Project Activities
The researchers conducted a multi-site, multi-cohort randomized controlled trial to compare the Supporting Knowledge of Language and Literacy (SKILL) intervention program to a business-as-usual counterfactual condition. Teachers provided SKILL to small groups of 2 to 4 students in 36 lessons that were 30-minute long across a 3-month period. Examiners, who were kept unaware of students’ assignment to treatment or control, administered measures of reading comprehension, oral narration, and written narration before randomization, immediately after treatment, and 5 months after treatment ended. The researchers then performed multi-level modeling with students nested within teachers and teachers nested within schools to compare the outcomes of the treatment and control groups and structural equation modeling to assess moderation and mediation.
Structured Abstract
Setting
This study was conducted in urban and rural school districts in Utah and Texas.
Sample
Participants included 357 students who were at risk for language and literacy difficulties in grades 1 through 4 in 13 schools across 7 school districts. A large percentage of the 357 participating students were Latino (57 percent) with the remaining students identified as Caucasian (32 percent), African American (5 percent), and Asian (1 percent). A substantial percentage (42 percent) of parents also self-reported that the students were English learners with a home language other than English. Slightly more than half (54 percent) of the sample was male. Thirty-seven percent of participants were receiving special education services for a previously identified disability. The two most common disability classifications of participants were speech language impairment (19 percent) and learning disabilities (16 percent). Overall attrition ranged from 12 percent to 15 percent at posttest and from 22 percent to 23 percent at follow-up across outcomes. Rates of attrition were comparable across treatment conditions. At posttest, the difference in attrition rates for SKILL and BAU ranged from 0.1 percent to 1.7 percent across outcomes. At follow-up, differential attrition ranged from 4 percent to 5.7 percent.
Intervention
The manualized SKILL program consists of three phases: (1) teaching story structure and causal language, (2) teaching strategies for creating a situation model, and (3) teaching strategies for integration into long-term memory. SKILL leverages two language intervention techniques, parallel story production and vertical structuring, extensively during all three phases of instruction. In parallel story construction, instructors first model a story, then help students co-construct a similar story using scaffolding techniques. In vertical structuring, instructors use scaffolding techniques to help students connect short, simple sentences to form complex sentences. The primary purpose of these activities is to teach children to use complex sentences (microstructure) to communicate causal connections between story grammar elements (macrostructure). Each phase contains scaffolded lessons, curricular objectives, materials, and instructions. Lessons progress in each phase resulting in the use of literature-based activities that are closely aligned to the kinds of texts children read and write in general education classrooms. Each phase ends with a series of literature-based activities designed to help students transfer new skills into authentic contexts they will encounter in school.
Research design and methods
The researchers conducted a randomized controlled trial in which at-risk students within classrooms were randomly assigned to receive SKILL or business as usual. In a population sample, the researchers screened 3,380 students using the Gates-MacGinitie Reading Test and the Test of Narrative Language–2 to assess at-risk status. Those scoring at or below the 33rd percentile on both tests were invited to participate in the study. The researchers assessed outcomes pre-intervention, post-intervention, and 5 months post-intervention.
Control condition
In the control condition, students received standard classroom practices in place at the school. The researchers conducted 121 classroom observations for 30-minutes each to examine the amount and types of narrative instruction (i.e., story comprehension, oral storytelling, and story writing instruction) that occurred in classrooms containing BAU participants. The amount and type of story instruction provided to students varied across classrooms. The researchers found that 44 percent of observed minutes were devoted to story comprehension and 10 percent of the minutes addressed story writing. Teachers spent no time working with students on oral storytelling. Findings suggest that story production may not be an instructional focus in many primary-grade classrooms.
Key measures
Narrative outcome measures for this study included the Text of Narrative Language–2 and Narrative language samples, which were oral and written self-generated stories produced by the students. Literacy outcome measures included the Gates MacGintie Reading Test, 4th Edition, the Test of Word Reading Efficiency–2, and the Test of Early Written Language–3. Researchers administered the Auditory Working Memory subtest of the Woodcock Johnson III: Test of Cognitive Abilities as a potential mediator measure.
Data analytic strategy
The researchers used multilevel modeling and structural equation modeling to test the efficacy of SKILL on student outcomes, with students at level 1 and classrooms at level 2. The researchers assumed an intent-to-treat model.
People and institutions involved
IES program contact(s)
Products and publications
ERIC Citations: Find available citations in ERIC for this award here.
Project website: A basic summary of the project and information about how to access the project data are available at https://www.meadowscenter.org/projects/detail/supporting-knowledge-in-language-and-literacy-skill
Select Publications:
Capin, P., Gillam, S.L., Fall., A.M., Roberts, G., Dille, J., & Gillam, R.B. (2022). Understanding the nature and severity of reading difficulties among students with language and reading comprehension difficulties. Annals of Dyslexia. 72, 249-275.
Capin, P., Vaughn S., Gillam, S.L., Fall, A-M., Roberts, G., Holbrook, S., Wada, R., Dille, J., & Gillam, R.B. (2023). Evaluating the efficacy of a narrative language intervention for bilingual students. American Journal of Speech-Language Pathology, 1-22.
Hall, C., Capin, P., Vaughn, S., Gillam, S. L., Wada, R., Fall, A. M., Roberts, G., Dille, J. T., and Gillam, R. B. (2021). Narrative instruction in elementary classrooms: An observation study. Elementary School Journal, 121(3), 454-483.
Gillam, S. L., Vaughn, S., Roberts, G., Capin, P., Fall, A.-M., Israelsen-Augenstein, M., Holbrook, S., Wada, R., Hancock, A., Fox, C., Dille, J., Magimairaj, B. M., & Gillam, R. B. (2023). Improving oral and written narration and reading comprehension of children at-risk for language and literacy difficulties: Results of a randomized clinical trial. Journal of Educational Psychology, 115(1), 99-117.
Magimairaj, B. M., Capin, P., Gillam, S. L., Vaughn, S., Roberts, G., Fall, A., Gillam, R. B. (2022). Online Administration of the Test of Narrative Language-Second Edition: Psychometrics and Considerations for Remote Assessment. Language, Speech, and Hearing Services in Schools, v53 n2 p404-416.
Related projects
Supplemental information
Co-Principal Investigators: Gillam, Ronald; Gillam, Sandra Laing; Vaughn, Sharon
- Students who received the SKILL treatment significantly outperformed students in the business-as-usual (BAU) condition on measures of oral narrative comprehension and production immediately after treatment and at follow-up 5 months later (Gillam et al, 2023).
- Improvements in oral narration generalized to a measure of written narration at posttest and the treatment advantage was maintained at follow-up (Gillam et al, 2023).
- Grade level did not moderate effects for oral narration, but it did for reading comprehension, with a higher impact for students in grades three and four (Gillam et al, 2023).
- Moderation results showed that the effects of SKILL did not vary for monolinguals or bilinguals based on their pre-intervention language performance (Capin et al., 2023).
- Structural equation modeling showed that children’s narrative ability at posttest mediated the relationship between treatment and reading comprehension suggesting that narrative ability is the mechanism by which treatment generalized to reading comprehension (Gillam et al., 2023).
- Many students identified with oral language and reading comprehension difficulties, even those in upper elementary grades, present word-reading difficulties that may need to be addressed in reading interventions (Capin et al., 2022)
Questions about this project?
To answer additional questions about this project or provide feedback, please contact the program officer.
