Efficacy of constant time delay implemented by peer ... - Springer Link

2 downloads 0 Views 1MB Size Report
Mark Wolery, Ph.D., 1,4 Margaret G. Werts, Ed.S., 2 Erin D. Snyder, ... KEY WORDS: peer tutoring; constant time delay; inclusive schooling; students with.
Journal of Behavioral Education, Vol. 4, No. 4, 1994, pp. 415-436

Efficacy of Constant Time Delay Implemented by Peer Tutors in General Education Classrooms Mark Wolery, Ph.D., 1,4 Margaret G. Werts, Ed.S., 2 Erin D. Snyder, B.S., 3 and Nicola K. Caldwell, B.S. 3

We evaluated the effects of teaching 13 typically developing children to use constant time delay when tutoring three of their peers who had substantial disabilities. We measured the extent to which the tutors used the constant time delay procedure correctly and the tutees" correct performance on the skills being taught. A multiple probe design across behaviors was used. The results indicated that (a) the tutors used each step of the constant time delay with a high degree of procedural fidelity; (b) the students with disabilities acquired the skills that were taught during peer tutoring," and (c) modifications of the procedure were required for 2 of the 3 students with disabilities, and the peers tutors were able to implement the modifications. Implications for practice and future research are described. KEY WORDS: peer tutoring; constant time delay; inclusive schooling; students with disabilities.

Many children with disabilities currently are served in general education classrooms. Such placements often require school systems and teachers to make adaptations in the manner in which instruction is delivered. Increases in the amount of individualization and flexibility frequently are necessary in the selection of curriculum content, in the use of teaching strategies, in scheduling, and in staffing patterns. Ideally, such adaptations tprofessor, Department of Psychiatry, Medical College of Pennsylvania and Hahnemann University, and Senior Research Scientist, Allegheny-Singer Research Institute, Pittsburgh, PA. 2project Associate, Child and Family Studies Program, Allegheny-Singer Research Institute, Pittsburgh, PA. 3Data Research Technician, Child and Family Studies Program, Allegheny-Singer Research Institute, Pittsburgh, PA.

415 1053-0819/94/12C0-O415507.00/0C 1994HumanSciences Press, Inc.

416

Wolery, Welts, Snyder, and Caldweil

occur without the loss of educational benefits to students in the class who do not have disabilities. Salend (1990) suggests that one barrier to the success of inclusive classrooms is the excessive demand on teachers' time for exceptional learners. Peer tutoring is a potential classroom adaptation for addressing the time demands of inclusive schooling. Peer tutoring has been used in both special and general education classrooms, at many grade levels, and with students having a wide range of abilities. An advantage of peer tutoring is that the proportion of time students are engaged in academic instruction can be increased (Greenwood, Delquadri, & Hall, 1989; Passe & Beattie, 1994). However, most previous studies with peer tutoring have not employed systematic response prompting strategies. Response prompting strategies have a rich history of success in special education settings (Demchak, 1990; Wolery, Ault, & Doyle, 1992). These strategies include procedures such as the system of least prompts (Doyle, Wolery, Ault, & Gast, 1988) and constant and progressive time delay (Handen & Zane, 1987). With adult implementation, these procedures also have been effective in inclusive classrooms--general education classrooms that enroll at least one child with disabilities (Fox & Hanline, 1993; Werts, Wolery, Venn, Demblowski, & Doren, 1994; Venn et al., 1993). A logical extension of the response prompting research is to teach peer tutors to use these strategies. A relatively simple response prompting procedure is constant time delay. The constant time delay procedure is a near-errorless teaching strategy that systematically transfers stimulus control from a prompt to the natural discriminative stimulus. This transfer is accomplished by initially pairing a new stimulus and a task request (e.g., "What is this?") with a controlling prompt (i.e., assistance that ensures the child will respond correctly) (Snell & Gast, 1981). Subsequently, the prompt is faded by systematically inserting a fLxed interval of time between the delivery of the task direction (stimulus presentation) and the delivery of the controlling prompt. Constant time delay procedures have been used to teach useful skills to students of many different ages and with many different disabilities (Wolery, Ault, & Doyle, 1992; Wolery, Holcombe, et al., 1992). In nearly all the constant time delay studies, adults served as the instructor. Two constant time delay studies used peer tutors who had identified disabilities to teach peers who also had disabilities (Koury & Browder, 1986; Telescan, 1990). However, both studies occurred in special education classrooms, were heavily supervised by the teacher, and did not employ typically developing peers as tutors. The purpose of this study was to evaluate peer tutoring using constant time delay in general education classrooms with typically developing students serving as tutors for their peers with disabilities. Training tutors to use response prompting strategies may serve to enhance the effectiveness

Peer Tutoring

417

of the tutoring as well as to provide the tutors with sufficient structure to achieve the desired educational outcomes for the students being taught. In this study, the specific research questions were: (a) Can elementary-school students reliably implement a constant time delay procedure? and (b) Will students with disabilities acquire the skills being taught through peer tutoring using constant time delay?

METHOD

Participants Three students, 2 second graders and 1 fourth grader, from two suburban school districts were selected to receive instruction. All had an identified disability and were included in general education classrooms for the majority of the school day. They were caucasian and from two-parent homes in the middle to lower-middle income range. They possessed adequate auditory and visual acuity for the experimental tasks, and all had corrected vision. They were responsive to verbal requests, imitated their peers, displayed expressive language, read some words, and were generally similar to some other children who participated in studies of constant time delay procedures when implemented by adults. None of these students had a history with constant time delay procedures. Additionally, 13 students served as peer tutors. The tutors were selected from teachers' recommendations based on regular attendance, a history of positive interactions with the student who had disabilities, and the permission from the parent/guardian for participation. Multiple peer tutors were used for each student with disabilities to reduce the time demands on each tutor. Darla was an 8-year-6-month old female with Down's syndrome and a visual impairment resultant from congenital cataracts--she had undergone corneal transplants. Darla was placed in a general education, second grade classroom for the majority of her school day. The remainder of her school day was spent in a resource room. On the WechslerPreschool Primary Scale of Intelligence--Revised (WPPSI-R) (Wechsler, 1974) given at age 5-years-5-months, Darla obtained a full scale IQ of 65 (VIQ = 82, PIQ = 54). At age 7, she was given the Stanford Binet Test of Intelligence-IV (Thorndike, Hagen, & Sattler, 1986) obtaining a full scale IQ of 77. On the Developmental Test of Visual Motor Integration (DTVMI) (Beery, 1982), her age'equivalent was 5-years-2-months. Peer tutors selected for Darla were Connie, Augustine, Clark, and Edith. Darla was taught to name sight words expressively.

Wolery, Werts, Snyder, and Caldweii

418

Sharon was an 8-year-9-month old female with diagnoses of mental retardation and avoidant disorder of childhood and phobic-like behaviors. She was in the same second grade class as Darla, and she also received resource room support. She was tested approximately 6 months prior to the onset of the study. On the Wechsler Intelligence Scale for Children-Ill (WISC-III) (Wechsler, 1993), she scored a full scale score of 54 (VIQ = 52; PIQ = 64). On the DTVM1, Sharon scored in the 5th percentile showing a significant deficit. Her expressive language, measured by the Expressive One-Word Picture Vocabulary Test (Gardner, 1979), showed a significant deficit (below the 1st percentile). Peer tutors for Sharon included Horatio, Patty, Alma, Marion, and Edith. Sharon was taught to name sight words expressively. Vince was a 10-year-11-month old male with a diagnosis of a seizure disorder and mental disability resultant from maternal x-rays at 3 months gestation. Vince went to several classes with other fourth grade students, and the remainder of his day was spent in a special education classroom. A paraprofessional accompanied him to classes. On entrance to kindergarten, he was given the verbal subtests of the WPPSI-R scoring a verbal IQ of 59. On the Kaufman Assessment Battery for Children (K-ABC) (Kaufman & Kaufman, 1983), Vince scored a mental processing composite score of 60 (sequential processing = 56; simultaneous processing = 69). His academic performance measured on the Wide Range Achievement Test (WRAT) (Jastak & Wilkinson, 1984) was below first grade level. The Kaufman Test of Educational Achievement (K-TEA) (Kaufman & Kaufman, 1985) was administered a year prior to the onset of the study. Vince scored an ageequivalent score of 6-years-0-months with stable performance across subtests. Peer tutors selected from Vince's class were Natalie, Wally, Terrell, Rhonda, and Claudine. Vince was taught to point to the correct spelling of words in a receptive task format.

Setting Two general education classrooms were used for the study. One was a second grade classroom with 18 children enrolled, 2 with disabilities (Darla and Sharon). The teacher did not have extra assistance in the class when the study was being conducted, but a special education teaching assistant was assigned to the class for other portions of the day. The sessions occurred at students' desks (60 cm x 90 cm) arranged in pairs. All of the children in the class at the time of the experimental sessions were involved in peer tutoring.

Peer Tutoring

419

The second class was a fourth grade, language arts classroom with 24 children enrolled, four of whom had disabilities (including Vince). All the children moved from class to class. The language arts class lasted approximately 40 minutes. It was staffed by the language arts teacher, a special education teaching assistant, and a student teacher. The sessions occurred at a table (1 m x 3 m) at the back of the classroom with the students seated side by side. The remainder of the class worked with the teacher or on independent activities during the experimental sessions.

Materials The target stimuli were selected individually for each student. Each target student and peer tutor had a book of stimuli that was composed of white cards (10 cm x 15 cm) bound with a black plastic comb binding that allowed for rebinding. The first card in the book had the word, "Ready?" printed on the front. The second and subsequent pages each had one stimulus centered on the page. On the back of each stimulus card, a script was printed for the tutor. For Darla and Sharon, sight words were printed in CG Times black type, 48 point, in lower case letters. For Vince, there was an array of words (six words for the first set of stimuli and four words for the second and third sets) printed in CG times, 24 point, black type in lower case letters. One word was spelled correctly and the remaining words had one letter substitution resulting in an incorrect spelling. The order of stimuli in the books was changed for each session. For probe sessions, cards identical to those used, in instructional sessions were used but the cards were not bound into books. During instruction, a variety of small toys and stickers were used as reinforcers and were delivered at the end of each session.

Experimental Design A multiple probe design across behaviors (Tawney & Gast, 1984) was used to evaluate the effectiveness of the constant time delay procedure implemented by peers in teaching expressive word naming (Darla and Sharon) and recognition of correct spellings (Vince). Initial probe sessions established a pre-tutoring level of performance. Daily progress during tutoring was monitored by the investigator through direct observation of the tutoring sessions. Probe sessions after acquisition of each set of stimuli measured the level of acquisition and maintenance of behaviors.

Wolery, Werts, Snyder, and Caldwell

420

Procedures

General Procedures Initially, the students with disabilities were screened to identify unknown stimuli. T h e peers were trained in one or two sessions to implement the constant time delay procedures. Prior to tutoring, the students with disabilities were probed to ensure that stimuli were still unknown. The children were instructed in one-to-one, daily sessions conducted by one peer tutor each day. Tutors for each child rotated across days. One word pair was taught to criterion before instruction on subsequent pairs was initiated; criterion was 3 of 4 days at 100% correct responding before the prompt. An investigator observed each session and collected data on the tutor's implementation of the constant time delay procedure and on the responses of the student with disabilities. The investigator also provided feedback to the tutors as needed.

Response Definitions and Data Collection Responses during probe conditions were scored as correct, the child said the correct word (Darla and Sharon) or pointed to the correct word choice (Vince) within 4 seconds of the task direction; or incorrect, the child did not respond, indicated that the response was unknown, or said an incorrect word or made an incorrect point. During tutoring, five responses were possible, and the data were collected by the investigator. Correct responses before the prompt were defined as the child saying the correct word (Darla and Sharon) or pointing to the correct choice (Vince) within 4 seconds after the delivery of the task direction. Correct responses after the prompt were defined as the child correctly imitating the tutor's p r o m p t (model). Errors before the prompt were defined as the child saying an incorrect word (Darla and Sharon) or pointing to an incorrect word (Vince) within 4 seconds following the task direction. Errors after the prompt were defined as the child not imitating the tutor's model correctly. No responses were defined as the child not speaking (Darla and Sharon) or not pointing (Vince) within 4 seconds of the tutor's model. D a t a were collected by the investigator on the tutors' fidelity of implementing the constant time delay procedures. T h e investigator noted the o c c u r r e n c e or non-occurrence of each of the planned steps in the constant time delay procedure: ensuring attention, showing the stimulus, delivering the task direction, providing the appropriate delay interval, delivering the

Peer Tutoring

421

controlling prompt (model), delivering the appropriate consequent events, and providing the appropriate inter-trial interval.

Probe Condition Procedures Prior to tutoring, each student was tested individually by the investigator to ensure that the stimuli to be taught were unknown. All six stimuli for each student were intermixed during probe sessions. For Darla and Sharon, three probe sessions occurred over two days with each session consisting of 18 trials (3 trials • 6 stimuli). Probe conditions were repeated each time a child reached criterion on a set of stimuli. Vince was probed on the following schedule: two full probes over two days with two trials per 6 stimuli (Sets 1, 2, and 3). In an additional session, trials were implemented for the set to be in instruction immediately following the probe and for the set upon which Vince had just reached criterion. These were conducted to increase the n u m b e r of trials on each stimuli. Each probe trial was conducted in the following manner: The investigator placed the card in front of the student, presented an attentional command (i.e., "Look."), secured the student's attention, delivered the task direction (i.e., "What's this?" for Darla and Sharon or "Which word is ?" for Vince), and provided a response interval. Responses during probe conditions were not reinforced; periodically a non-committal answer (i.e., " G o o d working.", "O.K.", or " T h a n k you for answering.") was delivered.

Tutor Training In the second grade class, the training was conducted as a large-group, whole-class activity. In the fourth grade class, the five students selected as tutors were trained by the investigator in a small group. In both cases, tutors were given a brief verbal introduction to the constant time delay procedure. Samples of the materials (stimuli books, pencils, and data sheets) were displayed and explained. Next, the procedure for each trial was demonstrated using the stimulus booklets and the students were given time to practice individual trials with the investigator and with each other. At the end of the training session, tutors were allowed to ask any questions or express any concerns. All of the tutors could correctly perform the target behaviors to be taught to the students with disabilities.

Wolery, Wefts, Snyder, and Caldwell

422

Peer Tutoring Procedures Once the peer tutors mastered the procedure, the peer tutoring was initiated on the first set of behaviors. The tutors used the constant time delay procedure in which they had been trained. The session began when the tutor asked the student with disabilities to indicate he or she was ready. The tutor then turned the first page in the booklet and followed the script to direct attention to the stimulus, deliver the task direction, provide the appropriate response interval, deliver the prompt when appropriate, and then praise correct responses and ignore errors. Darla and Sharon each received 12 trials (6 trials on each of two stimuli) for the first set of behaviors. In subsequent sets, Darla received 2 review trials and Sharon received 4 review trials in addition to the 12 target trials. Vince received 16 trials (8 trials on each of two stimuli) for all sets. The first two sessions in each tutoring condition involved no delay of the tutor's model (i.e., 0-second trials). The subsequent sessions in each condition used a 4-second interval. The tutors' scripts were adjusted for the 4-second interval by inserting "l-banana, 2-banana, 3-banana, 4 - b a nana" after the task direction. The tutors were instructed to read that portion of the script silently and to deliver the controlling prompt only if the child had not responded. If the child responded correctly before or after the prompt, the tutors praised the child and continued to the next trial. If the child responded incorrectly or did not respond, the tutors ignored the error and proceeded to the next trial.

Review Trials Darla received 2 review trials, intermixed into the target trial sequence, during instruction for Sets 2 and 3. During instruction for Set 2, she received one trial per day on each of the Set 1 stimuli. During instruction for Set 3, she received 2 trials (1 trial per stimuli) each day alternating stimuli from Set 1 and stimuli from Set 2. Sharon received 2 review trials, intermixed into the target trial sequence, during the first four sessions of instruction on Set 2. She made a consistent error on one of the stimuli; therefore, the number of review trials was increased to 4 (2 trials x 2 stimuli) until she reached criterion on Set 2. During Set 3 tutoring, the four stimuli from Sets 1 and 2 were intermixed as review trials. Vince did not receive any review trials during tutoring sessions.

Peer Tutoring

423

Reliability Interobserver agreement data were collected during tutoring sessions on two measures: tutor's procedural fidelity with the constant time delay procedures (Billingsley, White, & Munson, 1980) and responding by the student with disabilities. Interobserver agreement percentages were calculated using the point-by-point method in which the number of exact agreements was divided by the number of agreements plus the number of disagreements and the quotient was multiplied by 100. At least 26% of the sessions for each child were scored by two observers: 26.3% for Darla, 28.4% for Sharon, and 36.0% for Vince. For student responding, interobserver agreement was 99.0%. For the tutor's procedural fidelity, at least 20.0% (range 20.0-38.5%) of the sessions for each tutor were scored by two observers. All the sessions were at 100% agreement for ensuring attention, presenting the stimuli, and delivering the task direction. The observers agreed on 97.5% of the trials for providing the appropriate delay interval, 99.6% for providing the controlling prompt, 96.3% for providing the appropriate consequent event, and 99.6% for providing the correct intertrial interval. Investigator procedural fidelity during probe sessions also was assessed on 25% of the probe sessions. All steps were used correctly on all trials.

RESULTS

Peer Tutors' Fidelity with the Constant Delay Procedure The fidelity with which the tutors used the constant time delay procedure was assessed during each session with each tutor. The results are shown in Table 1 for each tutor and each step of the constant time delay procedure. The percentage of correct use of the procedure was consistently high. For Darla, all tutors implemented each step correctly on more than 95% of the trials. For Sharon, 3 of 5 tutors implemented each step correctly on more than 95% of the trials. The two remaining tutors did so on all steps with two exceptions: Patty correctly delivered the attending cue (described later) on 93.7% of the trials, and Marion delivered the appropriate consequent event on 90.6% of the trials. For Vince, 2 of 5 tutors delivered each step correctly on at least 95% of the trials. The remaining three tutors did so on all steps with the following exceptions: Natalie delivered the correct delay interval on 87.5% of the trials, Wally delivered the correct delay interval on 93.8% of the trials, and Rhonda delivered the correct delay interval on 94.4% of the trials and delivered the appropriate consequent

Darla Conme Augustine Clark FAith Sharon Horatio Patty Alma Marion FAith Vince Natalie Wally Terrell Rhonda Claudine

Student Tutor 100.0 100.0 98,5 100.0 100.0 100.0 100.0 100.0 100.0 100.0 99.4 100.0 98.1 99.0

100.0 100.0 100.0 100.0 100.0 100.0 99.4 100.0 98.1 99.0

Show Stimulus

98.5 98.5 98.5 100.0

Ensure Attention

------

I00.0 93.7 100.0 100.0 --

-----

Deliver Attending Cue

100.0 98.3 100.0 98.1 99.0

100.0 100.0 100.0 100.0 100.0

100.0 100.0 98.5 100.0

Deliver Task Direction

87.5 93.8 96.4 94.4 95.7

100.0 97.3 99.2 99.4 98.8

100.0 100.0 98.5 100.0

Wait Delay Interval

Table 1. Tutors' Procedural Fidelity

100.0 97.7 98.7 98.8 99.0

99.1 98.1 99.5 99.1 99.6

100.0 97.0 98.5 100.0

Deliver Prompt

100.0 95.5 98.2 86.3 98,1

I00.0 98.6 98.9 90.6 95.3

100.0 100.0 98.5 96.3

Deliver Consequences

100.0 99.4 99.5 96.9 100.0

100.0 100.0 100.0 100.0 100.0

100.0 100.0 100.0 100.0

Wait Intertrial Interval

80 is.

r

gg

gl

es_ r

r

,la,. ItO

Peer Tutoring

425

Table 2. Number of Sessions to Criterion, Percentage of Errors Made, and Time of Instruction for Peer Tutoring Using Constant Time Delay

Student

Stimulus set

# of sessions through criterion

% of errors current set/ added review

Time of instruction (Min:Sec)

Darla Set 1-push; girls Set 2-pull; danger Set 3-exit: boys

5 8 6

0.0 2.1/0.0 0.0/0.0

9:21 20:23 14:06

Set 1-seven; three Set 2-five; nine Review Sets 1 & 2 Set 3-six; eight

27 16 26 26 a

1.8 2.1/14.1 2.8 3.5/5.7

79:22 33:55 57:35 47:44

Set 1-pencil; crayon Set 2-calendar; notebook Set 3-orange; surprise

27 9 14

9.0 3.5 1.2

Sharon

Vince

108:21 28:30 32:22

alnstruction halted at the end of the school year. Sharon did not reach criterion.

events on 86.3% of the trials. The steps that were delivered with the lowest overall procedural fidelity were waiting the correct delay interval and delivering praise. The percentage of correct implementation by the secondgrade tutors was equal to or better than that of the fourth-grade tutors. Effectiveness of Peer Tutoring with Constant Time Delay The probe and tutoring data across sets of sight words for Darla and Sharon and across sets of receptive spelling of words for Vince are shown in Figs. 1, 2, and 3, respectively. The open circles represent the percentage of correct responses after the prompt; closed triangles represent the percentage of correct responses before the prompt. The number of sessions to criterion, percentage of errors, and total instructional time for each student are shown in Table 2. The minimum number of sessions per condition was 5 (2 sessions at 0-second delay and 3 sessions to demonstrate criterion level performance).

Darla Prior to tutoring on each set of sight words, Darla performed at 0% correct responses on all stimuli. She achieved criterion level responding in

t

9

1-3

9

9 j J/ \ \ b

4-8

--o--o

Tutoring

9-11

Probe 2

Sessions

12 - 19

Tutoring ~,

20 - 22

A_

Probe 3

A

9

0

23 - 28

Tutoring

O

9

0

9

O

9

0

9

29 - 31

/

Probe 4

Set 3

Set 2

Set 1

Fig. 1. The percentages of correct responses before the prompt (triangles) and correct responses after the prompt (circles) for Darla across probe and tutoring conditions.

100 90 80 70 60 50 40 30 20 10 0

11O0 90 80 70 60 50 40 30 20 10 0

9O 8O 70 6O 50 4O 30

Probe 1

n~

en

O

427

Peer Tutoring

O

C

O

e,

o o o

o

o ,9

o

o o

~m

9

.o~

o

o

~m

o

o o

o o

o

,.~o

8

.

~,~ O

O

~O

~SE8

~ 0

~

~

Eo~.=

428

Wolery, Werts, Snyder, and Caldwell

4

s163 o

8 ~o~

P.E

8._=

p

0

0

8~

8~ ~'.s ~

~

--I e~ o

~

o

Ee-

~s

c~

&:-=

8~ o

U

i.,

e~

E

$

,4 8

~._~

sosuodo~l ~o.uo~) ~o ~ua3~d

Peer Tutoring

429

the minimum number of sessions for Set 1. No modifications were required and she maintained 100% correct responding on Set 1 during Probe II. She met criterion on Set 2 in 8 sessions and initially displayed variable performance on review trials of Set 1. During Probe III, she had 100% correct responses on Sets 1 and 2. She achieved criterion on Set 3 stimuli in 6 sessions. During Probe IV, she had 100% correct responding on Sets 1 and 3 during all sessions and during the last two sessions on Set 2. H e r percentage of errors during tutoring was quite low, and no errors occurred for Sets 1 and 3. The mean daily duration of the tutoring sessions was 2.3 minutes.

Sharon Prior to tutoring on each sight word set, Sharon had 0% correct responses. During the first 14 sessions of tutoring, Sharon consistently waited for the tutor's prompt and then imitated it correctly; thus, the percentage of correct responses after the prompt was high but unprompted correct responses were not occurring. T o encourage her to initiate correct responses before the prompt, trials were inserted on a known stimulus (i.e., her name). During the next four sessions, she responded correctly before the prompt to her name but not to the target stimuli. Thus, an attending cue (tracing the first letter of each word on the stimulus card with her finger) was added at the beginning of each trial; interspersal of the known stimulus continued. This modification resulted in high levels of correct responses before the prompt. During the last two sessions, the attending cue was eliminated. During Probe II, her correct responses on Set 1 was below criterion level performance but higher than Probe I performance; her performance on Set 2 and 3 stimuli remained at 0% correct. Tutoring was initiated on Set 2, and review trials on Set 1 stimuli were interspersed with Set 2 trials. She demonstrated criterion level performance on Set 2 in 16 sessions; however, her performance on Set 1 stimuli during review trials demonstrated correct performance on one stimulus and incorrect or waiting behavior on the other stimulus. During Probe III, she performed correctly on all trials for 2 of the 3 sessions for Set 2 stimuli but was at 50% correct on Set 1 stimuli. As a result, a condition was implemented where tutoring was provided with Set 1 and 2 stimuli intermixed. After the fourth session of this condition, an attending cue (tracing the first letter of the word on the stimulus card with her finger) was added for all stimuli. After the eighth session of this condition, the attending cue described above was used for the two Set 2 stimuli

430

Wolery, Wefts, Snyder, and Caldwell

and one Set 1 stimulus, but a different attending cue was added for the stimuli to which she was not responding correctly before the prompt. This second attending cue involved tracing each letter of the word on the stimulus card with her finger. After twelve sessions using these two attending cues, contingent reinforcement was delivered only for correct responding before the prompt; this resulted in a high level of correct responding on the Set 1 stimuli. Starting with the 15th session, the attending cues were removed, and criterion level performance on Sets 1 and 2 were demonstrated in 26 total sessions. During Probe IV, she demonstrated 100% correct responses on Sets 1 and 2 and 0% correct on Set 3. Tutoring was initiated on Set 3 stimuli and review trials were continued on Set 1 and 2 stimuli. After 10 sessions of this condition, reinforcement was made contingent only on correct responses before the prompt. Although she demonstrated 100% correct responding on Set 3 on three separate sessions, she did not demonstrate criterion level performance. The condition was ended due to the impending close of the school year, and Probe V was initiated. During Probe V, Sharon demonstrated 100% correct responding on Sets 1 and 2 and demonstrated correct responding on Set 3 above that of previous probe conditions. Sharon's percentage of errors was low (range of 1.8-3.5% per set) on the sets of stimuli being taught but was higher on review trials (5.714.1% per set). The mean daily duration of tutoring sessions was 2.3 minutes.

l/ince Since Vince's task was a receptive task, some correct responses occurred during probe conditions prior to tutoring. However, during Probe I, correct responding was below chance levels. Tutoring was initiated on Set 1 in a 6-choice task. After the two sessions of 0-second trials, the 4second delay trials were implemented. The first four sessions with a 4-second delay resulted high levels of errors b e f o r e the prompt. Thus, reinforcement at the end of the session was made contingent upon correct performance (before or after the prompt); in addition, a further explanation of the purpose of waiting for the tutor to deliver the prompt was provided by the investigator. These modifications resulted in high levels of correct responses after but not before the prompt. To determine whether Vince could make the discrimination, a single probe session was conducted with a 2-choice format. He chose the correct spelling on 87.5% of these trials. Tutoring was reinstated with differential reinforcement provided only for correct responses before the prompt. Nine sessions of this condition re-

Peer Tutoring

431

suited in a steady increase in correct responses before the prompt, but no sessions of 100% correct responses before the prompt. Thus, the task was simplified to a 4-choice task, and criterion was demonstrated in five additional sessions. During Probe II, Vince demonstrated 100% correct responses on Set 1, and below chance level responding on Sets 2 and 3. Tutoring was initiated on Set 2 stimuli in a 4-choice task and resulted in criterion level performance in nine sessions. No additional modifications were required. During Probe III, Vince had 100% correct responding on Set 1, 100% correct responding on 2 of 3 sessions for Set 2 stimuli, and below chance level responding on Set 3 stimuli. Tutoring was initiated on Set 3 stimuli in a 4-choice format. Criterion level performance was demonstrated in 14 sessions. During Probe IV, correct performance on Set 1 stimuli was above probe levels, but was at 100% on only one of five sessions. Vince demonstrated 100% correct performance on 4 of 5 probe sessions for Set 2 stimuli and for all sessions of Set 3 behaviors. A generalization probe session was conducted with a 6-choice format. Vince's percentage of correct responses was 100 for Set 1, 75 for Set 2, and 100 for Set 3. The percentage of errors for Vince was high (9.0%) during the tutoring on Set 1. During tutoring sessions on Sets 2 and 3, the percentage of errors was low, 3.5% and 1.2% respectively. The mean daily duration of the tutoring sessions was 3.4 minutes.

DISCUSSION The purpose of this study was to evaluate the effects of a constant time delay procedure implemented by typically developing children to tutor their classmates who had substantial disabilities in general education classrooms. A notable finding was that the second and fourth grade students who served as tutors implemented the constant time delay procedure with a high degree of procedural fidelity. The training procedures for the peer tutors in this study were similar (with modifications in language) to those used for adults in previous studies with one major exception: The tutors in this study were not asked to read descriptions of the procedure prior to training. Data from about 30 previous adult-implemented constant time delay studies indicated that high levels of correct use were common (i.e., range across studies of 98.0-100% for all assessed steps) (see Wolery, Holcombe, et al., 1992 for a review). Interestingly, for adult-implementation, the step with the lowest level of correct use was providing the appropriate delay interval. This step also presented difficulties for the tutors in this

432

Wolery, Werts, Snyder, and Caldwell

study; but, with the exception of three peer tutors, the correct delay interval was used on more than 95% of the trials. Results of an earlier adult-implemented study comparing high and low levels of correct implementation of the delay interval indicate that the effectiveness of the constant time delay procedure is not greatly influenced by low levels of procedural fidelity on that s t e p - - g i v e n that other steps are correctly implemented (Wilbers & Wolery, 1994). Thus, if the delay interval is too long on some trials, too short on other trials, and correctly implemented on still other trials, children's learning is not likely to be affected. On the other hand, failure to use the prompt when needed may interfere substantially with the effectiveness and efficiency of the constant time delay procedure for some children (Holcombe, Wolery, & Snyder, 1994). In this study, all peer tutors implemented the prompt on at least 97% of the trials in which it should be implemented. In addition to implementing the basic constant time delay procedure correctly, the tutors correctly implemented modifications of the procedure such as using attending cues, delivering differential reinforcement of correct responses before the prompt, and employing changes in the format of the stimuli (as with Vince). No decrements in the fidelity of implementation occurred with the use of these modifications. Qualifications to the above finding should be noted. All of the peer tutors demonstrated the ability to read aloud. Scripts for each trial were printed on the back of each stimulus card. The tutors relied on these scripts and delivered the trials in a relatively static manner. All modifications in the procedures required changes in the scripts. Because of these factors, the high level of implementation may not occur when tutors cannot read or when printed scripts are not available. Thus, future research should evaluate fading the use of the scripts. The tutors in the second grade class (those for Darla and Sharon) were described by their teacher as "a good class" and the number of children in this class was small (i.e., 18). Based on the investigators' observations, the teacher was organized and maintained high levels of control, and the children were consistently compliant and engaged in almost no disruptive behavior. Thus, the findings of this study may not be applicable to larger primary-grade classrooms in which the teacher does not have control of the children's classroom behavior. It is also important to note that all tutoring sessions were observed by the investigators. This observation was used to collect data on the tutors' implementation as well as students' learning. Nonetheless, the consistent surveillance may have contributed to the high levels of correct implementation, and generalizations may not be warranted to situations in which such observation does not occur. Future studies should focus on tutors' level of implementation when such data are collected covertly. Similarly, the effects

Peer Tutoring

433

of this observation on students' learning are not known; thus, this is a suitable question for future studies. In many of the adult-implemented constant time delay studies, the adult instructor collected data on students' performance. In the current study, the students were given materials with which to collect data; however, the data on student performance was collected by the adult observer rather than by the child tutor. The training sessions for the tutors did not emphasize data collection, and they did not receive feedback on their data collection. Some of the tutors did not record any responses, some began to do so and would stop during sessions, and others recorded only whether the response was correct before the prompt. In practice, teachers can collect the data on students' learning from daily or periodic probes. Future studies could focus on procedures for teaching peer tutors to collect reliable data on the five potential responses of the constant time delay procedure. Another notable finding is that the constant time delay procedure implemented by peer tutors was effective in teaching children with disabilities to read words and to choose the correct spelling of words. Previous research indicates that adult-implemented constant time delay was effective with similar students (in age and in disability) in teaching similar behaviors (e.g., Ault, Gast, & Wolery, 1988). The error percentages for the participants in this study were low and were within the ranges of the error percentages found for adult-implemented constant time delay studies (see Wolery, Holcombe, et al., 1992, for a review). Because the tutor-implemented constant time delay procedure resulted in relatively low error percentages (see Table 2), the tutors had many occasions to praise their peers. Thus, the nature of the procedure appears to set the stage for high levels of positive exchanges between tutors and tutees and seems to minimize the occasions on which the tutor would provide corrective or negative feedback to the tutees. In the current study, modifications in the constant time delay procedure were required for Sharon and Vince. These modifications included using specific attending cues and differential reinforcement of correct responses before the prompt. Such modifications also have been necessary and successful in adult-implemented constant time delay studies. For example, Doyle, Wolery, Ault, Gast, and Wiley (1989) used an attending cue; and Gast, Ault, Wolery, Doyle, and Belanger (1988) used differential reinforcement of correct responses. In this study, the tasks taught required discrete responses (naming and pointing); however, adult-implemented constant time delay also has been effective with chained tasks (e.g., Griffen, Wolery, & Schuster, 1992). Future research should focus on peer tutors' use of constant time delay to teach chained tasks.

434

Wolery, Wefts, Snyder, and Caldwell

Interviews with the teachers at the end of the study indicated that they believed the procedure was effective. They stated that the amount of class time used was minimal and did not detract from planned activities. However, they indicated that the amount of advance preparation of the materials (booklets with scripts) could be excessive. In this study, the investigators rather than teachers prepared these materials. They commented that the consistent praise and provision of prompts used with the constant time delay had benefits for the students. Both teachers had used peer tutoring or partnering previously. In summary, this study indicates that typically developing peers can learn to use constant time delay reliably within the limitations described above. Further, when children used this structured approach to tutoring, their classmates with disabilities acquired the skills that were taught. The amount of class time needed is relatively minimal, and the percentage of errors was generally low. Based on these findings, we recommend that if teachers wish to employ peer tutoring then they should consider teaching the tutors to use the constant time delay procedure.

ACKNOWLEDGMENTS This study was supported by the U.S. Department of Education, Office of Special E d u c a t i o n and R e h a b i l i t a t i v e Services ( G r a n t N u m b e r H086D20005). However, the opinions expressed do not necessarily reflect the policy of the U. S. Department of Education, and no official endorsement should be inferred. The authors are grateful for the assistance provided by Dr. John Esaias, Superintendent, North Hills School District; Mrs. Sue Palfey, Principal, West View Elementary School; Nancy Guzan, teacher at West View Elementary School; Dr. Frank Todd, Psychologist, Pine-Richland School District; Dr. Richard Psych, Principal, Hance Elementary School; and Betty Balla and Cheryl Raymer, teachers at Hance Elementary School.

REFERENCES Ault, M. J., Gast, D. L., & Wolery, M. (1988). Comparison of progressive and constant time-delay procedures in teaching communitysign word reading. American Journal on Mental Retardation, 93, 44-56. Beery, K. (1982). Revised Developmental Test of Visual-Motor Integration. Cleveland: Modern Curriculum Press. Billingsley,F. F., White, O. R., & Munson, R. (1980). Procedural reliability:A rationale and an example. Behavioral Assessment, 2, 229-241. Demehak, M. (1990). Response prompting and fading methods: A review.American Journal on Mental Retardation, 94, 603-615.

Peer Tutoring

435

Doyle, P. M., Wolery, M., Ault, M. J., & Gast, D. L. (1988). System of least prompts: A review of procedural parameters. Journal of the Association for Persons with Severe Handicaps, 13, 28-40. Doyle, P. M., Wolery, M., Ault, M. J., Gast, D. L., & Wiley, K. (1989). Establishing conditional discriminations: Concurrent versus isolation-intermix instruction. Research in Developmental Disabilities, 10, 349-362. Fox, L., & Hanline, M. F. (1993). A preliminary evaluation of learning within developmentally appropriate early childhood settings. Topics~in Early Childhood Special Education, 13, 308-327. Gardner, M. (1979). Expressive One-Word Picture Vocabulary Test. Austin, TX: Pro-Ed. Gast, D. L., Ault, M. J., Wolery, M., Doyle, P. M., & Belanger, S. (1988). Comparison of constant time delay and the system of least prompts in teaching sight word reading to students with moderate retardation. Education and Training in Mental Retardation, 23, 117-128. Greenwood, C. R., Delquadri, J. C., & Hall R. V. (1989). Longitudinal effects of classwide peer tutoring. Journal of Educational Psychology, 81, 371-383. Griffen, A. K., Wolery, M., & Schuster, J. W. (1992). Triadic instruction of chained food preparation responses: Acquisition and observational learning. Journal of Applied Behavior Analysis, 25, 193-204. Handen, B. L., & Zane, T. (1987). Delayed prompting: A review of procedural variations and results. Research in Developmental Disabilities, 8, 307-330. Holcombe, A., Wolery, M., & Snyder, E. (1994). Effects of two levels of procedural fidelity with constant time delay on children's learning. Journal of Behavioral Education, 4, 49-73. Jastak, S., & Wilkinson, G. (1984). Wide Range Achievement Test--Revised. Wilmington, DL: Jastak Associates. Kaufman, A., & Kaufman, N. (1983). The Kaufman Assessment Battery for Children. Circle Pines, MN: American Guidance Service. Kaufman, A., & Kaufman, N. (1985). The Kaufman Test of Educational Achievement. Circle Pines, MN: American Guidance Service. Koury, M., & Browder, D. M. (1986). The use of delay to teach sight words by peer tutors classified as moderately mentally retarded. Education and Training of the Mentally Retarded, 12, 252-258. Passe, J., & Beattie, J. (1994). Social studies instruction for students with mild disabilities: A progress report. Remedial and Special Education, 15, 227-233. Salend, S. (1990). Effective mainstreaming. New York: Macmillian. Snell, M. E., & Gast, D. L. (1981). Applying time delay procedure to the instruction of the severely handicapped. Journal for the Association for the Severely Handicapped, 6(3), 3-14. Tawney, J. W., & Gast, D. L. (1984). Single subject research in special education. Columbus, OH: Merrill. Telescan, B. L. (1990). Peer tutoring: Training students with learning disabilities to deliver time delay instruction. Unpublished master's thesis, University of Kentucky, Lexington. Thorndike, R., Hagen, E., & Sattler, J. (1986). Stanford-Binet Intelligence Scale. Chicago: Riverside Publishing. Venn, M. L., Wolery, M., Werts, M. G., Morris, A., DeCesare, L. D., & Cuffs, M. S. (1993). Embedding instruction in art activities to teach preschoolers with disabilities to imitate their peers. Early Childhood Research Quarterly, 8, 277-294. Wechsler, D. (1974). Manual for the Wechsler Preschool and Primary Scale of Intelligence-Revised. New York: The Psychological Corporation. Wechsler, D. (1993). Wechsler Intelligence Scale for Children-IlL New York: The Psychological Corporation. Wilbers, J. S., & Wolery, M. (1994). Effects of high and low procedural fidelicy during the delay interval of constant time delay. Manuscript submitted for publication. Werts, M. G., Wolery, M., Venn, M. L., Demblowski, D., & Doren, H. (1994). Effects of transition-based teaching with instructive feedback in mainstreamed kindergarten classrooms. Manuscript submitted for publication.

436

Wolery, Werts, Snyder, and Caldwell

Wolery, M., Ault, M. J., & Doyle, P. M. (1992). Teaching students with moderate to severe disabilities. New York: Longman. Wolery, M., Holcombe, A., Cybriwsky, C. A., Doyle, P. M., Schuster, J. W., Ault, M. J., & Gast, D. L. (1992). Constant time delay with discrete responses: A review of effectiveness and demographic, procedural, and methodological parameters. Research in Developmental Disabilities, 13, 239-266.2

Suggest Documents