A Meta-Analysis of Strategy Use and Performance in the Model of Domain Learning

• In Event: The Model of Domain Learning: Understanding the Development of Expertise

Abstract

Objectives

This meta-analysis examines the relation between strategic processing and performance in studies that have utilized the Model of Domain Learning (MDL). Herein we examine students’ strategic processing, as has been explored in MDL-driven studies and the relation between strategy use and associated MDL constructs (i.e., knowledge, interest). Specifically, this synthesis will examine: which strategies have been examined in the MDL and how these have been assessed; the relation between strategic processing and other forces in the model (i.e., knowledge and interest); how strategic processing manifests at different levels of expertise and its relation to performance; and, how contextual factors influence the relation between strategic processing and performance.

Theoretical Framework

The MDL hypothesizes that as individuals move from acclimation, through competency, and into proficiency, the use of surface-level strategies decreases, while the engagement of deep-level strategies increases. Changes in strategic processing along with co-occurring developments in domain knowledge and individual interest improve performance on increasingly complex tasks (i.e., ill- versus well-structured). Thus, as deeper-level strategic processing increase, performance in academic domains, particularly those that are ill-structured should improve.

Methods and Data Sources

A systematic search of the literature identified 17 studies that used the MDL to examine strategic processing. Studies were reviewed and coded for: (a) stage of the MDL examined; (b) age of the participants; (c) sample size; (d) levels of strategic processing examined; (e) measure of strategic processing used; (f) other MDL forces examined; (g) academic domain considered; (h) task type; (i) effect sizes of the relation between strategic processing and performance; and, (j) effect sizes between knowledge, interest, and strategic processing.
For the meta-analysis (8 of the 17 studies in the table), all effect size estimates were converted to Pearson correlation coefficients (r) and analyzed using Hunter and Schmidt’s (2004) random-effects model. Further, two categorical moderating variables (i.e., type of task and MDL stage) were entered into the model to determine the extent to which they change the nature of the relation between strategic processing and performance.

Results and Scholarly Significance

Findings from the review indicate that: retrospective self-report was the most common measurement of strategic processing (71%); undergraduate students in upper acclimation were the most common participants for which strategic processing was measured (71%); most tasks given were reading tasks about a domain, rather than domain performance itself; and, although only 8 of the 17 studies directly related, the population effect between strategic processing and performance bordered on a small to medium effect (r=.19; SD=.13).
Based on the paucity of studies that directly examine the relations between strategic processing and performance and the monolithic manner in which strategies are measured, we conclude that there is much to be done in the area of strategic processing, both within the MDL theoretical framework and beyond. We strongly encourage further use of the MDL to study strategic processing due to the ability of the model to explain the multidimensional and developmental nature of strategic processing in complex academic domains.

Authors

• Daniel Dinsmore, University of North Florida

• Courtney Hattan, University of Maryland - College Park

• Alexandra List, The Pennsylvania State University