Skip Navigation

Cognition and Student Learning


- OR -


- OR -


- OR -

FY Awards

- OR -

Developing a Teacher-Based Intervention Involving Memory-Relevant Language During Instruction

Year: 2012
Name of Institution:
University of North Carolina, Chapel Hill
Goal: Development and Innovation
Principal Investigator:
Ornstein, Peter A.
Award Amount: $1,199,987
Award Period: 3 years (7/1/2012-6/30/2015)
Award Number: R305A120402


Co-Principal Investigator: Jennifer L. Coffman

Co-Investigator: Patrick J. Curran

Purpose: With increases in age, children become more proficient in the use of strategies or plans for the storage and retrieval of information. Given the importance of remembering for success in school, it is essential that children develop strategies for remembering. Longitudinal findings by this research team reveal clear differences in the use of "memory talk" across teachers, as well as linkages between teachers' mnemonic style and children's memory and academic performance. Children taught by teachers whose talk during instruction in mathematics and language arts reflects a "high mnemonic" orientation show better skill in the use of strategies than do their peers who are taught by teachers who are "low mnemonic" in their orientation. Moreover, the researchers have found that teachers can be taught to teach in either a high or a low mnemonic style and that the language of instruction was associated with differences in children's strategy knowledge and use. The purpose of this project is to develop and pilot test an intervention intended to improve both teachers' mnemonic style and children's memory and academic performance.

Project Activities: The research team will conduct four experiments over 3 years that will first isolate the key components of mnemonic style, and then test the intervention as embedded into either science content, Things that Move, or language arts content, Reporting the News. Children will receive a series of 2-hour long lessons taught on consecutive weekday afternoons in an after-school program. After collecting baseline data on teachers' mnemonic style, the researchers will train teachers in the content of the intervention and in the use of alternative conversational styles modeled on the high and low mnemonic styles of instruction identified in the researchers' previous exploratory studies. The team will then vary children's exposure to high mnemonic language in a separate experiment, before piloting the fully developed intervention in first-grade classrooms.

Products: The products of this project will be a fully developed teacher-based intervention involving memory-relevant language during instruction, evidence of the promise of the intervention, and peer-reviewed journal publications.

Structured Abstract

Setting: Experiments will take place in after-school and classroom settings in elementary schools in central North Carolina.

Sample: Approximately 320 first-grade students will participate in this project. Teachers who are licensed to teach in the early grades of elementary school but who are not working in the schools at the time will teach the curricula.

Intervention: Researchers will develop a teacher-based intervention involving memory-relevant language embedded in either a science intervention, Things that Move, or a language arts intervention, Reporting the News. Professional development training will also be provided to teachers in the pilot study.

Research Design and Methods: The research team will carry out a series of four experiments—three carried out in afterschool programs and one in first-grade classrooms—that will support the iterative development of this intervention. In Experiments 1 and 2 during Year 1, the research team will "unpack" measures of mnemonic style and embed them in two domains of instruction, one focusing on science (Experiment 1 uses the unit Things that Move) and the other on language arts (Experiment 2 uses the unit Reading the News). In both experiments, contrasting groups of first-graders will be taught by teachers who use (a) all components of high mnemonic style, (b) a subset of the components that relate to the depth of processing of information, or (c) a subset that emphasizes metacognition. In addition, children in a control group will receive the instruction that mimics the low mnemonic style, and assessments of children's performance are made at three times during the unit. In Experiment 3 during Year 2, the teachers will use the language components (i.e., the full set of components, the subset dealing with cognitive structuring activities, or the subset focusing on metacognitive information) as well as the length of the instruction (two, three, or four weeks) associated with better outcomes in Experiments 1 and 2. The researchers will examine effects of exposing children to high mnemonic instruction in one versus two subjects taught in succession, using a crossover design with tests at the end of the first unit. Year 3 ends with a pilot study (Experiment 4) in which the research team will use what they have learned in Experiments 1 through 3 to craft an intervention to be delivered by first-grade teachers during regular instruction. In this pilot study, the researchers will contrast the performance of children who are and are not exposed to teachers who have been instructed in the use of memory-relevant language in the classroom.

Control Condition: In Experiments 1 and 2, children in a control group will receive the instruction that mimics the low mnemonic style identified in previous research. During Experiment 3, no control group is used; instead, the contrasts include reversing the order of the science and language arts units that use high mnemonic style (science first following by language arts, or the reverse) to see if exposure matters. In Experiment 4, the pilot study, there will be "natural" control classrooms in which teachers do not receive any instructions concerning the language to use during their teaching. Control teachers will receive instruction in high mnemonic style at the end of the study.

Key Measures: All procedures will be videotaped and/or audiotaped for subsequent analysis. Measures given at multiple intervals throughout the intervention include: (1) memory measures (Object Memory Task, Free Recall with Organizational Training Task, Content-Specific Sort-Recall Task); (2) indicators of academic skills (Woodcock-Johnson Tests of Achievement III), Basic Math Solving Strategy Assessment; and (3) baseline teacher measures collected via 1-hour observations of instruction in science or language arts to assess teachers' natural mnemonic style at baseline. Ancillary measures will include child- and home-level factors.

Data Analytic Strategy: In Experiments 1–3, the researchers will use a series of hierarchical linear models (HLMs—continuously distributed variables) and hierarchical nonlinear models (HNLMs model fitted to dichotomous and ordinally scaled variables) to analyze the three-level structure of time nested within children (i.e., Level 1—children's memory and academic skills over time predicting Level 2—the variability of children's developmental trajectories from home and child-level factors), and children nested within classrooms (Level 3). Preliminary methods such as principal factor analysis will be used as a data reduction strategy to form optimal measures of the researchers' constructs prior to the HLM analysis. Traditional analytics (analysis of variance, t-tests) will be used in the pilot study during Year 4. Because the HLM approach allows for partially-missing cases, data from all children will be used in the analysis, including the small number of children who may not complete all assessments.