‘Tests have error,’ test author warns

Print More

Getting into high school in Chicago depends on a 40-minute reading test and an 80-minute math test. Most transition center staff say that’s a bad way to make such an important decision, and test makers agree. “It’s just that one test, that one day, that one score,” laments math teacher Darlene Bell of Proctor Transition G in Englewood.

“Retention of a student is much too important a decision to be left up to a teacher saying, ‘This work is good enough to pass,'” counters Chief Education Officer Cozette Buckney. In contrast, the Iowa Tests of Basic Skills are objective, she says, and enable the board to set a uniform standard across the district.

But Bell feels that the test does not accurately reflect the achievement level of some of her students. Some of her low-scoring students don’t belong in remedial math, she contends. “The work is too easy. They’re frustrated. Why put them through basic skills again when I know they’re going to do well in algebra?”

On the flip side, she has had students who got lucky on a retest and advanced to algebra even though they had not performed well in elementary-level math. “If they fail my class during remediation, when they come to my algebra class, guess what? They don’t pass. That’s why I feel there should be teacher input.”

John Goldwyn, a social studies teacher at Sengstacke Transition D, says that typically he’ll look at a student’s score and say, “Yes, that is about right.” But he objects strongly to ignoring everything else a student does. “We don’t assess people in society based on one day,” he notes. “We are judged on the accumulated work that we do.” He thinks that’s an important message to send to transition students, who have trouble with sustained effort.

Sometimes the test-score requirement lets transition students off the hook, teachers note. Jena Payne, a reading teacher at Transition D, knows of two “who came to school once a week to keep from getting dropped, took the ITBS, passed and moved on.”

Teachers would like to use several criteria in addition to tests, including grades, samples of students’ work and faculty recommendations.

“I don’t think the ITBS should be the sole criterion,” concurs H. D. Hoover, senior author of the ITBS. “Tests have errors in them.”

He refers to his publisher’s manual, “Sound professional use of educational and psychological tests means that all test users must … avoid labeling students on a single test score.”

A student could get one score one day and another score another day, Hoover explains.

Indeed, for each test, test-makers calculate a standard error of measurement that shows how a student’s score can vary simply due to chance.

For example, the average 7th-grader in the second month of the school year ideally would score a 7.2 on the ITBS reading test. If the test’s standard error of measurement is sixth months, which is not unusual, the student’s score could be expected to fall anywhere between 6.6 and 7.8—six months on either side of his “true” score. However, it’s possible for the student to score even further away from his “true” score.

If 30 students reading at the 7.2 level took the this same test, 20 likely would score between 6.6 and 7.8, and 10 likely would score outside that range. By definition, a student or group of students, at any given time, have a 2-to-1 chance of scoring within the range established by the standard error measurement

In the Spring of 1997, Chicago’s 8th-graders took an ITBS reading test with a standard of error of about six months, according to Hoover.

However, Hoover notes that teachers’ opinions are not foolproof either. “Any sole criterion in a situation like this will be unfair to some kids,” he says, recommending a combination of tests, teacher input and other assessments. “The more sources of information you bring to bear on this situation—once you have decided to make such decision—the better off you are.”

Hoover stresses that he is taking no position on the policy of retaining students.

The board has set up a “waiver” process to consider information besides ITBS scores. A school principal or a parent who sees reasons for promoting a student who failed to meet cutoff scores may make a case to his or her regional educational officer (REO). Few have been successful. Last August, each of the six REOs granted only a handful—if any—waivers, REOs told Catalyst.

“The only thing I granted was the opportunity to retake the test if students were sick or there was a death in the family,” says Eva Nickolich, REO for Region 1. “Or for a student [whose scores] had been on grade level, and all of a sudden there was a drop of a year and a half. I asked the school to retest that child.”

Nickolich says that many waiver requests had little substance. “A teacher may write a note and say, ‘John did fine work in reading,’ but I look at the grade and say, Wait a minute, if he’s that good, why did he get a D in reading?”

“Teachers defend their young ‘uns,” agrees REO Jose Rodriguez of Region 4. Barely missing the cutoff score “and being a very fine, hardworking youngster was the principal reason for the request. They say, ‘Well he only missed it by a tenth [getting a 7.1].’ I say, yes, but he should have scored 8.8. That child is almost two years behind.”

When the School Board first set test-score minimums for admission to high school, it required a 6.8 in reading and math; that number represents how well the average student in the eighth month of 6th grade would score on the 8th-grade test. Later, it raised the bar to 7.0 and then to 7.2. At the end of this school year, the required score will be 7.4.

Transition center students, however, do not face a rising bar; they must achieve only the standard that was in effect the year they left 8th grade. For students in their first year at the transition center, that’s a 7.2. For students in their second year, the standard had been 7.0, but that was lowered to 6.8 in January.

Because Chicago gives the ITBS in the eighth month of the school year, a score of 8.8 is the national average that is uses for the 8th-grade test. (If it gave the tests in April, the national average would be 8.7.) Popularly, 8.8 is referred to as “grade level” and any score below it as below grade level.

“Below grade level is sort of a silly idea,” contends Hoover. Grade level is simply the average score among the students chosen to take the test for norming purposes, he explains, adding that it is not an absolute standard.

“If I took a typical group of 8th-graders, they have a lot of variability,” Hoover continues. “I could take some 8th-graders and say their grade equivalent in height is 7.2, but that’s just a description of where they are on the developmental continuum. It’s normal.”

There is one group of students who is judged on more than test scores. Eighth-graders with acceptable test scores who have excessive absences or fail reading or math are required to attend the summer Bridge Program. If they don’t show up, they are assigned to transition centers rather than high school. This fall, 96 transition students fell into this category, according to Joyce Bristow, who oversees the program.