BackHome
Refresh: Off
Paper Summary
Share...

Direct link:

Finding the Sweet Spot in Interdisciplinary Data Science Education: The Data Puzzles Model

Wed, April 8, 3:45 to 5:15pm PDT (3:45 to 5:15pm PDT), Los Angeles Convention Center, Floor: Level Two, Room 515B

Abstract

A re-envisioning of data science education requires investigation of how the disciplinary communities of data science, education, and science disciplines can contribute to shared aims (e.g., Mayernick, 2023). How can these communities pursue shared aims? The Data Puzzles Research Partnership investigates this convergence as we leverage an ongoing middle school teacher PLC (n = 19) to promote student epistemic agency in science classrooms. The theoretical framework of the project (Figure 1) finds shared priorities among disciplinary communities: in 1) climate science education (USGCRP, 2024), 2) data sensemaking (Hunter-Thomson, 2019; Lee & Wilkerson, 2018), and 3) the Data Puzzles Instructional Framework (Griffith & Braaten, 2023). Given that these disciplinary communities have different priorities, the central pursuit of the project is the overlapping epistemic practice (e.g., Rouse, 1996) of working with the material features of data: Students explore, analyze, interpret, and reason with data to investigate climate phenomena (Figure 1). Our design principles are: 1) The use of high quality, standards-aligned instructional materials is the locus for teacher learning (Harris et al., 2015; Hestness et al., 2014) using data from professional interdisciplinary earth scientists 2) Teachers and students learn by working with data in supported, consequential environments, and 3) Students should work with datasets that are connected to real-world phenomena, and 4) Data sensemaking opportunities should be flexible and open-ended such that students make decisions about how to visualize and derive meaning. We ask, Can teachers’ confidence in each domain be increased in one PL workshop?

Figure 1. Theoretical Framework

Following our project’s inaugural 3-day professional learning workshop, we investigated teachers’ confidence in the three domains that intersect in this theoretical space. The initial data consists of teachers’ responses to a 5-point retrospective Likert item self-assessment of their confidence related to data sensemaking, climate science education, and the Data Puzzles Instructional Framework before and after the professional learning. Wilcoxon’s signed-rank test was used because the data were ordinal (Likert). For data sensemaking practices, confidence significantly increased from pre-test (M = 3.37, SD = 0.90) to post-test (M = 4.11, SD = 0.66) on a 5-point scale, V = 0, p < .001. Similarly, confidence in ESS-related instruction significantly increased from pre-test (M = 3.42, SD = 0.69) to post-test (M = 4.16, SD = 0.69) on a 5-point scale, V = 7, p < .001. Confidence with the Data Puzzles Instructional Model also showed a significant –and the largest–increase from pre-test (M = 1.89, SD = 0.94) to post-test (M = 4.11, SD = 0.46), V =0, p < .001. These statistically significant improvements in teacher confidence across all three interdisciplinary areas demonstrate that interdisciplinary, epistemic-practice-oriented teacher learning supported by high-quality instructional materials are a promising method for an integrated, meaningful future in data science education. This study offers evidence that the future of data science education can be inclusive of multiple knowledge communities if the shared pursuits center the creation of knowledge among learners.

Authors