Reporter Teresa Hoffman of the Ralston Recorder (NE) notes what readers of this blog have known all along - "Opinions vary on tests".
Some parents just want more of the testing info:
Kriss Kriglstein, a parent of a Blumfield sixth-grader, said she doesn't mind her son taking standardized tests, but said she would like to see more information on those tests given to parents and teachers.
"I don't think we get enough information," she said. "We just get a score and I don't think that helps us as parents."
Though she sees a need for the tests, Kriglstein said she thinks it's more important to look at the results of an individual child or even a classroom.
Some parents feel testing is too expensive and not well-explained to the consumers:
Janis Dwyer, who has a daughter in seventh grade at Ralston Middle School, said she has many concerns about standardized testing.
Among her objections are the focus it takes off of individual students, the cost and the lack of understanding of the scoring system by the public...
Dwyer also doesn't like the cost associated with testing and the fact that districts don't receive help from the state and federal governments even though they are requiring the testing to be done.
"The testing process is an expensive use of time taken from true teaching and learning opportunities in the classroom," she said...
Finally, Dwyer said, there's not enough understanding of the scoring process of standardized tests.
"Percentiles are not the same as percentage grade," she said. "A 50th percentile is average and the average of all numbers should be in the 50th percentile range to find a bell curve, which is the goal of standardized testing."
Unfortunately for Ms. Dwyer, who is a former teacher herself, she illustrates her point well by giving confusing and at least partially inaccurate definitions here.
The 50th percentile is not the average or mean score; it's the median, the point at which 50% of test takers score above and 50% below. The reason it's important to make that distinction is shown in the following example.
We give a test that has a score range of 1 to 100 to two groups of 11 examinees each.
* Group A contains scores of 30, 40, 50, 60, 60, 60, 70, 70, 70, 80, and 90.
* Group B contains scores of 60, 60, 60, 60, 60, 60, 90, 95, 95, 100, and 100.
The median, or 50%, of each group is the same - 60. But the mean for Group A is 61.81, while the mean for Group B is 76.36. What's more, the kid who's at the 50th percentile in Group A is indeed near the middle of his class, while a kid who scores at the 50th percentile in Group B, which is bimodal, is a member of that part of the class which is just not "getting it."
I assume that Ms. Dwyer meant to say that, if the distribution of examinees is a bell-curve, the 50th percentile will be (as it is in Group A), close to the average. However, if a test is criterion-referenced, there's no reason to assume that a particular cohort of students will make up a bell-shaped curve, or anything close to it. There's no reason why, on a criterion-referenced test, to assume ahead of time that a class would look more like Group A than Group B.
Thus, for any parent interested in interpreting test scores, it's helpful to look at the 25th and 75th percentile scores as well as the median and average. For example, in Group A above, the 25th percentile is 50, while the 75th percentile is 70. This means that half the class scores between those two scores. But in Group B, the 25th and 75th percentiles are 60 and 95, respectively.
Thus, a parent whose kid got a score of 60 would know, if his kid were in Group A, that his kid was indeed scoring near the middle of the curve. But if his kid were in Group A, that would mean his kid was one of the bottom performers in the class, even though a score of 60 is at the 50th percentile in both groups.
I think this next statement by Ms. Dwyer is also confusing:
She said because a list of scores showing that many students do perform in the higher percentile ranges is never reported, it leads those outside education to believe that all children are doing poorly.
Um, there have to be students performing in the higher percentiles, so it's hard for me to understand how the scores in the higher percentiles could not be reported. And whether or not all children are doing poorly has nothing to do with ranking them using percentiles. If in fact every student is doing poorly, a kid with a score of 60 out of 100 could be at the 95th percentile. If it's a bell-shaped curve, that kid is likely to be near the 50th percentile. Every test taker will get assigned a percentile rank, but that only measures how they've done in relation to one another, not in relation to the material on the test.
It's hard to see what she means here, unless she is using "percentiles" incorrectly again and is claiming that scores in the upper categories of scores are not reported. But again, it's hard to believe that that is true, or if true, how it would affect public perception. If the newspaper reports only that 70% of students scored Below Basic on a test, so what? No amount of reporting that 30% scored higher than that will dispel the notion that most of the kids are doing poorly.
But, enough of the statistics lecture. The article winds up by noting that others see testing as a necessary evil, and the most they'll say is that they're not against it:
Merry Naviaux, the academic resources teacher for the Ralston School District and president of the Ralston Education Association, said she sees standardized tests as a necessary tool.
"We need to have some kind of testing," she said. "I'm not against it. I think it is really important because it gives you baseline information that is critical to making decisions."
If opinions on testing really vary a lot, and I think they do, couldn't the reporter have found one teacher or parent to make an unequivocally positive comment on testing?
Posted by kswygert at December 19, 2003 04:30 PM