Network News

X My Profile
View More Activity

Posted at 2:55 PM ET, 10/28/2010

How to evaluate students: Look at their work

By Valerie Strauss

This was written by Monty Neill, interim executive director at The National Center for Fair & Open Testing, known as FairTest.

By Monty Neill
In June, I outlined a school evaluation system to replace No Child Left Behind’s test-only accountability structure. I then described in more detail how each component would work: first school quality reviews, then annual state tests in a few grades, and now local assessments. Together, these interrelated elements provide comprehensive evidence of school progress, as well as richer information for teaching and school improvement efforts.

The best way to find out what students know and can do is to look at their actual work. In a local assessment system, teachers document student products and processes. Research projects, oral presentations, essays, problem solving using computers, and science experiments allow evaluation of higher order thinking skills and deep content knowledge that standardized tests cannot measure.

What is standardized in this system is not individual student work but the criteria for gathering and evaluating work products. With a clear structure, the work assessed can vary: what books students read, the specific choices of what to emphasize in history or biology or math, the extended projects students undertake, etc. Strong structures already exist, including the Learning Record and the Work Sampling System. Other countries have developed various approaches to this issue.

Classroom-based assessments can be adapted to students’ varying needs while maintaining high standards. Rescoring can ensure teacher accuracy and comparability across schools and districts. In essence, trained readers rescore work from a random sample of students from each classroom. This review provides useful feedback to the originating teacher, score adjustments where needed, professional development for the readers, and trustworthy data across a jurisdiction. It is a “trust, but verify” system.

Some argue that local assessments mean different standards in different places. Accountability, they say, cannot rely on classroom evidence. In fact, research in other nations and this country shows that such data can be gathered and evaluated with a degree of consistency more than sufficient for statewide comparability.

Teachers assess frequently, but they need ongoing education to develop their capabilities. And since no one teacher can develop all the good assessments she might use, the state should create an electronic library of high-quality projects and measures. These would be reviewed by expert educators. Teachers would access them as needed. Using these instruments also will contribute to ensuring the quality of evidence of student learning.

Attention to local assessment improves teaching, forges stronger communities of educators, and produces better student outcomes. Nebraska built a statewide system of local, state-approved assessments. The New York Performance Standards Consortium replaces state tests with a mix of school- and consortium-based performance tasks.

In this proposed system, schools would produce an annual report. It would include evidence of success and ongoing problems, along with improvement plans. Documentation of student learning across the curriculum would become publicly available. Such reports could be discussed by the school’s community and reviewed by higher governmental authorities.

One hundred fifty-three national education, civil rights, religious, disability, parent and civic groups have signed the Joint Organizational Statement on No Child Left Behind. It directs Congress to “help states develop assessment systems that include district and school-based measures.” Congress should fund states to construct such systems.

Rep. John Yarmuth introduced legislation in 2007 toward this end, and Rep. George Miller included it in his 2007 “discussion draft” for reauthorization of the Elementary and Secondary Education Act, better known as No Child Left Behind. Assessment reform advocates should encourage their congressional delegation to support this approach.


Follow my blog every day by bookmarking And for admissions advice, college news and links to campus papers, please check out our Higher Education page at Bookmark it!

By Valerie Strauss  | October 28, 2010; 2:55 PM ET
Categories:  Assessment, Guest Bloggers, Learning, No Child Left Behind  | Tags:  fairtest, learning record, local assessment, monty neill, nclb, nclb reauthorization, no child left behind, problem solving, standardized tests  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Poll: How science should be taught
Next: And now, a tragicomic conversation on 'differentiation'


I wish my school would go with this.

Instead we're crunching data and looking at RtI.

Posted by: neaguy | October 28, 2010 4:44 PM | Report abuse

"annual state tests in a few grades..."

I don't think so. NO MORE TESTING!!! Teachers are qualified to assess their own students' learning.

Posted by: PissedMama | October 28, 2010 7:51 PM | Report abuse

"In this proposed system, schools would produce an annual report. It would include evidence of success and ongoing problems, along with improvement plans."


And the school production of these annual reports would be honest, unbiased documentation of how the school actually performed?

Massachusetts has something like this already in place in the form of mandatory "School Improvement Plans,;" the majority of which are fraudulent and designed to make everyone look good. They aren't worth the paper they're printed on.

Posted by: phoss1 | October 28, 2010 9:00 PM | Report abuse

This post gets a giant, "DUH!!!" "HELLOO?!"
Throughout my career I've always expected my students to 'show their work' in Algebra, which allows me to assess more critically their knowledge, skills, and reasoning. Today, though, was another bad example of 'student data' we use regularly. I observed during our monthly 'benchmark' tests anywhere from 15-40% of students per period guessing, or worse. When we analyze this data subsequently their results would appear as if they answered incorrectly from their knowledge and skills, e.g., "they choose 'd' because they forgot the negative," which could be misleading completely if they chose 'd' because they just guessed.

What's right isn't often expedient or cheap, and that's another 800 lb. gorilla in the room few people want to face.

Posted by: pdexiii | October 28, 2010 9:29 PM | Report abuse

The reference to "trained readers" brings out the cynic in me. The problem with qualitative assessments is that there will be some teachers making these assessments who lack the skill and judgment to differentiate between the good and the bad except, perhaps, in the crudest terms.

Well-designed questions in multiple choice format are the most reliable assessment tool, but the key here is "well-designed." If school systems administer seven "interim" assessments and one or more end-of year assessments every year, there's just no way the tests are going to be valid or reliable -- and we're not even talking about the effects of testing fatigue on student performance (or is it a healthy refusal to perform on cue?). Your suggestion in an earlier piece of limiting standardized tests to a few grades makes the most sense.

Ed reformers need to stop focusing on the tail end of the process anyway and turn instead to the curriculum itself. Teachers will be the first to say that typically they are assigned classes at the last minute, often without clear learning goals and baseline teaching materials. (I recently heard about an AP World History class in Houston that used the "Lion King" to teach about Africa!) Public schools across the country will never achieve consistency, much less high-quality instruction, with this seat-of-the-pants approach.

Posted by: Jennifer88 | October 28, 2010 11:11 PM | Report abuse

These tests are designed to be tricky with 2, 3, or even 4 plausible answers. If too many kids get a question right, then they throw out the question for the next year. If more black kids than white kids get the question right, then they throw out the question. The gap is maintained for a reason. It keeps the test publishers in business.

Posted by: PissedMama | October 28, 2010 11:17 PM | Report abuse

Education policy makers and lobbyists write legislation to enrich the testing conglomerates, subcontractors, some university professors, and the canned curriculum industry. As a result, K-12 students enrolled in public schools and their teachers are held hostage to generate profits for the insiders.

Making the Grades: My Misadventures in the Standardized Testing Industry, by Todd Farley, provides a behind the curtain account about testing schemes across America.

Posted by: nfsbrrpkk | October 29, 2010 7:49 AM | Report abuse

"The best way to find out what students know and can do is to look at their actual work."

I have to agree with pdexiii on this one.

The education reform has gotten so turned around and confused that going back to the old standard way of assessing students by looking at their knowledge and work is a radical idea.

Money won't help. Tougher standards, harder tests, and higher level materials are the key. Our standards are too low, which helps no one.

Posted by: hebe1 | October 29, 2010 3:25 PM | Report abuse

The Chicago Sun-Times is rating CPS's schools by test scores. Such foolishness.

Posted by: educationlover54 | October 29, 2010 6:27 PM | Report abuse


Do you have a link? Thanks.

Posted by: educationlover54 | October 29, 2010 6:29 PM | Report abuse

Nice title. What a concept!

Posted by: celestun100 | October 29, 2010 9:40 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2010 The Washington Post Company