Establishing a new standard of care for calculus using trials with randomized student allocation

Calculus, the study of change in processes and systems, serves as the foundation of many STEM disciplines. Traditional, lecture-based calculus instruction presents a persistent barrier for students seeking STEM degrees, limits access to STEM professions, and blocks their potential to address society’s challenges. A large-scale pragmatic trial with randomized student allocation was conducted to compare two calculus instruction styles: active student engagement (treatment condition) versus traditional, lecture-based instruction (control condition). A sample of 811 U.S. university students was studied across 32 sections taught by 19 instructors over three semesters at a large U.S. Hispanic-Serving Institution. Large effect sizes were consistently measured for student learning outcomes in the treatment condition, which demonstrated a new standard for calculus instruction and increased opportunities for completion of STEM degrees.

Methods
Data contained in this archive were collected from an NSF funded study carried out as a large-scale pragmatic trial that randomly allocated students to the treatment and control conditions to establish a new standard of care for calculus instruction. The treatment group used the Modeling Practices of Calculus (MPC) pedagogy and curriculum with the control group employed pre-existing instructional practices (primarily traditional lecture). Identical end-of-semester learning measures were used to rigorously assess student learning outcomes in the two conditions in the trial. Sets of open answer, learning measure questions were collectively developed and embedded as part of the identical end-of-semester learning measures given each semester to the treatment and control sections in the MPC Calculus 1 trials in order to assess end-of-course student learning in each of the groups. 

A total of 1,019 students were randomly assigned to either an active learning based curriculum in calculus or a traditional lecture-based approach. 516 students were assigned to the treatment group and 417 remained that in turn provided 373 outcome data points while 503 students were assigned to the control group and 394 remained in the treatment section at the drop/add deadline in the control condition providing 310 data points for analysis.

Each data point contains an instance of student allocation in the study including the demographic data collected, indicators of study arm assignment, and outcome data.