Like a Lead Balloon
Assessment informs instruction.
Assessment has other purposes, too. Assessments measure rote learning and mastery, they compare students, schools and states, they theoretically separate the most accomplished students from the pack for college admissions and special perks. And--if you believe the Blueprint for re-constructing ESEA--standardized assessment data should be the yardstick that rewards teachers and school leaders for their brilliant lessons and school management.
However. The foremost purpose of testing--requiring students to demonstrate applied skills and knowledge--ought to be figuring out what comes next in the instructional cycle. If they've got it, you move ahead. If they're still clueless, you must regroup.
It's pointless to go forward when a significant number of the students you're teaching will be, umm, left behind. You begin the dance all good teachers are familiar with: repeating, revising, reinforcing--while simultaneously extending, enhancing and engaging. Two steps forward and one step back, using different strategies, coming at content from different angles. A worthy curriculum is full of complex, spiraling and overlapping competencies and ideas; reaching true proficiency is always an individual thing and even rigorously tracking students does not produce lockstep success.
This is why we have human teachers.
Reading some of the language in the NGA/CCSSO's Designing Common State Assessment Systems, one might begin to think we won't have this assessment thing under control in America until all public school students are regularly evaluated by identical, efficient, computerized, artificial intelligence-based assessments. Which may give us more precise and reliable numbers to slice and dice, telling us with exquisite accuracy where the poorest and most underserved students are. But don't we already know who those kids are and how far they lag behind the kids in the advantaged schools? The problem isn't acquiring more data; it's raising achievement.
I'm all for richer, performance-based assessment models--for the sole reason that completing authentic tasks not only measures knowledge acquisition, but also reinforces the fact that learning is most enduring and valuable when you can demonstrate some proficiency in using what you've learned. We should be assessing kids regularly, through more authentic models, so we can teach them better--and they can learn more. The rest of it-- group-to-group statistical comparison, economies of scale realized by national curricula, materials and tests, and deciding which teachers need to be fired--contradicts the core purpose of assessment.
Assessment informs instruction. Assessment should also inform students and their parents about how much or little has been mastered. Everyone in the teacher/student/home/school quadrangle can respond: How is this student doing? What does she need next?
My state used performance assessments, hand-scored and tied to the state curriculum framework, for a time. They were expensive, but quickly began to accomplish what they were designed to do: drive instruction.
I administered state tests for years--and remember one in particular, a middle grades science assessment which called for narrow-necked beakers and long, skinny balloons. The task involved developing a written hypothesis about what would happen when the air in the beaker was heated then cooled, performing the lab, answering some constructed-response questions about collecting data via a scientific method, plus some multiple choice items on why temperature change causes gasses to expand and contract.
At lunch, the discussion was about clueless test developers at the State Department who obviously never spent time in a classroom full of middle school boys, watching hot dog-shaped balloons go up and down. But the data returned was valuable. Our kids clearly had absorbed useful knowledge about the process of testing hypotheses. They didn't do as well with remembering the correlated physics content. The test provided a template for further instruction. All good.
We eliminated statewide performance exams, post-NCLB, because all states are doing a lot more testing now-- no matter how useful, they became too costly to design, and the scoring process too slow. Teachers still conduct complex performance assessments in their classrooms--teachers have been doing authentic learning evaluation since Socrates--but you wouldn't know that by reading the NGA/CCSSO document.
And I guess that's what bothers me most about the $350 million set aside for state consortia to develop assessments aligned with the Common Core standards. There's a proposed role for teachers in suggesting authentic performance tasks--and mention made of how funding will be used to "invest heavily in other professional development activities so teachers become more adept at applying a range of assessment strategies in their own classrooms."
Still, it's hard for me to understand how regular, mandated, uniform performance assessments, attached to regular, mandated, uniform standards--all created by politically motivated psychometricians and education publishers-- are going to do anything except continue the professional isolation and de-skilling of teachers.
Common standards and assessments are something we can do, while adequately addressing child poverty and neglect seems impossible. It's the elevation of standardization and stuff over substance.
What do you think about common assessments? Post a comment--or join the discussion at Teachers' Letters to Obama.