We're learning that there are many ways to cheat.
The legitimacy of test score increases in District of Columbia Public Schools (DCPS), in particular those at Crosby S. Noyes Education Campus, are the focus of the latest installment in USA Today's "Testing the System," a multi-part series exploring the extent and causes of cheating by teachers, principals and schools on standardized tests. At the heart of the allegations are multiple erasures -- presumably adults correcting test answer sheets -- that were detected by the test scanning computers that grade the multiple choice tests. In one classroom at Noyes, USA Today reports, seventh-graders averaged 12.7 wrong-to-right erasures per student. Perhaps the students were lucky, but statisticians say they would have been more likely to pick winning lottery ticket numbers than to make that many wrong-to-right erasures by chance.
If the USA Today allegations are true, then the adults who changed students' answers did much more than just cheat on a test. They also cheated those students, by allowing them -- and their families -- to think that they had learned material they clearly hadn't.
But these deceptions are not new. For decades, less was expected from students attending schools in poor communities. Expectations were even lower for children with disabilities or in special education programs. It was virtually impossible to learn how these children were doing. Because scores were reported only as averages, it was easy to mask the fact that entire groups of students were not learning. Educators and policymakers knew--but the absence of hard data made it easy to turn a blind eye.
The federal No Child Left Behind Act (NCLB), which mandated yearly testing and public reporting of schools' results in grades three through eight and once in high school, was written in part to respond to these issues. Lawmakers wanted to ensure that test results would be comparable from student to student and create common standards for all students, regardless of their backgrounds. That's why the law requires that tests align with state content standards and that students be assessed at their official grade level. It's also why the law requires that schools report results for smaller groups of students -- those who don't speak English, for example, or those with disabilities -- separately.
So, before we heed reactionary calls to end standardized testing, it's important to remember the decades of willful neglect prior to NCLB. Most notably, the implementation of standards-based reform, first in the states in the 1990s and then by the federal government under NCLB in 2002, spurred an unprecedented focus on the deficiencies of schools that serve poor and minority students who were long ignored and whose outcomes were mostly hidden from view.
Still, the law is almost a decade old and its flaws are increasingly clear. Few people would defend the quality of most state tests and the low bar that they set to proclaim a student "proficient." Part of the solution is a better assessment system. The Obama Administration agrees and is investing $350 million in two different consortia of states to develop these assessments. Ideally, improved assessment practices will show us not only what students are learning but also how they are learning and why they may or may not be gaining particular skills or knowledge. We also need to continue explorations of data of all types (not just test scores), building on, for example, important research that's helping us develop early warning indicators to prevent students from dropping out.
We need voices and ideas from many places to continue to improve our understanding of how well students are learning in our public schools. But there's no excuse for cheating, whether it's done by children or adults. And let's remember that the system before NCLB also cheated children, denying them a chance to get a good education. Ignoring the past and taking us back to those days -- however you score it -- is definitely a wrong answer.
Follow Bill Tucker on Twitter: www.twitter.com/@billtucker01
Todd Farley: Standardized Testing: A Decade in Review
Cheating fears cast doubt on Rhee's legacy in DC public schools ...
Freakonomics » Have D.C.'s “Best Schools” Been Cheating?
DC School Chancellor: There Was No Cheating | wusa9.com
NBC News EducationNation.com - Cheating scandal rocks D.C. school ...
DC School Chancellor: There Was No Cheating | wusa9.com « Snow ...
The reading portion had fourth graders testing over 3 hours the other day. 3 hours.
This is the flaw with the current system, and I can't conceive of a multiple choice test that could overcome it. The whole underlying assumption, that we can gauge the quality of teaching by testing students, is unsound.
Kid A gets question #1 wrong. Did the teacher not teach the concept, or teach it badly? Was the kid absent for a week when it was introduced? Was he sleeping? Did he not do his homework? Is he just plain dumb? Did his parents feed him on the morning of the test? Or when it was taught? Or did they beat him? Did his dog die the day before? The answer to all of these questions is "maybe."
Kid B gets #1 right. Does he have a great teacher? Does he have a mediocre teacher who happens to have spent a bunch of time on the subjects the test happens to test this year? Does he have a horrible teacher whose kids haven't learned much of anything but how to bubble in tests? Did he learn the answer on Sesame Street? Maybe his teacher is incompetent, but he's got great parents and they worked with him in the evenings to make up for it? The answer, again, to all of these things is "maybe."
For example, research in the United Kingdom is testing how technology-enabled assessments, combined with advanced statistical and cognitive models, allow teachers to identify groups of readers with different patterns ofperformance even though the students’ raw test scores may be similar. Teachers can then tailor instruction to four types of readers—reluctant readers, developing readers, reasoning readers, and involved readers.
As for the multiple choice thing: yes, it's true that tests don't have to be multiple choice. But given the expense involved and the subjectivity introduced when they're not, we can be fairly sure that they're GOING to be multiple choice.
Over a period of 4 years, 100 students who passed their third grade standardized test had Mr X as their fourth grade teacher. Of these 100 students, only 20 passed their fourth grade standardized test.
Statistics is very helpful in making sense of large datasets. If you demand who is responsible for student's A etc, you end up arguing yourself into a knot.
You are absolutely correct in defining statistics to large datasets. You are absolutely correct that explaining individuals is a "knotty" situation with statistics.
So, let's follow your example of Mr. X in the 4th grade. According to a lot of research, a great teach cover about 1.5 years of material in a school year, while a bad teacher covers only 0.5 years. (a Good Teacher - 1.0)
Scenario: 100 students
1st Grade - Great Teacher (1.5 years) (Total = 1.5 years)
2nd Grade - Great Teacher (1.5 years) (Total = 3.0 years)
3rd Grade - Bad Teacher (0.5 years) (Total = 3.5 years)
At this point, it is easy to see why 100% of the students passed. They were well ahead from Years 1 & 2. They were coasting coming into 3rd grade. They could have passed the test at the beginning of the year. Now, the 3rd grade teacher actually regresses them. They lose motivation.
4th Grade - Good Teacher (1.0 years) (Total = 4.5 years)
Back on track, without loss, but let's punish Mr. X. Doesn't make sense. But this is how it is done.
I think if you know the data has been gamed and you neglect to mention that while claiming that it's actually your reforms, you are being dishonest at best, and, outright lying at worst.
Let me put it another way. If this accountability, firing teachers, fighting unions, charter schools, and emphasis on standardized testing is not improving things, why are we continuing to do it? Rhee's results were one of the justifications for doing this stuff, because, we certainly can't look to successful education systems in the rest of the world for proof, they, for the most part, do the exact opposite.
Also, if teachers are responsible for their students, isn't the chancellor responsible for those they are overseeing? Should accountability apply to the admin as well as the teachers? Or, does Rhee not have to hold herself as accountable as those that are being paid far less than her?