Manifesto: Three decisions, three kinds of data

This week, I’m working through Rick Stiggins’ Assessment Manifesto. Yesterday’s post noted that assessment for learning, done well, leads to impressively higher achievement. Today, a key point on assessments and decisions made from their results.

Assessment information should support:

• Classroom decisions on next learning activities for each student.
• Program/school level decisions on whether to keep or change strategies.
• Outside accountability decisions on whether a school is delivering acceptable work.

No one assessment can do all that. Those different decisions need different information, on different timetables:

• Classroom work needs evidence about individual students continuously.
• Program planning need evidence about which standards are being mastered, looking at groups of students, on a “periodic but frequent” basis.
• Accountability checks whether enough students are meeting the standards, annually.

Two common Kentucky concerns are that CATS does not provide detailed diagnostics on individual students and that CATS results do not reach schools in time to change curriculum and other strategies before the school year begins. According to the Manifesto, we should not expect the annual test to fill those needs. Instead, diagnostic data should be gathered in the classroom, and planning data taken from program-level testing that happens multiple times a year. Stiggins would say we're right about needing more information, but wrong to think we can fill those need simply by changing the annual state-wide test.

Manifesto Point 2: Don’t expect a single assessment to to support three different levels of decisions that need to be made on three very different timetables. It takes a balanced system.