Mapping design stages and methodologies for developing STEM concept inventories: a scoping review

Background: Concept inventories (CIs) have become widely used tools for assessing students’ learning and assisting with educational decisions. Over the past three decades, CI developers have utilized various design approaches and methodologies. As a result, it can be challenging for those developing new CIs to identify the most effective and appropriate methods and approaches. This scoping review aimed to identify and map key design stages, summarize methodologies, identify design gaps and provide guidance for future efforts in the development and validation of CI tools. Methods: A preliminary literature review combined theoretical thematic analysis (deductive, researcher-driven) focusing on specific data aspects, and inductive thematic analysis (data-driven), using emerging themes independent of specific research questions or theoretical interests. Expert discussions complemented the analysis process. Results: The scoping review analyzed 106 CI articles and identified five key development stages: define the construct, determine and validate content domain; identify misconceptions; item formation and response processes design; test item selection and validation; and test application and refinement. A descriptive design model was developed using a mixed-method approach, incorporating expert input, literature review, student-oriented analysis, and statistical tests. Various psychometric assessments were employed to validate the test and its items. Substantial gaps were noted in defining and determining the validity and reliability of CI tools, and in the evidence required to establish these attributes. Conclusion: The growing interest in utilizing CIs for educational purposes has highlighted the importance of identifying and refining the most effective design stages and methodologies. CI developers need comprehensive guidance to establish and evaluate the validity and reliability of their instruments. Future research should focus on establishing a unified typology of CI instrument validity and reliability requirements, as well as the types of evidence needed to meet these standards. This effort could optimize the effectiveness of CI tools, foster a cohesive evaluation approach, and bridge existing gaps.