Project Activities
The research team started by identifying articles focused on the use of AAC for individuals with autism and/or IDD. For articles that met inclusion criteria, the research team evaluated the quality of research designs and evidence to determine whether AAC can be considered an evidence-based practice. Finally, the research team conducted several meta-analyses of the effects of AAC-based interventions on social/communicative and behavioral outcomes for students with autism and/or IDD and whether effects were moderated by instructional context and intervention targets. . Meta-analyses were based on multiple effect size metrics with complementary strengths, including log response ratio and Tau-AB.
Structured Abstract
Setting
The research articles reviewed in this systematic review and meta-analysis included studies conducted in natural settings, including the home, classroom, community, place of employment, or clinic.
Sample
The research articles reviewed in the systematic review and meta-analysis included individuals ages 0-21 with IDD, including autism spectrum disorder (ASD), intellectual disability (ID), other developmental disabilities (DD), and complex communication needs (cannot communicate via age-expected speech).
The malleable factors under investigation were variables related to instruction in AAC (i.e., participant characteristics, type of instructional protocols implemented, and targeted outcomes). AAC includes any means of communication intended to replace or supplement communication, particularly speech, in individuals with complex communication needs.
Research design and methods
Researchers used systematic review and meta-analysis, a method of quantitative research synthesis that uses prior study results as the unit of observation. First, they conducted a literature search that focuses on the use of AAC among individuals with disabilities. To identify studies, researchers: (1) searched electronic databases (i.e., Academic Search Complete, ERIC, PsycINFO, Conference Proceedings Citation Index — Social Science & Humanities [Web of Science], and Proquest Dissertations & Theses Global); (2) searched reference lists of identified articles and systematic reviews of AAC; and (3) reviewed additional publication of first authors of identified articles and those who regularly publish single-case experimental design research related to ASD/IDD and AAC. Studies were included if they (1) included at least one participant with an IDD, ASD, ID, or other DD; (2) used a single-case experimental design or between-group design; (3) reported the results of an AAC intervention and, for group design studies, include a comparison to a no-treatment or business-as-usual control condition; (4) included social-communicative and/or behavior outcomes; (5) involved implementation of AAC in natural settings, including home, classroom, community, employment, or clinic; and (f) were in English. Published and unpublished manuscripts were included in the review. Second, for studies that meet inclusion criteria, researchers evaluated the quality of the designs and evidence based on the What Works Clearinghouse standards, the Council for Exceptional Children's standards, and other sources. Third, they reviewed the quality of the literature base of single-case and group design studies on the use of AAC among individuals with IDD. Fourth, they conducted meta-analyses to examine effect sizes describing AAC outcomes for individuals with IDD and whether effects vary by participant characteristics; instructional contexts, including setting (e.g., home, classroom, clinic), type of instructional features implemented (e.g., reinforcement, modeling, prompts, prompt fading) and behavioral intervention strategies (e.g., child or interventionist initiated, dispersed versus massed teaching opportunities, contrived versus embedded activity contexts); and targeted outcomes and communication modes implemented, including number and type of communication modes employed, communicative functions taught, and verbal/vocal versus non-verbal/vocal outcomes selected for treatment.
Control condition
Due to the nature of the research design, there was no control condition.
Key measures
The outcomes in this systematic review and meta-analysis were the effect sizes for each eligible study.
Data analytic strategy
The researchers coded eligible articles for methodological issues (related to population and sample, design quality, measurement quality, data quality and missing data, and statistical method) and malleable and moderating factors, including participant characteristics, instructional protocol, and targeted outcomes. Effect sizes were calculated for each eligible study. Multi-level meta-analysis regression models using separate intercepts for each level of the moderator were used to investigate the magnitude of effects of AAC interventions and factors that moderate these effects. Separate analyses were conducted for each of the single-case design effect sizes and for group design effect sizes. For single-case experimental design studies, both log response ratio and Tau-AB effect size metrics were calculated using available data. For group design studies, standardized mean difference effect sizes and associated sampling variances were calculated from reported summary statistics.
Key outcomes
The main findings of this study, as reported by the principal investigator, are as follows:
- Single-case and group design studies were reviewed for quality of reporting. Recent single-case studies adequately reported some quality indicators (defining outcome variables, describing materials, describing baseline and intervention procedures, and evaluating procedural integrity), but reported others poorly or sparsely (participant descriptions, settings, maintenance, generalization, and social validity). Across the literature base, many participants were reported to use at least 2 communication modes (e.g., AAC and vocalizations). Single-case studies were primarily implemented in natural settings (i.e., home and classroom), while group studies relied more heavily on clinical/research settings. Behavioral instructional features and behavioral strategies were implemented more frequently than naturalistic strategies across designs. Race, ethnicity, and home language of participants were infrequently reported.
- Effect sizes in single-case studies were highly variable, indicating a high degree of variation in effects across individuals. Few statistical differences were identified based on participant characteristics. Outcomes were higher for participants who used formalized AAC prior to intervention than those who relied on verbal/vocal or gestural communication. The highly heterogeneous results indicate a need to individualize AAC interventions based on individual participant needs.
- Analyses of instructional contexts of single-case studies indicated that most studies occurred in classroom or clinical settings. The most common instructional features implemented were reinforcement, prompt fading, and systematic arrangement, used in about 90% of included studies. The majority of studies implemented intervention strategies based on more behaviorally oriented approaches including interventionist initiation, massed teaching opportunities, contrived (rather than naturalistic) activity context, and a one-on-one instructional format. Average effects indicated that AAC interventions are generally effective. Few systematic differences were found between settings, instructional features, or behavioral versus naturalistic intervention strategies. However, there were significantly stronger effects and less heterogeneity for outcomes related to AAC use versus other social-communication outcomes (e.g., non-targeted outcomes). For participants with ASD, there were effects for interventions conducted in more controlled environments, although this difference was not significant when controlling for participant characteristics and instructional context variables. There were larger effects for participants with ASD in studies that used physical prompts versus those that did not and for those that did not use modeling versus those that did.
- Analyses related to targeted outcomes for single-case studies indicated that researchers implemented aided AAC with more participants (53%) than unaided AAC (20%; e.g., manual sign language, gestures). Due to the heterogeneity of the effects, few statistically discernible differences were found. Controlling for differences based on measures of participant characteristics and intervention characteristics did not explain a substantial amount of the variation.
People and institutions involved
IES program contact(s)
Project contributors
Products and publications
Study registration:
Publications:
ERIC Citations: Find available citations in ERIC for this award here.
Available data:
The researchers have created a project replication archive using the Open Science Framework (OSF). The archive includes raw data extracted from primary source articles, study characteristics coded as part of the project, and computer code for reproducing all of the analyses reported in manuscripts produced as part of the project. The archive can be accessed at https://osf.io/b5ydr/.
Related projects
Questions about this project?
To answer additional questions about this project or provide feedback, please contact the program officer.
