTT embedded in the context of a game
– Same components of instruction
– Fun, relaxed style in naturalistic
environment
• Based on client interests
– More similar to instructional games for
typically developing children
Purpose
• Compare embedded and traditional DTT
to teach receptive language to children
with autism
– Rate of acquisition
– Affect
– Preference
Traditional DTT
Listener Responding
Before SD Arrange array
SD “Point to the rhino.”
Learner Response Point to rhino
Prompt Gestural
Consequence Praise + edible
Embedded DTT
Ben
Sawyer
Session Duration
Affect
*
Choice Assessment
1 Session
of
Embedded
1 Session
of
Traditional
2 min
of no
interaction
Choice
Overall Results
• Both procedures comparable in effectiveness
• Both procedures comparable in efficiency
(slight edge to Embedded DTT)
• Similar positive and negative affect & preference
for Ben
• More positive affect and higher preference with
Embedded DTT for Sawyer
• Main clinical implication? … Have fun!
Study 7
Discriminations
• Most commonly targeted skills in early
intervention programming (Smith, 2001)
• Types of discriminations
– Simple
– Conditional
“Eating”
“Running”
“Swimming”
Conditional Discriminations
Auditory-Visual
Conditional
Discriminations
Conditional Discrimination Programs
• Also known as
– Receptive Labeling / Identification
– Listener Responding
• Targets
– Objects
– Actions
– Colors
– Emotions
– Feature, function, & class
• Two methods for teaching conditional
discriminations in applied settings
– Simple/Conditional Method
• Based on procedures described by Lovaas
(2003)
– Conditional Only Method
• Based on procedures described by Green
(2001)
Conditional Discrimination Training
“Show me puppy.”
Simple Conditional Discrimination Training
STEP 1
“Show me apple.”
Simple Conditional Discrimination Training
STEP 2
“Show me puppy.”
Simple Conditional Discrimination Training
STEP 3
“Show me apple.”
Simple Conditional Discrimination Training
STEP 4
“Show me apple / puppy.”
Simple Conditional Discrimination Training
STEP 5
“Show me guitar.”
Simple Conditional Discrimination Training
STEP 6
“Show me guitar / puppy.”
Simple Conditional Discrimination Training
STEP 7
“Show me guitar / apple.”
Simple Conditional Discrimination Training
STEP 8
“Show me guitar / apple / puppy.”
Simple Conditional Discrimination Training
STEP 9
?
Recommended
Clinical Practice
Research
Simple/Conditional
Method
Love, Carr, Almason,
& Petursdottir, 2009
Depends S/C
CD Only
The Current Study
• Concern with the Simple/Conditional method
– Steps requiring only simple discriminations may
interfere with subsequent conditional discriminations
(Green, 2001)
• Purpose
– To compare the Simple/Conditional and Conditional
Only training methods with children with autism
• Adapted Alternating Treatments Design
– 3 stimuli taught under each condition
Animals
Animals
Evaluation 3: Step 5 Error Analyses
0
10
20
30
40
50
60
70
80
90
100
0 3 6 9 12 15
Session
Per
cen
tage
of
Acr
oss
Ses
sio
n
Win
-Sta
y R
esp
on
ses
0
10
20
30
40
50
60
70
80
90
100
0 3 6 9 12 15
Blocks of Sessions
Per
cen
tage
of
Wit
hin
Ses
sion
Win
-Sta
y R
esp
on
ses
Per
cen
tag
e of
Res
po
nse
s fr
om
Pre
vio
us
Ste
p
Blocks of Sessions
Error Analysis
Error Analysis
Animals
Overall Results
• Acquisition in 7 of 8 evaluations for the
Conditional Only method
• Faulty stimulus control in 4 of 8 evaluations for
the Simple/Conditional method
• Maintenance was better in the Conditional Only
method
`
A Comparison of Methods for Teaching
Receptive Labeling to Children with
Autism Spectrum Disorders:
A Systematic Replication
Grow, Kodak, & Carr (in press)
JABA
Study 8
Clinical/Educational Implications?
• Teach A-V Conditional Discriminations as
A-V Conditional Discriminations
jim-carr.net/grow2013.pdf
Unanswered Questions
Measurement
PECS
Mastery Criteria Blocking Stereotypy
Fading Methods Array Size
Program Order
Task Interspersal
Reinforcer Schedule Thinning
Maintenance Programming
Lovaas (1987)
0
25
50
75
100
ABA (40 hrs/wk) ABA (10 hrs/wk)
% w
ith
Best
Ou
tco
mes
After 24+ mos.
• The resources available from the behavioral research community do not match:
- the pace of clinical dissemination
- the degree of unanswered questions
• Clinicians are potentially well suited to contribute to this mission
- access to many clients
- services are already being provided
- influence over the training context
• Clinicians might collaborate with established researchers to help answer even more important questions
Getting the Job Done
There are MANY journals!
There are MANY journals! Access to journal content is expensive
$50-500/yr Limited searchable access to the archive
Journal web sites (e.g., JABA)
▪ Incomplete
PsycINFOTM
▪ $149-500/year Contacting the new literature is effortful
Obtain access to PsycINFOTM
Personal/organizational subscription; alumni membership?
Compile a list of specific journals
Subscribe to JABA, BAP + others Sign up for ToC alerts
Obtain access to PsycINFOTM
Personal/organizational subscription; alumni membership?
Compile a list of specific journals
Subscribe to JABA + Sign up for ToC alerts Review ToC alerts for potentially relevant articles
Save in a Word file for later review
Obtain articles from journals, ILL, or authors Bookmark journal pages that have “in press” articles or do
not have ToC alerting
Periodically check for articles
For those who work with others …
Journal club meetings
Assign a supervisor to act as a liaison with the literature
For those who work alone …
Self-management strategies
▪ e.g., goal setting, self-monitoring, contingency contracts
Thank you.