james e. carr - irfid · conversation-basic yes/no joint attention assertiveness sharing ......

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.