Acquisition and Generalization of PECS Behaviors in Children with Autism

Laurie S. Hanley and Ann R. Beck
Illinois State University
Normal, IL

Abstract
This study investigated whether children with autism would generalize the use of PECS to untrained settings and across communication partners. An additional question of interest was whether the children's spoken language would increase as a result of using PECS. This study utilized a single subject design with multiple baselines in which 4 childre, ages 7-9 years, were trained to use PECS during a 6-week period. All participants generalized PECS behaviors and increased spoken language.

Narrative Summary
Children with autism often display a variety of social and communicative deficits that preclude their use of functional communication. The presence of unusual social and communicative behaviors presents challenges to professionals attempting to find a functional means of communication and an appropriate approach to implement intervention. Individuals with autism who do not have functional communication skills are often considered to be candidates for augmentative and alternative communication (AAC).

A variety of aided and unaided AAC interventions have been used with individuals with autism with varying levels of success. There are, however, many limitations that may arise when using AAC systems with individuals with autism. Many of the AAC approaches to possible interventions require prerequisite skills such as attentional skills, ability to imitate, and match-to-sample knowledge, in order to be successful. These prerequisite skills are often skills that individuals with autism do not have. Additionally, many AAC interventions often require prompting and use of social reinforcers. Because persons with autism can become prompt dependent, the use of prompts can reduce the occurrence of spontaneous communication and social reinforcers are weak motivators for most individuals with autism.

The Picture Exchange Communication System (PECS) was developed by Bondy and Frost (1994) in response to their observations that children had difficulty acquiring spontaneous communication skills through other AAC techniques. PECS is a training program that intends to teach children self-initiated functional communication. PECS is different than other AAC systems in that it does not require any prerequisite skills, it uses tangible rewards, and verbal prompting is not used. PECS is taught by having the child request desired items by giving a picture of the item to a communication partner in exchange for the tangible item. PECS includes six phases which teach the physical exchange of a picture for an item, persistence in communication, discrimination between pictures, building sentence structure, responding to questions, and commenting.

Previous studies have shown that children with autism have rapidly acquired functional communication and developed speech through PECS training. As a result of success with many individuals, PECS has become a popular AAC system for individuals with autism. While data on the effectiveness of PECS have been reported, most of the literature is anecdotal in nature. Although PECS is commonly used and information on PECS use has been published, there is still a need for empirical research to validate the effectiveness of PECS. Additionally, there is a lack of controlled data regarding the generalization of PECS behaviors across settings and partners.

The purpose of this study was to collect empirically derived data regarding the acquisition and generalization of PECS behaviors. More specifically, this study investigated whether the use of PECS would generalize to untrained settings and across communication partners. An additional question of interest was whether spoken language would increase as a result of using PECS. A single subject with multiple baselines across participants design was utilized in which 4 children with autism were trained to use PECS during a 6-week period.

Participants.
The 4 participants were all males, between the ages of 7 and 9 years, who had a diagnosis of autism. Prior to the initiation of the study, in-depth language and sensory assessments were administered to the participants. They showed varying levels of language ability, however, all participants demonstrated an absence of consistent functional communication. In addition, reinforcer inventories were conducted in order to determine which items the children desired. The participants were randomly assigned to pairs to which intervention sessions were applied.
Baseline. The experimental sessions followed the same protocol for all participants. The order in which each participant’s baseline data were recorded was randomized across participants. Pictures of desired items and the corresponding items were placed on a table and were available to the child. Without any other interaction with the child, the investigator observed and recorded the child’s use of PECS behaviors and productions of spoken verbalizations for 10 minutes. No verbal cues were given and reinforcement was not provided. At least 3 baseline sessions were documented for each participant prior to the initiation of intervention.

Intervention.
Intervention was provided 3 days per week for 6 weeks and consisted of approximately 30 trials (i.e., picture exchanges) per day. Intervention primarily focused on PECS Phases 1, 2, and 3 and was implemented as specified by Frost and Bondy (1994). The only exceptions to the protocol suggested by Frost and Bondy were that intervention was conducted in a discrete trial format and the investigator served as the communication partner during all trials. Intervention was conducted individually with each child, but was started on the same day for both children in a pair. Intervention was initiated for pair 1 after a stable baseline was established for both children. Intervention was initiated for pair 2 after a stable baseline was established for both children and at least one participant in pair 1 had met criterion (i.e., 90% accuracy 2 out of 3 days) to advance to the next level of intervention (Phase 2). Once a participant was in the first level of intervention (Phase 1), continuation through intervention levels (PECS Phases) depended on each individual participant’s ability to meet criterion.

Generalization.
Generalization sessions were conducted twice a week in both the home and school settings. The same procedure was employed for baseline and generalization sessions (i.e., observation for 10 minutes). If, however, the child did initiate a picture exchange, the investigator accepted the picture and gave the child the pictured object. The partner with whom the child exchanged was recorded in order to document generalization across communication partners. The same preferred items were used across baseline, intervention, and generalization sessions. Data on the acquisition of PECS behaviors and production of verbalizations were recorded across baseline, intervention, and generalization sessions. Verbalizations were further coded as either spontaneous or prompted, and intelligible or unintelligible.

Reliability.
Reliability was measured by interrater agreement on a minimum of 20% of all sessions. Procedural reliability was also documented to determine whether intervention sessions were conducted as described in the method of the study.

Results.
All participants acquired PECS behaviors in a relatively short period of time. Within the 6-week period, 3 of the 4 participants advanced to Phase 3 and one participant advanced to Phase 4. Although the amount of generalization of PECS behaviors varied across participants, all participants generalized PECS behaviors to both home and school settings. Moreover, all participants generalized PECS behaviors across communication partners. Additionally, all participants’ verbalizations increased in number, however, increases in spontaneity and intelligibility varied across participants.

Possible reasons why the participants acquired PECS behaviors, generalized PECS behaviors, and increased their spoken verbalizations will be emphasized in the paper presentation. Clinical implications and directions for future study will also be discussed.

PowerPoint Slides

Return to 2003 Recipients