Discussion 5

AN ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES FOR LEARNERS WITH AUTISM SPECTRUM DISORDER

KATE A. JOHNSON AND JASON C. VLADESCU

CALDWELL UNIVERSITY

TIFFANY KODAK

UNIVERSITY OF WISCONSIN MILWAUKEE

AND

TINA M. SIDENER

CALDWELL UNIVERSITY

Differential reinforcement procedures may promote unprompted correct responding, resulting in a quicker transfer of stimulus control than nondifferential reinforcement. Recent studies that have compared reinforcement arrangements have found that the most effective arrangement may differ across participants. The current study conducted an assessment of differential rein- forcement arrangements (i.e., quality, schedule, and magnitude) and nondifferential reinforce- ment to identify the most effective arrangement for each participant. The assessment phase showed that the quality arrangement was the most efficient for all participants during auditory- visual matching. Next, a validation phase was conducted to evaluate whether the assessment would predict the most effective arrangement across multiple skills. The results from the assess- ment phase were validated for all participants for the same skill. However, the results were only validated for one participant during the other skills (i.e., tact and intraverbal). The results are discussed in light of previous research and future areas of research. Key words: autism spectrum disorder, differential reinforcement, reinforcement magnitude,

reinforcement quality, reinforcement schedule

Researchers have recommended the use of differential reinforcement during skill acquisi- tion programs (e.g., Grow & LeBlanc, 2013). Differential reinforcement may be arranged by delivering higher-quality, higher-magnitude, or denser schedules of reinforcement following unprompted correct responses and delivering lower-quality, smaller-magnitude, or leaner schedules of reinforcement following prompted correct responses (Vladescu & Kodak, 2010).

A group of studies have involved the compari- son of nondifferential reinforcement to a single type of differential reinforcement; that is, schedule (Hausman, Ingvarsson, & Kahng, 2014; Olenick & Pear, 1980; Touchette & Howard, 1984) or quality (Karsten & Carr, 2009) manipulations. Although these studies demonstrated the potential usefulness of differ- ential reinforcement in promoting unprompted correct responses, they do not provide infor- mation about the relative effectiveness of such arrangements. Direct comparisons of differen- tial reinforcement arrangements are important, as the outcomes of these studies may identify the conditions under which one procedure is superior to another and help guide clinicians in the selection of one arrangement over others.

This article is based on a thesis submitted by the first author, under the supervision of the second author, at Caldwell University in partial fulfillment for the require- ments of the Master of Arts in Applied Behavior Analysis.

Address correspondence to Jason C. Vladescu, Depart- ment of Applied Behavior Analysis, Caldwell University, 120 Bloomfield Avenue, Caldwell, NJ 07006. E-mail: [email protected]

doi: 10.1002/jaba.372

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2017, 50, 290–303 NUMBER 2 (SPRING)

© 2017 Society for the Experimental Analysis of Behavior

290

Although the findings from research in the area of differential reinforcement during skill acquisition programming have been positive, one reinforcement arrangement has not consist- ently produced superior results over other arrangements across learners with autism spec- trum disorder (ASD). We identified three stud- ies that involved the evaluation of multiple differential reinforcement arrangements (Boudreau, Vladescu, Kodak, Argott, & Kisa- more, 2015; Cividini-Motta & Ahearn, 2013; Fiske et al., 2014). The results of these compar- ison studies suggest that the most effective rein- forcement arrangement was inconsistent across participants. For example, Boudreau et al. (2015) compared two differential rein- forcement arrangements (i.e., quality and mag- nitude manipulations) to nondifferential reinforcement in teaching tacts to three partici- pants with ASD. For one participant, the most effective condition was nondifferential rein- forcement. For the other two participants, dif- ferential reinforcement was most efficient; the magnitude manipulation was most efficient for one and the quality manipulation was most efficient for the other. This variation across learners creates the need

for clinicians to alter their teaching practices for individual learners instead of using one rein- forcement arrangement across all learners as suggested by some early intervention manuals (e.g., Lovaas, 2003). As such, it seems worth- while to explore whether an assessment could be created to identify the most efficient learner- specific reinforcement arrangement. Recent studies have evaluated the usefulness of assess- ments to identify the most efficient procedures to use during subsequent skill acquisition pro- gramming (e.g., Kodak, Fisher, Clements, Paden, & Dickes, 2011; McGhan & Lerman, 2013; Seaver & Bourret, 2014). For example, McGhan and Lerman conducted an assessment to determine the most efficient of four error correction (EC) procedures and then validated the results of the assessment with five

individuals with ASD. The results indicated that the assessment correctly predicted the least intrusive, most efficient EC procedure with 79% accuracy. To date, however, no studies have compared a manipulation of reinforce- ment schedule, quality, and magnitude to non- differential reinforcement. Therefore, the purpose of the present study was to develop and conduct an assessment similar to the one created by McGhan and Lerman to identify the most efficient of three reinforcement arrange- ments within the context of arbitrary auditory- visual (AV) matching and subsequently deter- mine if the outcome of this assessment would predict the most efficient reinforcement arrangement within the context of AV match- ing, tact, and intraverbal instruction.

METHOD

Participants Three boys diagnosed with ASD partici-

pated. Jacob was a 10-year-old who obtained standard scores of 62 (qualitative description: extremely low) and 68 (extremely low) on the Peabody Picture Vocabulary Test-Fourth Edi- tion (PPVT-4; Dunn & Dunn, 2007) and the Expressive Vocabulary Test-Second Edition (EVT-2; Williams, 2007), respectively. Jackson was a 9-year-old who obtained standard scores of 68 (extremely low) and 75 (moderately low) on the PPVT-4 and EVT-2, respectively. Aaron was an 8-year-old who obtained scores of 71 (moderately low) and 75 (moderately low) on the PPVT-4 and EVT-2, respectively. Jacob scored in level three on the mand portion; and level two on the tact, listener, visual perceptual and match-to-sample, motor imitation, and intraverbal portions of the Milestones Assess- ment from the Verbal Behavior-Milestones Assessment and Placement Program (VB- MAPP; Sundberg, 2008). Jackson and Aaron scored in level three for these portions of the VB-MAPP. All participants had a history of instruction that incorporated differential

291ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

reinforcement that most closely resembled a reinforcement quality manipulation, albeit not as structured as arranged in the current evaluation.

Setting and Materials All sessions were conducted in a self-

contained classroom for individuals with devel- opmental disabilities in a suburban public school. Sessions for each participant were con- ducted at their individual work areas, which contained a table, chairs, session materials (e.g., data sheet, timer, edibles, stimuli bin- ders), and an iPad on a tripod to record most sessions. The experimenter created a stimulus binder for each condition to present trials for each session. A sheet of colored paper (condi- tion specific colors were assigned based on results of color preference assessment) was attached to the cover of each binder. Inside each binder were trial sheets that consisted of a sheet of colored paper (specific to each condi- tion) containing zero, one, or three pictures (for the intraverbal, tact, and AV matching tasks, respectively) for Jacob, and zero, one, or six pictures for Jackson and Aaron. Addition- ally, a blank piece of colored paper was placed on top of each trial sheet. This piece of paper prevented participants from viewing the stimuli prior to the delivery of the antecedent verbal stimulus and provided an opportunity for an observing response. The experimenter sat adja- cent to or across from each participant at the table.

Dependent Variables and Interobserver Agreement During the assessment and validation phases,

the experimenter scored unprompted correct, unprompted incorrect, prompted correct, and prompted incorrect responses on data sheets prepared for each session. Only unprompted correct responses are depicted in the figures, and were defined as the participant emitting

the predefined target response (saying the name of the item in the picture during tact trials, touching the comparison stimulus correspond- ing to the sample stimulus during AV matching trials, and providing the Spanish word for words presented in English by the experimenter during intraverbal trials) prior to the delivery of the prompt (definitions for other responses are available from second author). The experi- menter also recorded session duration using a digital timer. The experimenter started the timer immediately prior to the presentation of the antecedent stimulus on the first trial of the session and stopped the timer immediately fol- lowing the completion of the last trial of the session. A second observer independently scored par-

ticipant responding during a minimum of 33% of sessions across conditions from the assess- ment and validation phases in vivo or from video for interobserver agreement (IOA) pur- poses. IOA was calculated on a trial-by-trial basis by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. Mean IOA scores for Jacob were 95% (range, 67% to 100%) during the assessment phase and 99% (range, 78% to 100%) during the validation phase. For Jack- son, mean IOA scores were 98% (range, 83% to 100%) during the assessment phase and 99% (range, 92% to 100%) during the valida- tion phase. For Aaron, mean IOA scores were 98% (range, 83% to 100%) during the assess- ment phase, and 99% (range, 92% to 100%) during the validation phase.

Preference Assessments The experimenter conducted a color prefer-

ence assessment to identify colors to be arranged as condition-correlated stimuli during the reinforcer assessments (to aid in the dis- crimination of the sessions), and the assessment and validation phases (details available from second author). The experimenter conducted a

KATE A. JOHNSON et al.292

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

paired-stimulus preference assessment (PS; Fisher et al., 1992) using 12 edibles identified by the participants’ parents prior to the onset of the assessment phase. The experimenter con- ducted a brief multiple-stimulus without replacement preference assessment (MSWO; Carr, Nicolson, & Higbee, 2000) using the top five edibles from the PS prior to each reinforcer assessment, assessment phase, and validation phase session. The first item chosen by the par- ticipant during the MSWO was used during the ensuing session.

Edible Amount and Size Assessment Using procedures identical to Boudreau

et al. (2015), the experimenter conducted an assessment to determine the number of pieces of an edible item (hereafter referred to as the edible amount) that were provided following unprompted correct responses during the qual- ity, magnitude, schedule, and nondifferential conditions, and following prompted correct responses during the nondifferential condition. The experimenter also conducted an assessment to determine the size of the edible that was pro- vided following prompted correct responses during the magnitude manipulation condition (hereafter referred to as the small edible) for the top five edibles identified in the PS preference assessment. For all participants, the edible amount was one whole piece of their top five edibles, except for one edible for Aaron. Aar- on’s edible amount for fruit snacks was two. Additionally, the small edible for all partici- pants was 1/8th the original size for all edibles.

Reinforcer Evaluation Each participant completed three

progressive-ratio (PR) reinforcer evaluations (Roane, Lerman & Vorndran, 2001): praise compared to no consequence (i.e., extinction); praise compared to praise and edible amount; and praise and edible amount compared to praise and the small edible. The experimenter

selected an arbitrary response (e.g., button pushing; additional details available from sec- ond author) for each participant to complete. The materials associated with the arbitrary task incorporated a color (based on the color prefer- ence assessment) specific to each condition in the PR reinforcer evaluations and were placed on the table prior to each session. The experi- menter said the name of the color and simulta- neously prompted the participant to touch the colored paper. We incorporated this observing response in an attempt to increase the probabil- ity that the participants discriminated the con- ditions. Prior to each session, the experimenter conducted one forced-exposure trial. The num- ber of correct responses required to obtain the putative reinforcer increased by two following the completion of the preceding requirement. Sessions were terminated after 5 min, following 1 min of no response, or following a verbal response indicating the participant no longer wanted to engage in the task.

Target Identification Three (Jacob) or six (Jackson and Aaron)

arbitrary AV matching targets were assigned to each condition for the assessment phase. Arbi- trary stimuli were used to control for target dif- ficulty, learning history, and any learning that might occur outside of the assessment. The validation phase consisted of evaluating

the four reinforcement arrangements across three skills: AV matching, tacts, and intraver- bals. For each skill, three (Jacob) or six (Jackson and Aaron) targets were identified for each condition. Targets were selected with input and approval from parents of the partici- pants and included labeling states and people, and English-Spanish translations (description of the pretest available from second author). Sti- muli were assigned to conditions such that no two auditory stimuli in a target set sounded similar (e.g., “cat” and “rat”), no two visual sti- muli in a target set shared physical similarity

293ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

(e.g., lizard and alligator), and all conditions contained targets with an approximately equal number of syllables.

General Procedure During the assessment and the validation

phases, each target was presented an equal number of times during a session and on no more than two consecutive trials. All sessions consisted of nine (Jacob) or twelve (Jackson and Aaron) trials. Sessions were conducted 1 to 5 days per week with a minimum of 5 min between each condition. Sessions for each con- dition were discontinued upon reaching mas- tery criterion, defined as two consecutive sessions with unprompted correct responding on 100% of trials. Similar to McGhan and Ler- man (2013), once mastery was achieved in one condition, the participant had up to five addi- tional sessions to achieve mastery in the remaining conditions. Each reinforcement arrangement condition

was correlated with a different color (based on the color preference assessment), which remained the same throughout the assessment and validation phases. At the start of each assessment session, the experimenter placed the four stimuli binders (one binder for each of the four reinforcement arrangements) in a horizon- tal line and physically prompted the selection of a binder. The experimenter repeated this procedure prior to beginning each session. Conditions were conducted in a random order without replacement. Prior to each trial, an observing response was physically prompted (i.e., removing the cover sheet to reveal the trial sheet). Across all trials, a 3-s constant prompt delay was used. Following unprompted incor- rect responses, the experimenter modeled the correct response, and allowed the participant 3 s to respond. If the participant engaged in a prompted incorrect response following the model, the experimenter presented the next trial.

Assessment phase. The purpose of the assess- ment phase was to identify the most efficient of four reinforcement arrangements during train- ing of arbitrary AV matching targets. Each trial sheet displayed a horizontal array of three (Jacob) or six (Jackson and Aaron) stimuli, including the target stimulus and two or five comparison stimuli. Each stimulus served as the target on one-third or one-half of the trials and as a comparison stimulus on the remaining trials. The reinforcement contingencies associ- ated with unprompted and prompted correct responses for each condition are described below. Nondifferential. The experimenter delivered

the edible amount (e.g., two pieces of a fruit snack) and verbal praise following unprompted correct and prompted correct responses. Quality manipulation. The experimenter

delivered the edible amount and verbal praise following each unprompted correct response and provided verbal praise only following a prompted correct response. Magnitude manipulation. The experimenter

delivered the edible amount and verbal praise following unprompted correct responses and the small edible (e.g., 1/8th of a fruit snack) and verbal praise following prompted correct responses. Schedule manipulation. The experimenter

delivered the edible amount and praise on a fixed ratio (FR) 1 schedule for unprompted correct responses and on an FR 3 schedule for prompted correct responses. The experimenter delivered no consequence on trials in which the participant engaged in a prompted correct response that did not meet the FR 3 schedule requirement. We arranged these specific sche- dules based on what has been included in pre- vious studies (e.g., Hausman et al., 2014; Touchette & Howard, 1984). Validation phase. For each participant, an

adapted alternating treatments design (Sindelar, Rosenberg, & Wilson, 1985) embedded within a multiple baseline design was used to validate

KATE A. JOHNSON et al.294

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

the results of the assessment phase. During the validation phase, all four reinforcement arrange- ments were compared across three skills (i.e., AV matching, tact, and intraverbal). AV trial sheets displayed stimuli in an identical fashion to the assessment phase. Tact trial sheets contained one stimulus per page. Intra- verbal trial sheets were blank to maintain con- sistency of the presentation across tasks. Baseline. During each trial, the experimenter

presented the relevant antecedent verbal stimu- lus (i.e., said the name of the target for AV matching trials, said, “What is it?” during tact trials, and said an English word for intraverbal trials). Across all baseline conditions, the exper- imenter delivered generic praise at the end of each trial (e.g., “nice work”), regardless of the participant response. No other consequences were provided and no condition-correlated col- ors were presented during baseline sessions. Training. At the start of each session, the

experimenter placed the four stimuli binders in a horizontal line on the table. The experi- menter presented the binders in a quasirandom order such that no binder was presented in the same position during two consecutive trials. The experimenter prompted the selection of a binder. All other procedures remained the same as in the assessment phase.

Procedural Integrity and Procedural Integrity IOA An independent observer collected proce-

dural integrity data (specific steps are available from second author) during a minimum of 25% of sessions for each participant in vivo or from videos. Treatment integrity was calculated by dividing the number of correctly implemen- ted steps by the number of correctly implemen- ted steps plus the number of incorrectly implemented steps and multiplying by 100. Mean treatment integrity scores were 100% for Jacob, 99% (range, 92% to 100%) for Jackson, and 99% (range, 83% to 100%) for Aaron. A

secondary observer measured procedural integ- rity during a minimum of 25% of sessions across participants for IOA purposes. Proce- dural integrity IOA was calculated by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. Mean treatment integrity IOA scores were 100% for Jacob, 99% (range, 92% to 100%) for Jackson, and 99% (range, 92% to 100%) for Aaron.

RESULTS

During the reinforcer evaluations, partici- pants engaged in an average of 332 (range, 87 to 596) more cumulative responses in the praise condition compared to the no conse- quence condition, an average of 176 (range, 115 to 412) more cumulative responses in the praise plus edible condition versus the praise only condition, and an average of 255 (range, 109 to 343) more cumulative responses in the edible amount plus praise condition compared to the small edible plus praise condition. These results demonstrate that the edible amount plus praise had more reinforcing value than no con- sequence, praise only, and small edible plus praise (individual participant figures available from second author). Figure 1 illustrates the results of the assess-

ment phase for Jacob, Jackson, and Aaron. Jacob and Jackson demonstrated mastery in the quality condition in 8 sessions (72 trials, 50 min) and 16 sessions (192 trials, 46 min), respectively. Neither achieved mastery in the other conditions. Aaron demonstrated mastery in 12 sessions (144 trials, 41 min) in the qual- ity condition and 17 sessions (204 trials, 58 min) in the nondifferential condition. Mas- tery level responding was not observed in the other conditions. Figures 2, 3, and 4 illustrate the results of

the validation phase for Jacob, Jackson, and Aaron, respectively. During baseline, all partici- pants engaged in zero or chance levels of

295ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

unprompted correct responses across skills. Fol- lowing baseline, unprompted correct respond- ing increased across all conditions for all participants. Jacob demonstrated mastery in 12 training sessions (108 training trials, 29 min) in the quality condition during AV matching, 21 training sessions (189 training trials, 71 min) in the magnitude condition for the intraverbal, and 19 training sessions (171 training trials, 59 min) in the schedule condition for the tacts. Jacob did not achieve

mastery in the other conditions for any of the skills. For AV matching, Jackson demonstrated

mastery in 8 training sessions (96 training trials, 24 min) in the quality condition, 11 training sessions (132 training trials, 37 min) in the magnitude condition, 12 training sessions (144 training trials, 39 min) in the schedule condition, and 14 training sessions (168 training trials, 43 min) in the nondifferential condition.

0

20

40

60

80

100

Jacob

Quality Magnitude Schedule Nondifferential

0

20

40

60

80

100

P er

ce n ta

g e

o f

U n p ro

m p te

d C

o rr

ec t

R es

p o n se

s

Jackson

0 10 20 30 40 50 60 70 80 0

20

40

60

80

100

Sessions

Aaron

Figure 1. Percentage of unprompted correct responses for Jacob, Jackson, and Aaron across all conditions during the assessment phase.

0 20 40 60 80 100 120 140 160

0

20

40

60

80

100

AV Matching

Sessions

0

20

40

60

80

100

P er

ce n ta

g e

o f

U n p ro

m p te

d C

o rr

ec t R

es p o n se

s

Tact

0

20

40

60

80

100

Intraverbal

Magnitude

Schedule Nondifferential

Quality

BL Training

Figure 2. Percentage of unprompted correct responses across the intraverbal, tact, and AV matching skills during baseline (BL) and training of the validation phase for Jacob.

KATE A. JOHNSON et al.296

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

For the intraverbals, he demonstrated mastery in 16 training sessions (192 training trials, 39 min) in the quality condition. For the tacts, he demonstrated mastery in 7 training sessions (84 training trials, 24 min) in the magnitude condition. Jackson did not demonstrate mastery in any of the other conditions for the intraverbal and tact skills.

For AV matching, Aaron demonstrated mas- tery in 6 training sessions (72 training trials, 17 min) in the quality condition, 7 training sessions (84 training trials, 21 min) in the non- differential condition, 8 training sessions (96 training trials, 21 min) in the magnitude condition, and 11 training sessions (132 train- ing trials, 30 min) in the schedule condition.

0

20

40

60

80

100

Quality Magnitude Schedule Nondifferential

AV Matching

0 10 20 30 40 50 60 70 80 90 100 110

0

20

40

60

80

100

Tact

Sessions

0

20

40

60

80

100

Intraverbal

P er

ce n ta

g e

o f

C o rr

ec t R

es p o n se

s

TrainingBL

Figure 3. Percentage of unprompted correct responses across the AV matching, intraverbal, and tact skills during baseline (BL) and training of the validation phase for Jackson.

297ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

For intraverbals, he reached mastery in 7 train- ing sessions (84 training trials, 17 min) in the schedule condition. He reached mastery in 12 training sessions (144 training trials, 33 min) and 15 training sessions (180 training trials, 45 min) in the magnitude and nondiffer- ential conditions, respectively, for tacts. Tables 1, 2, and 3 summarize the total train-

ing sessions, training sessions to mastery,

training trials to mastery, total training time, and average training time per target during each condition during the assessment and vali- dation phases for Jacob, Jackson, and Aaron, respectively. We included total training ses- sions, total training time, and average training time per target for conditions in which partici- pants did not demonstrate mastery to show that these measures exceeded those for

0

20

40

60

80

100

P er

ce n ta

g e

o f

U n p ro

m p te

d C

o rr

ec t R

es p o n se

s

AV Matching

0

20

40

60

80

100

Tact

Quality Magnitude Schedule Nondifferential

0 10 20 30 40 50 60 70 80 90 100

0

20

40

60

80

100

Sessions

Intraverbal

TrainingBL

Figure 4. Percentage of unprompted correct responses across the tact, AV matching, and intraverbal skills during baseline (BL) and training of the validation phase for Aaron.

KATE A. JOHNSON et al.298

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

conditions in which participants did demon- strate mastery. Further, we included a calcula- tion of average training time per target, as this is a more sensitive measure of instructional effi- ciency than training sessions to mastery (Yaw et al., 2014). When evaluating these data, the lowest average training time per target was asso- ciated with the condition in which participants demonstrated mastery level responding in the fewest number of sessions across phases and skills.

DISCUSSION

Recent evaluations have demonstrated that the most efficient reinforcement procedure to arrange during skill acquisition programming may vary across participants (e.g., Boudreau et al., 2015). These results suggest clinicians

may need to implement different reinforcement arrangements based on specific consumer responding. Without a methodology for select- ing the most appropriate reinforcement arrangement, clinicians may arbitrarily select one procedure over others regardless of relative efficiency. As an alternative, an initial assess- ment of reinforcement procedures conducted prior to instruction may be useful. The current study evaluated whether an initial assessment of four reinforcement arrangements during AV matching instruction may be useful in identify- ing the most efficient procedure for subsequent instruction across multiple types of skills (i.e., AV matching, tacts, intraverbals) for three children with ASD. The results indicated that the assessment successfully predicted the most efficient reinforcement arrangement for the matched skill (AV matching), but was not gen- erally predictive for the other skill types (tacts

Table 1 Amount and Duration of Training for Jacob.

Total Average

Condition Sessions Trials Time Time per target

Assessment AV Matching Magnitude 13 117 44 14 Quality 8* 72 40 13 Schedule 13 117 59 20 Nondifferential 13 117 60 20

Validation AV Matching Magnitude 17 153 44 15 Quality 12* 108 29 10 Schedule 17 153 41 14 Nondifferential 17 153 39 13

Tact Magnitude 24 216 63 21 Quality 24 216 63 21 Schedule 19* 171 59 20 Nondifferential 24 216 75 25

Intraverbal Magnitude 21* 189 71 24 Quality 26 234 88 29 Schedule 26 234 86 29 Nondifferential 26 234 99 33

Note. Asterisks indicate that mastery was achieved in the number of sessions depicted (in the absence of an asterisk, mastery was not achieved); time is depicted in minutes.

Table 2 Amount and Duration of Training for Jackson.

Total Average

Condition Sessions Trials Time Time per target

Assessment AV Matching Magnitude 21 252 66 11 Quality 16* 192 46 8 Schedule 21 252 65 11 Nondifferential 21 252 67 11

Validation AV Matching Magnitude 11* 132 37 6 Quality 8* 96 24 4 Schedule 12* 144 39 7 Nondifferential 14* 168 43 7

Tact Magnitude 7* 84 24 4 Quality 12 144 27 4 Schedule 12 144 30 5 Nondifferential 12 144 30 5

Intraverbal Magnitude 21 252 56 9 Quality 16* 192 39 6 Schedule 21 252 54 9 Nondifferential 21 252 64 11

Note. Asterisks indicate that mastery was achieved in the number of sessions depicted (in the absence of an asterisk, mastery was not achieved); time is depicted in minutes.

299ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

and intraverbals). This outcome suggests that the most efficient reinforcement arrangement may vary across consumers and skill types. We arranged the assessment phase to include

only one skill because we hypothesized that the most efficient reinforcement arrangement for one skill would predict the most efficient arrangement across other skills. As this was not the case, including a variety of skills (e.g., AV matching, tacts, intraverbals) in the initial assessment is an important consideration for future studies. Such studies would be helpful in determining if an initial assessment involving a variety of skills would be predictive of subse- quent performance across skill domains. Alter- natively, given that consumers may acquire skills under various reinforcement arrange- ments, future researchers could attempt to identify a reinforcement arrangement that is

generally effective and efficient for a large num- ber of consumers and across a variety of skills. Researchers may then delineate situations (e.g., consumer learning is stagnant) in which conducting an assessment of alternative rein- forcement arrangements would be beneficial. The findings of the current evaluation also

suggest the need to further evaluate the predic- tive validity of assessments for other instruc- tional components (e.g., EC for incorrect responses). For example, McGhan and Lerman (2013) arranged their EC assessment to include only AV matching, and validated the results across the same skill type. Thus, it is unclear if the assessment would be predictive of the most efficient EC procedure across different skill types. Until such studies are conducted, we urge practitioners to consider the relative value of these assessments in their clinical work. The lack of predictive validity between the results of the assessment with one skill type to the results of comparisons with other skill types calls into question the benefits of these assessments. Future studies may identify assessment models that have strong discriminant and predictive validity, but until that time we caution practi- tioners in the use of these assessments, particu- larly for consumers whose learning is progressing well. Further research is needed to determine why

the assessment only predicted the most efficient reinforcement arrangement for a matched skill, but did not generally predict the most efficient arrangement for different skills. One potential variable may be instructional history. Previous research (e.g., Coon & Miguel, 2012; Free- man & Lattal, 1992) has demonstrated the effect of proximal history on subsequent responding. In the current evaluation, the qual- ity condition was associated with the most effi- cient acquisition during the assessment phase for all participants. Efficiency was quantified and demonstrated by not only training sessions and trials to mastery, but total training time and mean training time per stimulus. Following

Table 3 Amount and Duration of Training for Aaron.

Total Average

Condition Sessions Trials Time Time per target

Assessment AV Matching Magnitude 17 204 58 10 Quality 12* 144 41 7 Schedule 17 204 60 10 Nondifferential 17* 204 58 10

Validation AV Matching Magnitude 8* 96 21 4 Quality 6* 72 17 3 Schedule 11* 132 30 5 Nondifferential 7* 84 21 4

Tact Magnitude 12* 144 33 6 Quality 17 204 46 8 Schedule 17 204 45 7 Nondifferential 15* 180 45 8

Intraverbal Magnitude 12 144 25 4 Quality 12 144 22 4 Schedule 7* 84 17 3 Nondifferential 12 144 25 4

Note. Asterisks indicate that mastery was achieved in the number of sessions depicted (in the absence of an asterisk, mastery was not achieved); time is depicted in minutes.

KATE A. JOHNSON et al.300

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

from this, it seems reasonable to suggest that average reinforcement density across training sessions was highest in the quality condition as participants demonstrated unprompted correct responding across trials more quickly in this condition relative to the other conditions. A recent history of a greater density of reinforce- ment with a particular reinforcement arrange- ment may function to establish different degrees of stimulus control across conditions, which in turn could have influenced subse- quent responding across conditions under a similar context. By arranging our assessment phase to include only one skill (AV matching), it may be possible that the differential degrees of stimulus control were only established for this skill and explain why the assessment phase did not predict the most efficient arrangement within the context of different skills (tacts and intraverbals). It is also worth considering the degree to which outcomes of variations in rein- forcement arrangement are related to skill defi- cits and motivational variables. When motivational variables are influencing consumer responding, the reinforcement arrangement used may be of increased importance and the optimal arrangement may be consumer specific. This may be the case if one reinforcement arrangement promotes unprompted responses by more effectively increasing the value of a reinforcer for that type of response relative to the other reinforcement arrangements. On the other hand, when consumer responding is a function of skill deficits, the reinforcement arrangement used may be of less importance, as unprompted responding related to novel skills has little history with reinforcement. Finally, participants may have had difficulty discrimi- nating among the conditions. We attempted to increase the likelihood of discrimination by incorporating condition-correlated colors; how- ever, this may not have been sufficient. One variable not consistently reported in the

extant assessment-based instruction literature is the time required for assessment. This appears

to be an important variable to consider if clini- cians are to adopt these assessment procedures. One notable exception was Seaver and Bourret (2014), who reported a mean assessment dura- tion of 8 hr. For the current evaluation, the average duration of the assessment across parti- cipants was 3 hr 45 min (range, 2 hr 23 min to 4 hr 4 min). This duration includes the cumulative duration of training across the four conditions included in the assessment phase, but did not include the time required to com- plete other procedural components (e.g., the color and edible stimulus preference assess- ments). The length of the assessment phase is a considerable limitation, particularly considering the outcomes were not generally predictive across skill types. Future studies could avoid this limitation by arranging functional targets (rather than arbitrary targets) during the assess- ment phase. The inclusion of such targets would be beneficial because consumers would acquire functional skills no matter the outcome of the assessment. For example, if all conditions included in the assessment phase are similarly efficient, this outcome may not provide strong evidence for the use of one instructional arrangement over others, but the consumer would have acquired functional skills nonethe- less. In addition, future researchers should eval- uate ways to reduce the duration of the assessment, even if functional targets are included. For example, it may not be necessary for individuals to achieve mastery level responding during a condition in the assess- ment phase. Perhaps evaluating consumer responding during a fixed number of trials (e.g., 50 trials) or instruction time (e.g., 10 min) would be valid alternatives. We also did not collect IOA data for session dura- tion during the assessment phase. This would be an important measure to include in future endeavors. The current study differed from some recent

studies in a couple of ways. First, unlike some previous studies (e.g., Boudreau et al., 2015;

301ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

Karsten & Carr, 2009), we arranged differential reinforcement in the relevant conditions from the onset of treatment. Further research may be needed to evaluate the effects of the timing of the implementation of different arrangements of reinforcement. Second, it is unclear how edi- ble amounts and small-edibles values were selected in previous studies (e.g., Cividini- Motta & Ahearn, 2013; Fiske et al., 2014). Similar to other previous studies (e.g., Boudreau et al., 2015), we used an empirical method for selecting the edible amount and small edible size. Future studies are needed to evaluate procedures for identify- ing different reinforcer qualities and magni- tudes when broader, nonedible classes of stimuli are arranged as reinforcers. For example, it would be important to evaluate how to arrange differential reinforcement when tokens, toys/activities, or social interactions are arranged as reinforcers. Fiske et al. (2014) arranged different durations of access to social interaction or toys for a subset of participants based on teacher report. Additional studies are needed to evaluate procedures for identifying optimal reinforcer qualities and magnitudes. Future research related to identifying the

most efficient differential reinforcement arrangement should consider a number of fac- tors. First, the current study did not collect maintenance data. It is possible that some con- ditions may be more difficult to program for maintenance (e.g., magnitude manipulation) and will therefore affect the efficiency of skill acquisition. Paden and Kodak (2015) demon- strated that a larger magnitude reinforcer may result in more efficient acquisition of a skill (based on number of training sessions to mas- tery) than a smaller magnitude reinforcer for one participant. However, it is possible that longer periods of instructional time may be required to fade the larger magnitude reinforcer following acquisition, thus potentially eliminat- ing the initial efficiency of this arrangement. Future research is needed to evaluate what, if

any, effect differential reinforcement arrange- ments have on skill maintenance. Second, the feasibility and ease of implementation of the different reinforcement arrangements should be taken into account. For instance, a manipula- tion of reinforcement schedule may be rela- tively more difficult for clinicians to use, as it requires the continuous evaluation as to whether the reinforcement requirement (e.g., FR 3) was met across trials, whereas the other conditions do not require such effort. It is unclear how this and other variables may influence treatment integrity. Future studies may also assess consumer preference for one reinforcement arrangement over another. In instances in which multiple conditions are found to be equally effective, an additional evaluation of consumer preference for a condi- tion (Hanley, 2010) may be used as the basis for selecting one arrangement over others.

REFERENCES

Boudreau, B. A., Vladescu, J. C., Kodak, T. M., Argott, P., & Kisamore, A. N. (2015). A comparison of differential reinforcement procedures on the acqui- sition of tacts in children with autism. Journal of Applied Behavior Analysis, 48, 918–923. doi:10.1002/ jaba.232

Carr, J. E., Nicolson, A. C., & Higbee, T. S. (2000). Evaluation of a brief multiple-stimulus preference assessment in a naturalistic context. Journal of Applied Behavior Analysis, 33, 353–357. doi:10.1901/ jaba.2000.33-353

Cividini-Motta, C., & Ahearn, W. H. (2013). Effects of two variations of differential reinforcement on prompt dependency. Journal of Applied Behavior Analysis, 46, 640–650. doi:10.1002/jaba.67

Coon, J. T., & Miguel, C. F. (2012). The role of increased exposure to transfer-of-stimulus-control procedures on the acquisition of intraverbal behavior. Journal of Applied Behavior Analysis, 45, 657–666. doi:10.1901/jaba.2012.45-657

Dunn, M., & Dunn, L. M. (2007). Peabody Picture Vocabulary Test (4thed.). Circle Pines, MN: AGS.

Fisher, W. W., Piazza, C. C., Bowman, L. G., Hagopian, L. P., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying rein- forcers for persons with severe and profound disabil- ities. Journal of Applied Behavior Analysis, 25, 491–498. doi:10.1901/jaba.1992.25-491

KATE A. JOHNSON et al.302

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

Fiske, K. E., Cohen, A. P., Bamond, M. J., Delmolino, L., LaRue, R. H., & Sloman, K. N. (2014). The effects of magnitude-based differential reinforcement on the skill acquisition of children with autism. Journal of Behavioral Education, 23, 470–487. doi:10.1007/s10864-014-9211-y

Freeman, T. J., & Lattal, K. A. (1992). Stimulus control of behavioral history. Journal of the Experimental Analysis of Behavior, 57, 5–15. doi: 10.1901/ jeab.1992.57-5

Grow, L. L., & LeBlanc, L. A. (2013). Teaching receptive language skills: Recommendations for instructors. Behavior Analysis in Practice, 6, 56–75.

Hanley, G. P. (2010). Toward effective and preferred pro- gramming: A case for the objective measurement of social validity with recipients of behavior-change pro- grams. Behavior Analysis in Practice, 3, 13–21.

Hausman, N. L., Ingvarsson, E. T., & Kahng, S. W. (2014). A comparison of reinforcement schedules to increase independent responding in individuals with intellectual disabilities. Journal of Applied Behavior Analysis, 47, 155–159. doi:10.1002/jaba.85

Karsten, A. M., & Carr, J. E. (2009). The effects of dif- ferential reinforcement of unprompted responding on the skill acquisition of children with autism. Journal of Applied Behavior Analysis, 42, 327–334. doi:10.1901/jaba.2009.42-327

Kodak, T., Fisher, W. W., Clements, A., Paden, A. R., & Dickes, N. R. (2011). Functional assessment of instructional variables: Linking assessment and treat- ment. Research in Autism Spectrum Disorders, 5, 1059–1077. doi:10.1016/j.rasd.2010.11.012

Lovaas, O. I. (2003). Teaching individuals with develop- mental delays: Basic intervention techniques. Austin, TX: Pro-Ed.

McGhan, A. C., & Lerman, D. C. (2013). An assessment of error-correction procedures for learners with autism. Journal of Applied Behavior Analysis, 46, 626–639. doi:10.1002/jaba.65

Olenick, D. L., & Pear, J. J. (1980). Differential rein- forcement of correct responses to probes and prompts in picture-name training with severely retarded chil- dren. Journal of Applied Behavior Analysis, 13, 77–89. doi:10.1901/jaba.1980.13-77

Paden, A. R., & Kodak, T. (2015). The effects of rein- forcement magnitude on skill acquisition for children with autism. Journal of Applied Behavior Analysis, 48, 924–929. doi:10.1002/jaba.239

Roane, H. S., Lerman, D. C., & Vorndran, C. M. (2001). Assessing reinforcers under progressive sched- ule requirements. Journal of Applied Behavior Analysis, 34, 145–167. doi:10.1901/jaba.2001.34-145

Seaver, J. L., & Bourret, J. C. (2014). An evaluation of response prompts for teaching behavior chains. Jour- nal of Applied Behavior Analysis, 47, 777–792. doi:10.1002/jaba.159

Sindelar, P. T., Rosenberg, M. S., & Wilson, R. J. (1985). An adapted alternating treatment design for instructional research. Education & Treatment of Chil- dren, 81, 67–76.

Sundberg, M. L. (2008). Verbal behavior milestones assess- ment and placement program: The VP-MAPP. Con- cord, CA: AVB Press.

Touchette, P. E., & Howard, J. S. (1984). Errorless learn- ing: Reinforcement contingencies and stimulus con- trol transfer in delayed prompting. Journal of Applied Behavior Analysis, 17, 175–188. doi:10.1901/ jaba.1984.17-175

Vladescu, J. C., & Kodak, T. (2010). A review of recent studies on differential reinforcement during skill acquisition in early intervention. Journal of Applied Behavior Analysis, 43, 351–355. doi:10.1901/ jaba.2010.43-351

Williams, K. T. (2007). Expressive Vocabulary Test (2nd

edition). Minneapolis, MN: Pearson Assessments.

Yaw, J., Skinner, C. H., Skinner, A. L., Maurer, K., Cihak, D., Wilhoit, B., … Booher, J. (2014). Measurement scale influences in the evaluation of sight-word reading interventions. Journal of Applied Behavior Analysis, 47, 360–379. doi:10.1002/ jaba.126

Received October 23, 2015 Final acceptance November 29, 2016 Action Editor, Bridget Taylor

303ASSESSMENT OF DIFFERENTIAL REINFORCEMENT PROCEDURES

19383703, 2017, 2, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.372 by Purdue U

niversity G lobal, W

iley O nline L

ibrary on [13/11/2025]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense