Using Multiple Measures to Understand Science Inquiry Behaviors in Open-Ended Learning Environments in Two Countries

• In Event: Transfer for Learning in the Classroom: Developing Contexts to Examine STEM Learning Skills and Strategies

Abstract

In the past three decades, different programs and resources have been implemented to promote scientific inquiry processes in the classroom (Minner & Levy, 2002; NGSS 2013). Yet, the study of scientific inquiry has been mainly conducted in environments with prescribed learning sequences and pre-defined skills, and is often assessed with measures like test scores and survey data (Gobert, Sao Pedro, Raziuddin, & Baker, 2013). The increase of digital environments provides new inquiry environments for students to learn to navigate and new opportunities for the assessment of inquiry skills (Henrie, Halverson, & Graham, 2015).

We contribute to this growing body of work by studying students’ behaviors in an open-ended learning environment which simulates a tug-of-war competition. This builds on previous work studying the same environment, in which a probabilistic model representing students’ knowledge and strategy use was developed and ultimately used to cluster students into groups based on in-game behavior (Authors, 2017). We conducted a first study in Colombia in the summer 2018 with five public schools, five teachers, and 223 eighth graders. Three measures were used for this study: i) clusters based on game logfiles, ii) student learning strategies ranked by the teacher, and iii) science teaching practices in the classroom.

A high number of students were in cluster with low experimentation (N=130), while only 7 students were in a cluster with high levels of experimentation. We also looked if students explore systematically. In previous studies, they found US schools have more students who explore systematically (around 10% of the student sample) and that around 30% fall in the low exploration cluster. In the case of Colombia, 3% of the students are systematic explorers and 58% of the students show a low number of exploratory behaviors.

When students did explore, their experimental set-ups were rated as low, medium, or strong based on the amount of information the experiment could provide to the students. In the US dataset, the number of set-ups rated as strong was significantly correlated with students' science grades. Whereas in this first study conducted in Colombia, there was not a significant association. One plausible explanation is that if teachers do not promote inquiry in the classroom there is a lower probability to assess this skill. Therefore science grades might not reflect good inquiry.

We conducted a second study in summer 2019 with a sample twice the size of the first study. We added more items to better understand the classroom practices of science teachers and a survey to measure students’ perceptions of science class and themselves as learners.

Results from the first study show the importance of understanding better how students develop inquiry behaviors in open-ended learning environments but also differences between countries. We expect that results from the second study allow us to identify how multiple measures provide different types of data to better understand and support students’ inquiry behaviors in open/ended learning environments in different countries.

Authors

• Ana Saavedra, Stanford University

• Tanja Kaeser, Institute for Dynamic Systems and Control

• Daniel L. Schwartz, Stanford University