Date of Award

12-25-2024

Degree Name

Doctor of Philosophy

Department

Science Education

First Advisor

William Cobern, Ph.D.

Second Advisor

Betty Adams, Ph.D.

Third Advisor

Ya Zhang, Ph.D.

Keywords

Biology education, concept mapping, meta-analysis, research synthesis, science education, systematic review

Abstract

Contemporary education research seeks causal inferences in order to guide policy and practice. Often well-controlled quasi-experimental and experimental designs are used as they are very useful for supporting and contextualizing causal inferences. Nevertheless, individual studies, even high-quality, resource-intensive randomized controlled trials, are not able to establish generalized claims. Because primary studies vary over dozens of variables (e.g., grade, subject, duration) simply conducting additional primary studies cannot change this. Furthermore, the informal assessments of primary studies common in literature reviews are unable to resolve certain challenges (e.g., publication bias) that are described in the emerging field of meta-research. These challenges require systematic methods, as well as expert knowledge, to assess generalized claims and accumulate knowledge. In this three-paper dissertation, systematic review and meta-analyses methods are applied to the case of concept mapping instructional interventions in science education.

Concept mapping (CM) is a pedagogical tool which was first innovated in the 1970s and has been widely regarded as an effective learning activity. A concept map consists of conceptual nodes with connecting verbal links; each node-link-node connection forms a proposition. CM has been applied in a variety of ways and for a variety of instructional purposes such as collaborative learning, group discussion, directed reading, and formative assessment. With the move of classes to online settings during the COVID-19 pandemic, CM has received even more attention in science education as a tool for active learning in virtual settings (Gerber Hornink and Costa 2021; Choe 2020). Several meta-analyses have already been performed assessing the efficacy of CM interventions and have found significant positive effects. However, these meta-analyses have significant methodological limitations and focus on mean effects rather than on heterogeneity. Also, CM has had less impact on the educational profession than some of its advocates would have hoped (e.g., Kinchin, 2001). For these reasons, CM was deemed a good topic to investigate, serving as a case study of instructional interventions in science education.

The first paper discusses the challenges to knowledge accumulation in education research and reviews how research synthesis methods may be applied to address these challenges. This informs and motivates the second two papers which each serve as case studies of meta-regression used to explore heterogeneity in science education research. The second paper is a systematic review and meta-analysis of CM interventions in biology education which applies recent innovations in meta-regression modeling and informs the last paper. The last paper is another systematic review and meta-analysis of CM interventions which expands the scope to all science education research, implements an improved systematic review process, and applies machine learning to generate the best possible meta-regression model for explaining heterogeneity. From the results of these two meta-analyses, it is argued that there is little evidence supporting the use of CM interventions in science education and that primary studies should change certain methodological practices. Science educators and policy makers should read the science education research critically and should not accept generalized claims without sufficient evidence.

Access Setting

Dissertation-Open Access

Share

COinS