New York Teacher logo. 3K gif.
January 17, 2001
The use and abuse of tests; consultant explains right, wrong ways to use data


With more state tests every year, educators must be aware of the real value and potential pitfalls. Testing consultant Giselle O. Martin-Kniep says in this interview with New York Teacher assistant editor Ned R. Hoskin that state assessment results can be good levers for change, but should not be considered targets for change.

Question: What are inappropriate ways test data are used?

Answer: It is inappropriate to isolate teachers at specific grade levels where the test is being given, and then ascribe blame or success to them. That's inappropriate because most state tests are cumulative tests on skills students acquired over time. To assume that a specific grade-level teacher is the one to blame or praise is fairly ridiculous, especially when the test is given early in the school year. Even if it is given late in the year, the skills those tests are measuring are acquired over a long period of time and not moreover in one particular school year.

In addition, test data are being used to infer teacher incompetence, again, in grade levels where the test is given. If those skills are acquired over time, you can't assume (from poor test results) that a specific teacher is incompetent.

(Another issue) is using the data to develop lessons and assessments that are basically teaching and measuring isolated skills or items that are in the state tests. To teach a skill or a concept in a decontextualized, isolated way does not produce learning or transfer of any kind.

Q. What are appropriate ways to use the data?

A. You can use test data to identify the strengths and weaknesses of a program. For example, looking at test data - scores, raw scores, item analysis - you can see where students overall are doing well and where they are struggling. They are very good for letting us see what our weakest students seem to have in common and what they need, and what our strongest students can do or cannot do and what they need.

Test data can be a great diagnostic device to develop intervention strategies and program evaluation strategies, and get teachers involved in action research to see relative merits or effects of their strategies in terms of helping students remedy their limitations.

Q. Can the data help teachers improve practice in teaching?

A. If we were to use test data to identify the lowest performing students - what they know, what they can do, what they don't know, what they cannot do - then we could have teachers pretest those students and develop very specific intervention strategies targeted to the weaknesses of the students. We could help teachers focus on specific strategies. Then we could post-test the students to evaluate which strategies had what effect.

So we can use test data actually to help teachers assess the strengths, the weaknesses, the effects of strategies on different students. That's the best use of test data.

Q. Individual teachers can do that?

A. Teachers could do that if they had access to the test data; teachers could do that if they were given the time to analyze the data and to devise and target specific strategies, and teachers could do that if they were given time to then look at the student work and to look at the impact of those strategies on those students. They definitely could do that. (See sidebar on training, at right.)

Q. What is a useful way to proceed if your students are struggling on standardized assessments?

A. There are three kinds of things that teachers need to do. One is they've got to teach students how to take tests effectively, especially standardized tests. Teach the tricks and rules of test taking - how to identify distractors, how to time one's self effectively, how to make good guesses. They should also help students become very familiar with the form of the tests. Finally, they should address the deficits that the students have, or the skills, the concepts, the knowledge that are obviously weak in those students.

One of the problems I see in many schools is an emphasis on the second kind of test prep, at the exclusion of the other two. There's an overemphasis on getting kids to become super-familiar with the test or with items on that test, and there's not enough attention to helping students take any test well.

Data-use training

A collaborative group of teacher centers, the State Education Department, and its Staff Curriculum Development Network and Special Education Teacher Resource Center is conducting training on the uses of statistical and non-statistical data to drive decisions by teachers that affect student achievement. Contact a regional representative of one of these networks for information.


NYSUT.org. Copyright New York State United Teachers. 800 Troy-Schenectady Road, Latham, New York, 12110-2455. 518.213.6000. http://www.nysut.org. For questions about this web site, contact the webmaster at bthomas@nysutmail.org.