Abstract
Gillam, S. L., Gillam, R. B., & Reece, K. (2012). Language outcomes of contextualized and decontextualized language intervention: Results of an early efficacy study. Language, Speech, and Hearing Services in Schools, 43, 276–291.
Approaches to language intervention may be categorized as being contextualized or decontextualized in nature. Contextualized language intervention (CLI) approaches train multiple linguistic targets or curriculum-related skills within activities that involve rich, meaningful, and coherent references to people, objects, and actions. In CLI, clinicians model and elicit target language structures during functional activities that promote a great deal of clinician–child interaction. CLI is based on Vygotskian social–interactionist principles (1978) that children learn best when they engage in activities with more skilled participants who provide them with models and support within authentic communicative interactions. CLI is consistent with current frameworks supporting the use of curriculum-based instructional approaches to language intervention (Justice, 2006; Whitmire, 2002).
Results of multiple studies, however, have shown that children with learning disabilities and/or specific language impairment (SLI) often require explicit skill instruction before they are able to generalize new knowledge in contextually relevant contexts (Fuchs & Fuchs, 2001). Thus, speech-language pathologists must strike a delicate balance between teaching specific linguistic structures and facilitating the development of functional language in naturalistic contexts.
In decontextualized language intervention (DLI), sessions typically consists of a series of short clinician-led activities that focus on a variety of specific skills with minimal continuity across the topics of the interactions. For example, children may define words in one activity, followed by an activity in which they use past-tense verbs to create sentences to describe pictures on cards, followed by an activity in which they explain how two words are alike or different. In each activity, the topics and interactive contexts are different. There is little evidence about the comparative outcomes of CLI versus DLI approaches to improving language skills in school-age children.
The authors of this study report on a small-scale study that compared the language outcomes of a CLI designed to teach language skills during authentic, topic-centered discussions of children’s literature with the outcomes of a DLI designed to teach language skills in drill-based contexts with minimal topic continuity across activities. The specific research question was as follows: Will children who receive CLI or DLI present greater improvement on sentence-level or narrative-level language measures than children in a no-treatment control (CON) group?
Study
Groups
A total of 16 first- and second-grade students with language impairment (LI) who were receiving treatment for LI were randomly assigned to the CLI and DLI experimental groups. In all, 8 children of the same age with LI who were in a no-treatment phase of a study were in the CON group. All children were given the same measures separated by a 6- to 8-week period of time.
Measures
All participants had a standard score of less than 85 on the Clinical Evaluation of Language Fundamentals–Fourth Edition (CELF-4; Semel, Wiig, & Secord, 2005).
The Recalling Sentences and the Formulating Sentences of the CELF-4 were used as outcome measures because they assess the type of memory and sentence production skills taught and practiced in the DLI.
The Test of Narrative Language (TNL; Gillam & Pearson, 2004) was used for four outcome discourse measures:
○ TNL narrative language index, ○ TNL narrative comprehension, ○ A measure of microstructures (use of conjunctions, mental/linguistic verbs, adverbs, elaborated noun phrases), ○ A measure of narrative macrostructure (children received points for included elements of a story: setting, character, initiating event, internal response, plan, action/attempt, consequence).
Procedures
Children in the intervention groups received 50-min sessions 3 times a week for 6 weeks (for a total of 15 hr of intervention). Children in the CLI group participated in a literature-based, narrative intervention. The CLI was structured around children’s literature and incorporated oral and written language. Children were provided with multiple opportunities to talk about and share knowledge of the story content, use vocabulary and grammatical structures from the model stories, answer comprehension questions, generate inferences, compare/contrast characters on physical attributes and personalities and actions and consequences, brainstorm solutions to problems, and retell the stories. Four different children’s trade books related to two topics, bedtime and parties, were used. Children were encouraged to link relevant world knowledge and experiences to the vocabulary and sentence structures that were taught in each session. During lessons, children were asked to define two words from the stories and to generate sentences containing the target words contained in the books. Children also practiced using conjunctions (e.g., and, or, but, so), modals, and question forms (e.g., auxiliary inversion) as they discussed and retold the stories.
Children who were randomly assigned to the DLI group answered questions and played games from the No-Glamour series published by LinguiSystems (LoGiudice, 2002, 2004; LoGiudice, McConnell, & Warner, 2004; McConnell & LoGiudice, 2003). These packaged game sets were chosen for the DLI program because they incorporated vocabulary, grammar, and narrative discourse targets that were similar to the targets of the CLI program, but unlike the CLI intervention, there was not a topic continuity within or across activities. Each question on each card had its own topic. The No-Glamour grammar cards (LoGiudice, 2004) target basic grammatical structures. The instructions with the card sets encourage clinicians to give children an explanation of each grammatical target, an example of how to use each target, and opportunities to identify and use the grammatical targets within a variety of question and answer activities.
The No-Glamour social language/behavior cards (LoGiudice et al., 2004) provide illustrations of social situations and are grouped into categories such as school, eating, feelings and self-control, getting along, conversations, being responsible, solving problems, and role-playing. Each card relates to a new topic. Clinicians were directed to select cards in one of the categories, read the situation on the card, and ask the children to paraphrase it before answering. Clinicians were encouraged to accept and facilitate discussion around answers that were reasonable and appropriate even if they did not match the answers presented on the cards. The No-Glamour category/definition cards (McConnell & LoGiudice, 2003) target categories such as functions, attributes, associations, comparisons, compound words, synonyms, antonyms, multiple-meaning words, and absurdities. The No-Glamour language cards (LoGiudice, 2002) are designed to improve a variety of skills, including asking and answering questions, exclusion, negatives, listening, semantics, grammar, making inferences, comparing and contrasting, paraphrasing, retelling, social communication, and problem solving. For all activities, clinicians were instructed to provide “therapeutic” feedback in the form of focused stimulation, explanations, growth-relevant recasts, and vertical structures, much like the feedback that occurred in the CLI condition.
Results/Discussion
○ The CLI group’s posttest scores on the Recalling Sentences and Formulated Sentences subtests were significantly larger than the CON group’s posttest scores. The effect sizes of the CLI versus CON differences were large for both Recalling Sentences and Formulated Sentences. ○ The DLI group’s posttest scores were significantly higher than the CON group’s scores for Formulated Sentences but not for Recalling Sentences. ○ Effect sizes for the CLI intervention were 81% larger on average than effect sizes for the DLI intervention.
○ Neither the CLI or DLI groups differed significantly from the CON group on the macrostructure score. ○ There were significant differences between the CLI and CON groups for the TNL Narrative Language Index, the TNL comprehension score, and the microstructure score. ○ The only significant difference between the narrative performance of the DLI and CON groups was on the microstructure score. ○ For the four narrative measures combined, the effect sizes for the CLI were nearly 3 times larger than the effect sizes for the DLI.
Neither the CLI or DLI explicitly taught the children the components of narrative macrostructure, so this may account for the lack of difference between the children in the experimental groups and the CON group. Both the CLI and DLI were effective for improving some sentence-level abilities, but only the CLI group performed significantly better than the CON group on a general measure of narrative language ability and on measures of narrative comprehension and narrative microstructure.
Both the CLI and DLI programs incorporated activities that involved listening to short stories, asking or answering questions, brainstorming solutions to problems, defining words, and generating sentences containing targeted vocabulary words. The primary difference between the interventions was that the CLI condition provided more “topic continuity” across activities that were presented in narrative-based intervention contexts. In the CLI program, children and clinicians engaged in discussions and therapeutic interactions that centered on themes that were introduced in children’s literature. These themes (topics) were included in all the language activities across multiple sessions. In the DLI program, the topic of the interaction changed when each new card was introduced. There were some topics that were continued across several turns, as when an activity included the use of social language cards, but the topics did not continue across activities and day. Although there were aspects of DLI that were not entirely decontextualized, the DLI condition differed substantially from the CLI condition in terms of the level of decontextualization that children experienced within and across sessions. The CLI resulted in the development of a wider range of language skills than the DLI; and in those areas where the DLI did result in language gains, the gains were not as great as those in the CLI. The results of this study advocate for use of contextualized approaches to language intervention rather than the training and practice of individual language skills and concepts.