Effects of Nonreciprocal Peer Tutoring With Preschool Students

Abstract

We investigated the effectiveness of nonreciprocal peer tutoring, a type of peer-mediated instructional intervention, with preschoolers. We used a multiple-probe design across 2 dyads with participants with and without disabilities. The dependent variables were the number of correct tact responses to the untaught stimuli for the tutor and the number of correct tact responses to taught stimuli for tutees. Results demonstrated that nonreciprocal peer tutoring was effective in the acquisition of untaught tacts for the tutors and directly taught tacts for the tutees.

Peer-mediated interventions involve one or more peers assisting other classmates academically, interpersonally, or socially. These interventions provide students with disabilities with greater opportunities to learn from peers and to enhance social development in general educational settings (Strain & Kohler, 1999). Peer tutoring is a type of peer-mediated intervention where students teach academic and social behaviors to each other. Peer tutoring has been shown to be effective across different subject areas and grade levels in increasing academic achievement for students with and without disabilities. Studies demonstrated that peer tutoring increases reading fluency and comprehension (Kamps, Barbetta, Leonard, & Delquadri, 1994) and enhances academic responding and test scores (Greenwood, Dinwiddle, Terry, Wade, Stanley, Thibadeau, & Delquadri, 1984). Additionally, researchers have found that peer tutoring can enhance the frequency and duration of social interactions for students with and without autism (Kamps et al., 1994), as well as increase social approvals and decrease disapprovals (Greer & Polirstok, 1982).

The majority of peer-tutoring research has studied its effects on school-age children, with fewer applications reported with preschoolers. The few studies conducted with preschoolers have taught academic skills (Stokes, Doud, Rowbury, & Baer, 1978) or social interactions (McGee, Almeida, Sulzer-Azaroff, & Feldman, 1992). These studies utilized a reciprocal peer-tutoring procedure where each participant functioned as both a tutor and tutee by exchanging roles. However, a recent meta-analysis revealed that nonreciprocal peer tutoring produces better outcomes than reciprocal peer tutoring when school-age participants including elementary, secondary, and college students were taken as a whole (Leung, 2019). Thus, the purpose of this study was to investigate the effects of a nonreciprocal peer-tutoring procedure with preschoolers on the acquisition of skills for both the tutor and the tutee.

Method

Participants and Setting

We selected four participants, one male and three females, from a Comprehensive Application of Behavior Analysis to Schooling (CABAS®) model preschool classroom. All participants ranged in age from 4.1 to 5.3 years old. Dyad 1 was composed of Brianna (5.1) as a tutor and Henry (5.3) as a tutee. Dyad 2 was composed of Amy (4.11) as a tutor and Helen (4.10) as a tutee. All participants were assessed based on their degree of bidirectional naming (BiN), a developmental cusp that signifies when an individual begins to acquire language incidentally (Greer, Stolfi, Chavez-Brown, & Rivera-Valdes, 2005). Brianna, Henry, Amy, and Helen showed 75%, 52.5%, 62.5%, and 72.5% of BiN, respectively. The selection criterion for tutors required that they textually respond to novel words with four to nine letters. Both tutors were typically developing and had no diagnoses. Selection criteria for tutees required that they attend to and follow vocal peer directions without interfering behavior (escape from directions or attention-seeking maladaptive behavior). Both tutees had an educational classification of a preschooler with a disability and had echoics, tacts, and intraverbal vocal verbal behavior in their repertoires.

All participants attended an integrated classroom in a publicly funded private preschool for children with and without disabilities. All sessions took place at a table in the participants’ classroom, with the participant sitting across from the instructor (during probe sessions) or peer (during peer tutoring).

Materials

During all sessions for Dyad 1, the researchers used 7.6 cm × 12.7 cm flash cards with pictures of presidents on one side and the written last name of the president on the back. The presidents included were Abraham Lincoln, Barack Obama, Richard Nixon, George W. Bush, and John F. Kennedy. For Dyad 2, three-dimensional shapes were used that were approximately 10 cm. The three-dimensional shapes included were a sphere, a pyramid, a cube, a hemisphere, and a cylinder. The researchers utilized tokens as consequences for correct responding; later, participants traded in tokens for back-up reinforcers. Back-up reinforcers included classroom activities such as playing in the toy area, coloring, and reading, all of which had been demonstrated, prior to the study, to function as reinforcers for behavior. Further, the researchers used Teacher Performance Rate and Accuracy (TPRA; Ingham & Greer, 1992) forms to mark the accuracy of each antecedent presentation and consequence presentation for the purposes of scoring treatment integrity.

Dependent Variable

The dependent variable was tact responses that corresponded with the discriminative stimulus, emitted within 5 s of the antecedent. During pre- and postintervention probe trials, the researchers recorded the percentage of correct tact responses (out of 10 opportunities) in the presence of corresponding pictures. Stimuli for Dyad 1 were pictures of U.S. presidents, and stimuli for Dyad 2 were three-dimensional shapes. During peer-tutoring sessions, the dependent variable was identical; however, this was only recorded for the tutee, using the same definition of tacts as during probes.

Procedure

Target Stimuli Selection

The researchers identified tacts of stimuli that were absent from both the tutor’s and tutee’s repertoires. The researcher presented pictures of coins, three-dimensional shapes, and U.S. presidents. Each stimulus set consisted of five specific stimuli. The researcher provided 10 trials with two opportunities for each stimulus. For each trial, the researcher presented the stimulus, allowed the student 5 s to respond, and provided no consequences. The researchers selected sets of stimuli that each member of the dyad emitted with 0% accuracy. Dyad 1 emitted with 0% accuracy for tacting three-dimensional shapes, and Dyad 2 emitted with 0% accuracy for tacting U.S. presidents.

Tutor Training

After selecting the stimuli, the researchers provided training to the to-be tutors on how to accurately conduct peer tutoring. The tutors sat across from the tutee and presented 20 trials of mastered tacts (e.g., colors). The researcher sat next to the tutor and tutee and evaluated the tutor’s behavior using the TPRA (Ingham & Greer, 1992). The researcher verbally praised the tutor when the tutor presented a correct instructional trial (antecedent and consequence), and the researchers provided a correction for the tutor contingent on inaccurate delivery of antecedents or consequences. For the correction procedure, the researcher modeled a correct instructional trial, had the tutor re-present the instructional trial, and delivered no consequence for the tutor’s corrected instructional trial. The tutor-training phase continued until the tutor emitted 90% correct teaching behaviors across two consecutive sessions or 100% correct responding for one session.

Once the tutor met criterion on the delivery of peer-tutoring instruction, the researchers taught the tutor to collect data on the tutee’s responses. The process was identical to the instruction presentation training, but the researchers required the tutor to mark a plus (+) for a correct response and a minus (−) for an incorrect response at the end of each trial. The mastery criterion for the second phase of tutor training was 90% correct responding across two consecutive sessions or 100% correct responding for one session.

Probe Assessments

Once the tutors mastered training, the researchers presented target stimuli and assessed the participant’s tacts of stimuli (three-dimensional shapes for Dyad 1 and pictures of U.S. presidents for Dyad 2). The researchers provided 10 probe trials with two opportunities for each of the five stimuli. The tutors and the tutees were required to emit tact responses within 5 s. Researchers provided no feedback.

Textual-responding Training for Tutors

Before the tutor started teaching the tutee, the researchers trained the tutor to textually respond to the target stimuli. The researchers presented flash cards with the stimuli names written in black marker and required the student to textually respond to the word. For each session, the instructor provided 20 trials with four opportunities for each stimulus. The criterion for mastery was 90% correct responding for one session.

Nonreciprocal Peer Tutoring

Once the tutor met criterion on textual responding, the tutor provided peer tutoring for novel stimuli to the tutee. The peer tutor sat across from the tutee and provided 20 trials with four opportunities for each stimulus. For each presentation of a target stimulus, the side with the written names faced toward the tutor, and the side with only the picture of a shape or president with no written words faced the tutees. Then, the tutor presented a stimulus card and recorded the tutee’s response. The tutor provided praise for a correct response and implemented the correction procedure contingent on an incorrect response. For the correction procedure, the tutor modeled the correct response, re-presented the antecedent, and delivered no consequence following the tutee’s response. Mastery criterion was 90% accuracy, demonstrated by the tutee, for one session. For the first session, the researchers sat next to the tutor and tutee to record the tutee’s response for interobserver agreement purposes. For the subsequent sessions, the researcher distanced herself so that the tutor independently taught the tutees. After each session, the tutors brought the data sheet to the researcher.

Experimental Design

The experimenters used a multiple-baseline design across dyads to evaluate the effectiveness of nonreciprocal peer tutoring on the acquisition of novel tacts. Preintervention probe data were compared to postintervention probe data to demonstrate the effectiveness of peer tutoring on the acquisition of untaught tacts. The experimenters started by collecting preintervention data for each participant, and each participant showed zero correct responding prior to the implementation of peer tutoring. Once we obtained zero correct responding for Dyad 1, we implemented the peer-tutoring intervention. After Dyad 1 demonstrated the criterion of the intervention, the experimenter conducted a postintervention probe for these participants and another preintervention probe for Dyad 2. Then, the experimenter began the peer-tutoring intervention with Dyad 2 until they demonstrated the criterion for peer tutoring, after which a postintervention probe was conducted.

Interobserver Agreement and Treatment Fidelity

A second independent observer collected data for the purpose of assessing interobserver agreement (IOA). The researchers calculated the IOA by adding the number of trials with agreement across both observers, dividing by the total number of trials, and multiplying by 100%. The researchers collected IOA data for 57% of all the probe sessions across all four participants. Agreement was 100%. A researcher collected IOA data for 38% of all peer-tutoring sessions, where the tutor’s data were considered the primary data. Agreement was 100%.

The researchers collected fidelity data for the tutor using TPRAs during the tutor-training sessions. For Tutor 1, researchers evaluated the tutor’s behavior during 66% of the training sessions, and fidelity was 98% (range 95%–100%). For Tutor 2, researchers evaluated the tutor’s behavior during 88% of the training sessions and fidelity was 87%, which began at a lower level and ascended to criterion levels (range 60%–100%).

Results

Tutor Training

Tutor 1 met criterion on the delivery of peer-tutoring instruction within two sessions and on data collection after one session. Tutor 2 met criterion on the delivery of peer-tutoring instruction within seven sessions and on data collection after one session.

Nonreciprocal Peer Tutoring

Figure 1 displays performance data before, during, and after peer tutoring. Performances for Dyad 1 are displayed in the top two panels. Prior to peer tutoring, both participants responded with 0% accuracy to probe trials, in addition to this level of accuracy during the target identification phase. During peer tutoring, Henry (second panel) required 10 sessions to meet the mastery criterion. During the post-peer tutoring probe, Brianna (top panel) responded with 100% accuracy to the target stimuli, and Henry responded with 90% accuracy.

Fig. 1.

The performances during probes (bars) and peer tutoring (data paths) for Dyad 1 (top two panels) and Dyad 2 (bottom two panels)

Performances for Dyad 2 are displayed in the bottom two panels. Prior to peer tutoring, both participants responded with 0% accuracy to probe trials, in addition to this level of accuracy during the target identification phase. During peer tutoring, Helen (fourth panel) required six sessions to meet the mastery criterion. During the post-peer tutoring probe, Amy (third panel) responded with 70% accuracy to the target stimuli, and Helen responded with 90% accuracy.

Discussion

The results of our experiment demonstrated the effectiveness of one type of peer-mediated intervention, a nonreciprocal peer-tutoring procedure, on preschoolers with and without disabilities. Two tutors without disabilities taught academic responses to two tutees with disabilities. Following the intervention, all participants, including tutors who received no direct consequences, demonstrated a marked increase in tact responses. Our study extended the existing research on peer tutoring by highlighting the effectiveness of nonreciprocal peer tutoring and peer-mediated intervention on students with disabilities. This study adds to the literature by demonstrating this outcome with preschoolers, which is a population with whom peer tutoring is not frequently implemented. The outcomes suggest that peer tutoring can be implemented with preschool students in such a manner that both the tutor and the tutee gain valuable academic skills (Stokes et al., 1978).

Further, there are possible collateral effects of peer-mediated interventions on social interaction. Studies have shown increased interaction between the tutor and tutee during noninstructional settings, highlighting the benefits of peer tutoring on social interactions (McGee et al., 1992). Additionally, collateral effects of peer-mediated interventions on other constructs, including self-confidence or self-efficacy, support the possible benefits of nonreciprocal peer tutoring (Spencer, 2006). Peers may benefit from peer tutoring or other peer-mediated interventions outside of the targeted training skill.

Although our study successfully demonstrated the effectiveness of a nonreciprocal peer-tutoring procedure on the acquisition of untaught tacts, one limitation of the current study was the small number of probe sessions across both dyads. An increased number of probes would have demonstrated greater control over maturation and instructional history. Another limitation was the variable teaching skills of the tutors, which may have affected the learning of both the tutor and tutee. Brianna had a quiet voice, and this could have negatively affected the reinforcing nature of praise. Also, tutors’ presentation of antecedents and consequences were less fluent compared to the researchers, which may have yielded the tutees’ slower rate of learning because the slower presentation of trials could influence the learner to become distracted and less motivated.

The limitations noted previously suggest future directions of research on effective peer-tutoring methods for preschoolers with and without developmental disabilities. Given the successful implementation of the nonreciprocal peer-tutoring procedure for learning untaught stimuli, future research should study the effectiveness of nonreciprocal peer tutoring in other areas, including social interactions and self-management skills. Also, through peer tutoring, the tutor acquired tact responses to stimuli that were not directly taught. The tutors observed the pictures, observed the responses made by the tutees in the presence of the pictures, and then emitted the responses themselves. Given that the tutors learned through observing the tutees’ responses, future studies should investigate how having observational learning in a repertoire affects the acquisition of novel responses through peer tutoring. Additional sessions after the postintervention probe sessions to assess maintenance and generalization of the skill should be encouraged as well.

Despite such room for improvement, our study successfully demonstrated the effectiveness of a nonreciprocal peer-tutoring procedure on the emergence and acquisition of untaught and taught tact stimuli for preschoolers with and without disabilities. Given that all four participants displayed increased correct tact responses as a function of peer tutoring, we expect to observe peers functioning as each other’s tutors throughout the day and, ultimately, a classroom environment where students foster each other’s learning.

Compliance with Ethical Standards

Conflict of Interest

All three authors claim no conflict of interest.

Ethical Approval

The school’s institutional review board deemed this research as exempt educational research. Each participant’s parent consented to the dissemination of data collected during regular educational activities.