Analytic agency framework: three stages of data analysis (exploration, interpretation, confirmation), each classified as primarily human-led or computer-led¶

Theory · All theories
Evidence · no claims cited

Description¶

The framework treats data analysis as proceeding through three stages: exploration (how data exist and are transformed or selected), interpretation (how patterns are made sense of, labelled, or theorized), and confirmation (how findings are tested, validated, or generalized). For each stage it asks which entity—human researcher or computational tool—most strongly shapes which questions are asked, which patterns are noticed, and which results are treated as meaningful. The authors situate common methods (statistical hypothesis testing, qualitative analysis, supervised and unsupervised machine learning, computational grounded theory, and their own case study) in a table classifying each stage's leadership, arguing this turns positionality into specific, nameable stage-leadership choices.

Design Implications¶

Context¶

Requirements¶

• Analysts must explicitly name which entity leads each of the three stages rather than leaving leadership implied

Constraints¶

• The authors state the method classifications are not prescriptive but placed by the default locus of decision-making most practitioners would recognize; the case study is one example, not a method recommendation

Target Learners¶

• Education researchers analyzing multimodal classroom data

Target Learning Objectives¶

• Transparent and defensible methodological design in educational data mining and learning analytics

Claims¶

Related Theories¶

• Multimodal analysis: studying learning with multiple integrated data streams

Examples¶

• Multimodal pipeline for detecting and analyzing teacher–student group interactions in classroom video
• Design teacher-facing analytics that support interpretation of student persistence within classroom contexts

Key Sources¶

• Hur, P., Palaguachi, C., Machaka, N., Krist, C., Dyer, E. B., D'Angelo, C., & Bosch, N. (2026). A Framework for Considering Exploration, Interpretation, and Confirmation During Data Analysis: Computationally Assisted Analysis of Teacher–Group Interactions. Journal of Educational Data Mining, 18(1). https://jedm.educationaldatamining.org/index.php/JEDM/article/view/1030