To explore the effects of the Calculus reform on retention we focused on whether or not students are retained at the university immediately subsequent to the year in which they encounter Calculus I. We divided 3002 student records into two groups: those who encountered the new version of Calculus and those who had the traditional experience. We then compared retention rates for the two groups. We found that the new Calculus course improved retention (relative to the old) by 3.4 percentage points; a modest, but statistically significant (*p* = 0.020) result. University retention rates for women, under-represented minorities (URM), and Pell-eligible students were also computed. All three subgroups showed gains, with URM leading with 6.3 percentage points of improved retention (p = 0.107)

We then considered retention within STEM as a measure of how the Calculus reform influenced students. For the same groups of students, we computed the rate at which STEM majors were retained in STEM. Once again we found a modest overall gain of 3.3 percentage points (*p* = .078). We found strong effects on women and underrepresented minorities (URM). The new Calculus course improved retention for both of these groups **by more than 9 percentage points**, a large effect. At this university, under the old Calculus, women used to lag men in STEM retention by about 8 percentage points. After the Calculus reform this gap nearly vanished, shrinking to 0.5 percentage points. Under the old Calculus, STEM retention of URM students used to lag that of non-URM. After the Calculus reform the gap flipped, so that underrepresented minority students are now retained in STEM at *higher* rates than non-URM.

As a final result we examined student success in courses that typically follow Calculus I. Here the metric is pass rate, and we compared pass rates between the students who took the new Calculus against those who took the old. For additional comparison we also included students who transferred into post-Calculus course work. Once again the reformed Calculus course led to better results.

]]>Suppose you want your students to know that a function has an inverse if and only if it is a bijection (both one-toone and onto). You could state the theorem, perhaps prove it, and work some related problems. Or, you could ask them to explore a set of carefully chosen examples, creating an opportunity for students to observe the relationship. We observed a college discrete mathematics class in which the second approach was taken. Students examined a set of nine functions to determine which functions had inverses; the functions were chosen to challenge assumptions about functions and their properties. Students determined whether the functions were injective (one-to-one), surjective (onto), or both (bijective). Data from students provided insight that only the functions with inverses were bijective. This type of mathematical activity served to review function concepts and provide opportunities for making significant mathematical observations, which can then be explored further or proven.

]]>Project leaders are seeking to expand these gains to other areas of the curriculum and to broaden the community of instructors who are fully accepting of the reforms. Common concerns expressed by faculty resistant to the overhaul include suspicion that pass rate gains might reflect grade inflation or weakened standards, and that altering the traditional content of Calculus I might leave students unprepared for Calculus II. External stakeholders also have a vested interest in ensuring students receive a solid preparation in Calculus I. In this paper we develop a response to ensure solid evidence of Calculus II readiness that we hope will be useful to change agents and campus leaders in many other settings.

We address concerns about Calculus II readiness by conducting a natural experiment, tracking two cohorts of students through Calculus I and into Calculus II. The “treatment” cohort consists of students who reach Calculus II after passing the reformed Calculus I. The “control” cohort consists of students who reach Calculus II after passing non-reformed Calculus I at Boise State University. The experiment has no designed randomizing, but enrollment data shows that both cohorts spread out across all sections of Calculus II with apparent randomness. Our research question is: “Does the treatment cohort perform any worse than the control cohort in Calculus II?” Data on pass rates and grades in Calculus II will show that the answer is “No.”

]]>