Improving the Targeting of Treatment: Evidence From College Remediation

This post is from Judith Scott-Clayton (Teachers College, Columbia University)

While there is widespread concern nationwide about low rates of college readiness among our high school graduates—approximately 50 percent of all entering college students take at least one remedial class—little attention is paid to how “college readiness” is actually determined.  Remediating students is expensive: colleges spend $7 billion annually on developmental education, and this estimate does not include opportunity costs for students. Yet at community colleges, where almost half of all students begin college, readiness is almost always determined by scores on relatively short standardized math and English placement tests. Often, these scores alone determine whether students can enroll in college-level courses, or must first go through remediation.

A few years ago, the Community College Research Center conducted research finding that a number of students with scores below the cut-off on placement tests had skipped remediation and had nonetheless performed well in college-level courses. These findings pushed us to take a closer look at the assessments themselves.

To do this, we analyzed data from tens of thousands of community college entrants. Incorporating rich information on students’ high school performance, placement test scores, and demographics, we developed statistical models to predict how students would have fared had they been placed directly into college-level courses. Our analysis indicated that a quarter to a third of students assigned to remedial classes based on standardized test scores could have passed college-level classes with a grade of B or better.

These findings are striking – but perhaps not surprising. A number of rigorous studies have found that assignment to remediation has little impact, or even a negative impact, on students who just barely failed their placement exam. If many students who are assigned to remediation do not actually need it, this could explain why prior research has failed to find more positive effects.

To explore alternative methods of remedial screening, we looked at whether high school performance might be a more accurate placement device—and how using different screening devices would affect the gender and racial composition of remedial classes.

We found that high school transcript information was surprisingly powerful: using high school achievement resulted in far fewer misplacements (both into college-level courses and into remediation), and higher success rates in college level courses. And the improvements were substantial: the analysis estimated that incorporating high school transcript information could reduce incorrect placements by 30 percent and result in a 10 percentage point increase in the likelihood that students placed into college-level in the relevant subject would complete the course with a grade of C or higher, by the end of their first semester.

The analyses indicated that these lower severe placement errors and higher college–level success rates held true for all racial and gender subgroups. The choice of using high school transcript information versus placement test scores alone also can affect the composition of students in remedial classes. For example, in the setting we examined, incorporating high school grades would increase the proportion of women in college-level math classes, but could increase the proportion of Black students assigned to English remediation.

Around the time we released these findings in 2012, Long Beach City College (LBCC) implemented a pilot program, the Promise Pathways, based on similar findings from internal research. LBCC, which previously had primarily used test scores for remedial placement decisions, placed 1000 students from the local school district students in college-level courses based on high school grades and CST scores and encouraged students to take those courses immediately. The impact was dramatic and immediate: the percentage of students placed in a college English course quadrupled, as did the percentage of students who completed the course with a C or better in their first year.  Similarly, in college math, placement rates and the percentage of students completing the course in their first year tripled.

The California Community College Chancellor’s Office has taken these findings seriously. It has requisitioning a multi-year study of student assessment, placement and success at 11 California community colleges from the RP Group and Cal-Pass Plus, and is supporting the development of a multiple measures warehouse to make high school transcript and test score data available to colleges seeking alternate methods of placement. The signs are promising that a more nuanced system of placement—based on multiple measures of academic achievement—is on its way in California. The evidence strongly suggests that this change will result in a more equitable, efficient, and effective system of remedial education.

Twitter

  •  
  • 1 of 21

PACE thanks these funders and sponsors for their financial support