Across California, community colleges are placing far too many students into remediation who don't need to be there.
See Fall 2015 Placement into college English
According to Judith Scott-Clayton of the Community College Research Center, the under-placement of students is often invisible to faculty. "When a student is placed into a college-level course and fails there, the fact that there has been a placement mistake is painfully obvious to all." But "when a student does well in a remedial course, it's unlikely to be perceived as a problem."
For every remedial course required, a student's chances of completing college English and math drop. Changing placement policies to allow more students to begin in college-level courses is a powerful strategy for increasing student completion and narrowing equity gaps.
- When Long Beach City College and Butte College dramatically increased student access to college English, completion of college English was 1.3 to 3 times higher for all students, and students of color saw the greatest gains.
- A randomized controlled study in the CUNY system found that students placed into elementary algebra were substantially more successful if allowed to enroll in college statistics than those who took elementary algebra. A year after the intervention, their completion of college math was over three times higher than students who took the traditional algebra course.
- When the Virginia community college system implemented a pathways approach to math placement, with different competencies required for different college-level courses (e.g., liberal arts math vs. STEM), completion of college math tripled.
One of the ways colleges are increasing student access to transfer-level courses is through considering high school transcript information in placement.
All California community colleges are legally required to use multiple measures to place students. In practice, however, most have relied almost exclusively upon standardized tests, which are incredibly weak predictors of students' performance in college. Indeed, one of the most commonly used tests -- Compass -- has been taken off the market, with the test-maker directly acknowledging its limitations.
Researchers working with the Multiple Measures Assessment Project (part of the statewide Common Assessment Initiative) are recommending a much more robust use of high school transcript information. They recommend a "disjunctive" approach, under which colleges provide multiple ways for students to qualify for a transfer-level course, such as through test scores OR key high school measures, whichever is higher.
The researchers studied a large data set from community colleges and high schools throughout California. They identified the factors that predict success in college and used these to develop placement rules colleges could adopt. Students who meet the criteria are predicted to pass transfer-level courses at a rate of at least 70%.
In English, students with an overall high school GPA of 2.6 (B-) or higher are recommended for placement into college English. In math, the placement rules are more complex because students take many different courses at the college level, depending upon their program of study. Overall, placement is determined by students' high school GPA, performance in high school math courses, and math scores on students' California Standards Test.
Issues to Consider
High School GPA: Overall high school GPA is the strongest predictor of students' performance in college. It reflects how students perform in multiple subjects, over multiple years, with multiple instructors and modes of assessment. In their study of one state's community college system, CCRC researchers recommended that students with a high school GPA of 2.3 or higher (C+) be allowed to bypass placement testing and enroll directly in college-level courses.
The Power of "OR": Setting up placement rules with "or" statements means that students have multiple ways to qualify for college level courses. Students who don't do well on standardized tests, for example, can demonstrate their readiness through high school GPA, or their grades in English and math courses (e.g., junior or senior English, Algebra 1 or 2). Placement rules that use "and," on the other hand, narrow the pool of qualifying students.
Using Self-Reported Transcript Information: Historically, community colleges have been constrained in considering students' high school records because colleges had to secure separate data-sharing agreements with all of their feeder high schools. But with studies showing that students' self-reported data is reliable, colleges are now asking students to report their high school GPA and course-taking history during the intake/assessment process. This makes it possible to consider high school measures in the placement of all students, not just those from certain high schools.
Disproportionate Impact: Under state rules for disproportionate impact, no racial/ethnic group should have less than 80% of the access to a transfer-level course that white students have. For example, if 60% of white students qualify for college English, access for other racial/ethnic groups should be no lower than 48% (.8 x .6). Examine your own data: English template, math template.
Differentiated Math Placement: Colleges are beginning to question the use of a single, algebra-based definition of "college readiness" in math and instead defining different criteria for different college-level courses (e.g., Statistics vs. Pre-Calculus). Given that fewer than 10% of Elementary and Intermediate Algebra topics are needed in the study of Statistics, there is growing scrutiny about using tests of these topics to deny students access to the course.
Outcomes for Students who Don't Meet Criteria: The multiple measures rules recommended by MMAP would place far more students into transfer-level English and math, but they are conservative, excluding students if their pass rate in these courses is predicted to be less than 70%. The problem is that excluded students have even worse outcomes when placed into a remedial course. For example, under MMAP rules, a student would be excluded from college Statistics if their pass rate in the course was predicted to be 60%. But requiring that student to begin in a math course one-level-below transfer means that at most colleges their likelihood of completing college Statistics would drop from 60% to less than 40%. Corequisite models -- in which students have access to transfer-level courses with additional support -- are a powerful solution to this problem.