Assessment


Assessment Guru Grant Wiggins on Measuring Student Progress

All the talk about changing the way we measure student progress raises important questions. What do we mean by assessment — as opposed to grading? How do you design reliable assessments? Should you "teach to" standardized tests? How can you evaluate an individual performance on a cooperative project? And how do we explain it all to parents? To get answers, Instructor senior editor Meg Bozzone interviewed assessment expert Grant Wiggins, president of the nonprofit Center on Learning Assessment and School Structure (CLASS).

What's the difference between assessment and grading?

When teachers assess student performance, they're not placing value or judgment on it — that's evaluating or grading. They're simply reporting a student's profile of achievement.

What's a "profile of achievement"?

It's like a baseball card that lists the batting average, runs batted in, and other data about a player without placing a value on performance. In order to evaluate a player's performance, you must factor in other information, such as: Is the player a rookie who has a low batting average in general, but a high average for a rookie? Therefore, for student performance you report about such things as absolute achievement, relative progress, scores for specific writing skills (not writing as a whole), and so on.

Should teachers trade in their traditional assessment methods for profiles of achievement?

It's not "Yeah, portfolios! Boo, multiple choice!" or any other either/or choice. Good assessment is about expanding the assessment repertoire because no single form is sufficient. There are reliability and validity problems with each. Every method has its strengths and weaknesses, and its place.

Are parents comfortable with profiles of achievement?

Parents at schools that have switched from report cards to narratives often say, "This is great, but how is Johnny doing?" I'm sympathetic to them and return to the baseball-card analogy: I don't know how to judge the value of a batting average of .280 unless I have some comparative information. It may turn out that hitting three out of ten, which sounds poor, would land a player in the majors.

In schools, teachers should report the profile of achievement and compare students to standards or norms. For example, a student in a second-grade class may be a year younger than classmates and therefore less experienced. There is no reason to compare this child flat out with others with no qualifying language or context of expectations. At the same time, teachers should report, for example, that a child who may not be a strong writer for his age has made progress in attitude and with some writing skills.

What are ways of putting students' performance in context?

Many schools report performance on a novice-expert continuum. Teachers draw up a rubric that describes what a student should have mastered from the novice to the expert level, with steps delineated and explained along the way. Teachers then put students' learning into context. They can point out to parents, for example, "Here's where we placed your child on this continuum. The evidence we used to support our assessment will help you see why."

The same rubric also can be used to report individual progress. Teachers can say, for example, "Relative to other five-year-olds, your child is still behind. But, as you can see from the last assessment, she has made gains and we're happy with that progress."

How can teachers expand their assessment repertoire?

Begin where you know you have a mismatch between an outcome you value and the way you now assess it. For example, if you know your 20-question multiple-choice quiz on the Civil War is inadequate, experiment with adding another type of assessment that's rich and interesting. What have you got to lose? You're not tossing out the test. You're finding a way to fill in a gap. Also, give students scoring rubrics and other insight into what criteria is applied to assessment — secrecy should be minimal — and have kids practice peer review. Self-assessment and self-adjustment are at the heart of better performance.

Back to top

How will teachers know if an assessment is reliable?

There's a big difference between a neat activity and valid assessment, and not all assessment is hands-on. Teachers need to design assessments backward from the task, asking at each step of the way "What's the evidence I need of children's understanding? Will this assessment get at it?" For example, if you're considering using a diorama of the Civil War as an assessment, ask "Could students do the diorama but not understand the Civil War? Could students not do the diorama well but understand the Civil War?" The answer to both is yes, so it's a good idea to either rethink using the diorama as assessment or use other types of assessments, such as oral presentations, in conjunction with it.

What about assessing cooperative projects?

Many parents worry about this because they want to know the specifics of their child's performance. I'm sympathetic to that, and again return to the sports analogy: Your team can lose ten games in a row, but the data for your best player still stands and vice versa. So teachers need to find assessments — such as presentations by each student on an aspect of a collective project — that pull out information about each child

If a school's policy is to give letter grades only, what can an individual teacher do?

There's nothing to prevent teachers from expanding their modes of assessment while still living in a letter-grade world. Assessment reform is about getting different and richer information about students' performance, all of which teachers can factor into a grade. It's a matter of expanding your pile of evidence, not necessarily changing the grading system.

Where does standardized testing fit in with assessment?

Many teachers think that they have to teach worse in order for their students to get better scores on standardized tests. Not true. The tests are usually simplistic and generic, so if teachers have a rigorous local curriculum and assessment system, their students should do very well. The test designers aren't interested in teaching through the test — all they're trying to do is find the quickest and easiest way of getting at some basic skills. Teachers' standards should be much higher than the test designers' standards, which are minimal.

If you want to know more:

Grant Wiggins & Associates
4095 US Route 1 € Suite 104
Monmouth Junction, NJ 08852

Phone: (732) 329-0641
Fax: (732) 438-9527

http://www.grantwiggins.org

Back to top