Association for Public Policy Analysis & Management

Researchers from CCRC and MDRC are currently investigating the effects of multiple measures assessment (MMA) and placement strategies on students’ academic performance, persistence, and progress toward college completion using randomized control studies. Research suggests that the standardized tests that most institutions use for placement do not always accurately determine which students will benefit from developmental education. A more accurate placement mechanism would ideally allow colleges to decrease the number of students in developmental courses without harming success rates in college-level classes. This would also save students money and time, by lowering the total number of classes they need to take in order to graduate and potentially increase the chances of student persistence towards graduation.

The learning objectives include:

1. The reasons why colleges may wish to consider using multiple measures for assessment and placement, and different options for implementing MMA systems.
2. Key implementation lessons from two recent MMA publications, with a focus on helping colleges with an interest in MMA be successful.
3. Sharing of early impact findings from the use two alternative assessment and placement systems.

During the session:

* First, we will explain why there is a need to change assessment and placement practices, and share information on our approaches to developing and testing alternative strategies based on data analytics.

* Second, we will share findings from two 2018 publications from two studies: the Ascendium Multiple Measures Assessment Project and the CAPR Assessment Project.

https://ccrc.tc.columbia.edu/publications/multiple-measures-placement-using-data-analytics.html

https://www.mdrc.org/publication/toward-better-college-course-placement

* We will go over key implementation findings from both reports.

* We will share findings on early impacts (such as placement into college level and passing the first college-level course in the first semester) from both studies - including new findings from 2019.

Examples of tips and lessons that we will share in this session include:

1. In analyses performed so far to set up algorithms, college placement tests are much less predictive of success than high school variables, especially high school GPA.
2. MMA can be based on precision-weighted models that produce a probability of success, or simpler decision rules that define cutoffs for combinations of measures.
3. The process of setting up an alternative placement system requires considerable planning across admissions, testing, advising, course registration, IT, and academic departments. We will share guidance on the following questions:
   1. How Do We Decide Whether Our Institution Is Ready for MMA?
   2. How Should We Organize the Work?
   3. How Should We Choose the Placement Criteria?
   4. How Do We Manage the Logistical Aspects of the MMA System?
   5. How Can We Make Sure That the System Works as Intended?
   6. What Are the Costs of MMA Design and Planning?