Many advocates have bemoaned the focus on multiple-choice testing that has grown since the advent of No Child Left Behind. They often argue for more meaningful assessments, ones that include more essay and performance-based items. But because those assessments are more expensive to score, the cost of such systems is typically seen as a stumbling block.
Well, maybe not. That’s the conclusion of a study due to be published next week as part of a Stanford University project. It finds that by implementing key savings strategies, states can actually save money by using assessment systems that tilt more toward essays and performance-based tasks.
The paper’s three authors estimated the cost of a typical state assessment system, defining that as once-per-year summative tests that are nearly all multiple-choice questions, with one or two extended constructed-response questions in reading, writing and mathematics. Such a system, they concluded, costs a moderate-sized state $52.3 million, or $19.93 per student, over a four-year period that includes the original year of development and field-testing.
A testing system they refer to as a “high-quality comprehensive assessment system,” which would include half as many multiple-choice items and more essay and performance-type questions, would cost $146.1 million, or $55.67 per student, over that same four-year period.
But certain cost savings can reduce that substantially, the authors found. Chief among them were joining in a consortium of 30 or more states to share costs, and using teachers instead of vendors to score test items. If states were to employ all the cost-reduction strategies outlined in the report, the cost of the “high-quality comprehensive assessment system” would drop to $10 per student, half the cost of the current typical state assessment system, the study found.
The authors, Barry Topol, John Olson, and Ed Roeber, are affiliated with the Assessment Solutions Group, an organization that provides consulting services on testing issues. They did the study as part of a Stanford University project that is exploring various aspects of performance assessment. All eight papers released as part of that project, including this one, are due for release on April 13, and can be viewed then at http://edpolicy.stanford.edu.