Trust Your Teachers, Not the Testers

As No Child Left Behind comes up for re-authorization (or termination, please!) this fall, I happen upon a study that calls into question the value of the standardized tests that are so central to the thinking (and the private profit potential) behind the law. [Credit to Andy Hilbert, a California teacher who blogs at HorseSense and Nonsense; he links to Research that Counts, where an abstract and link to the study of interest appear.]

Saul Geiser and Maria Veronica Santelices [University of California, Berkeley, "Validity of High-School Grades in Predicting Student Success Beyond the Freshman Year: High-School Record vs. Standardized Tests as Indicators of Four-Year College Outcomes," Research & Occasional Paper Series: CSHE.6.07, 2007] looked at the high school GPAs and SAT scores of 79,785 students admitted to the University of California system and compared those data to two basic measures of student success at university: university cumulative GPAs and four-year graduation rates. Geiser and Santelices found that high school GPA more reliably predicted student success at university than any of the five SAT scores used in UC admissions (Geiser p.9, Table 4; HSGPA also more reliably predicted student success than parents' education, family income and the "Academic Performance Index" rank of the students' high schools). The study also notes that standardized tests show more bias against minority students than high school GPA.

These results support the position that the Madville Times has advanced previously: if you really want to know how well your school district is performing, it's silly to rely on measures like the standardized tests required by No Child Left Behind and prepared by for-profit out-of-state companies. The people with the best perspective on how well your school district is performing are the people in your school district, the teachers and administrators who actually work with your kids. In the name of "accountability," No Child Left Behind imposes a lot of extra expense on school districts (money that ends up leaving the community to pay testing companies and outside consultants to process the results). Yet standardized tests don't provide the best measure of what a majority of parents are interested in: "Is Junior learning enough to succeed in college, or are we going to waste a bunch of tuition dollars?"

Now, before you rush over to Stephanie Herseth Sandlin's next August recess event and hand her this study as proof positive that she should vote to kill No Child Left Behind, keep in mind that this study doesn't look at NCLB tests. The Geiser study looks at scores from the SAT,which is an aptitude test, purporting to measure a student's potential for success in college. No Child Left Behind requires achievement tests that measure what a student as already learned. The SATs are used nationwide; the testing companies craft different tests for different states, so it will take a lot of researchers and research dollars to do comparable studies on the gaggle of tests proliferated by NCLB.

Nonetheless, the Geiser study does put the lie to the conventional wisdom that local GPAs are less reliable than standardized tests. The SAT, as the grand-dame of all standardized tests, fails to do exactly what it was created to do: to provide a measure of student success more objective and reliable than can those darned teachers, who surely must be biased in evaluating their students and can't be trusted to provide good scientific data. Hmm. Turns out the teachers are providing better data than the well-funded College Board after all. Maybe before we reauthorize No Child Left Behind and impose further expense on our school districts, we should ask for some data that shows the testing companies are providing us with instruments that give any better measure of our students' achievement than the tests and assignments we're already paying our teachers to create. If we don't have that data, then maybe we should scrap No Child Left Behind and get back to trusting our local teachers.

But be careful -- start trusting your teachers, and then you might start getting even more radical ideas, like paying your teachers what they are worth....

p.s.: For more on why we should let No Child Left Behind die this September, see Edward Humes's review of Linda Perlstein's new book Tested ["Review: Linda Perlstein's 'Tested' Examines School," Los Angeles Times, 2007.08.12, posted on Newsday.com]:

The endless regimen of testing, drilling, report filing, student bribing and student berating that Perlstein describes could only have been conceived by politicians and ideologues who rarely set foot in actual public schools (and would never subject their own children to the Frankenstein classrooms their policies have created)....

...Perlstein depicts a school obsessed not so much with educating as with measuring education, and with doling out a kind of pallid simulation of knowledge. Stories, for example, are always analyzed for their structure, almost never for their actual content....

...Creativity and spontaneity only get in the way of data collection. And so the author treats us to the awful moment when bright kindergartners identifying long vowel sounds are told to stop - because rigid lesson plans say they are supposed to know only short vowel sounds....

...The school lavishes attention on troubled and unruly children, while the most gifted and co-operative are ignored, one of No Child Left Behind's most destructive unintended consequences.

Reviewer Humes notes that Perlstein doesn't connect the dots as damningly as she ought, but the observations do that job for us: No Child Left Behind kills local control of education, and often education itself. It is predicated on the notion that we can't trust our local teachers to do their job and report fairly on the results. It dooms certain schools to failure and "takeover, closure, or privatization" [Humes]. Let's eliminate NCLB now before it does any more harm.