Bill Evers on TV on Standardized Testing

I joined Sean Reardon, an associate professor at the Stanford School of Education, for a half-hour discussion on statewide testing of K-12 students in California.

Within the over-arching topic of testing, Sean & I discuss:

-- Can you test such matters as civic understanding and ...
creativity?
-- Can you have effective school reform in the absence of
data from student test scores?
-- Is money per se the key to effective
school reform
-- Is the racial achievement gap narrowing?
-- Is there a "stereotype threat" that leads girls and African-American and Latino
students to underperform on tests? And if so, what should be done about it?
-- Should California renew its accountability system in 2010? Should it retain its statewide grade-by-grade tests? Its high-school exit exam?

The program is entitled "Class Action." The show aired on Sunday, November 8, NBC Bay Area (KNTV, Channel 11), and the video has now been posted online. The host is KNTV news anchor Jessica Aguirre.

California's Conditions on National Standards

Patrick Riccards, blogging at eduflack, reports that California has signed onto the national-standards effort, but (wisely in my opinion) has placed conditions on its commitment to adhering to the outcome of the process. A decade ago I was a participant in the drafting of California's academic-content standards. These standards have been judged (by the Thomas B. Fordham Foundation and by the American Federation of Teachers) as among the clearest and most rigorous in America , and it would be a shame to water them down in the name of national commonality. Congratulations to Governor Schwarzenegger and California ed officials for their sensible letter setting forth the conditions on the state's involvement.

Must We Give Up High School Exit Exams?

Education school researchers at U.C. Davis and Stanford assert that California's high school exit exam is keeping girls and non-whites from graduating from high school.

The study speculates that girls and non-whites have been failing the exit exam in high numbers because of a "stereotype threat." "Stereotype threat," as suggested by psychologist Claude Steele, is a situation in which blacks, women, etc. are put off their pace by awareness that others expect them to fail.

The Los Angeles Times story on the study paraphrases Sean F. Reardon, one of the study's co-authors:

Reardon said there was no other apparent reason [than stereotype threat] why girls and non-whites fail the exam more often than [those] white boys who are their equals in other, lower-stress academic assessments.

Here's what the executive summary of the study says:

Among students in the lowest quartile of achievement, the CAHSEE [exit exam] requirement has no effect on the graduation rate of white students, but a large negative effect on graduation rates of black, Hispanic, and Asian students.

On average, among students in the bottom quartile of achievement, graduation rates were 19 percentage points lower among Black students, 15 points lower among Hispanic students, and 17 points lower among Asian students who were subject to the CAHSEE requirement than among similar students not subject to the requirement.

Likewise, graduation rates were 19 percentage points lower among female students, but only 12 points lower among male students who were subject to the CAHSEE requirement than among similar students not subject to the requirement.

I wonder about several things:

• In theory, the stereotype threat should not be lowering scores for Asians.
• It should not be lowering scores for boys in the lowest quartile almost as much as for girls.
  
• Why is the stereotype threat working only on the lowest-quartile students? In theory, it should be lowering scores for blacks and girls in all quartiles on this high-stakes test.

There would seem to be many difficulties for stereotype threat as the explanation for the results.

(The authors anticipate these objections and attempt, not very persuasively, to respond to them in their paper.)

Without getting into my skepticism about some aspects of the stereotype-threat effect, let's assume that it's true or can sometimes be true. The need then is to accustom blacks and women to competition and challenges, and the potential would seem to be there for greater success when people have instead high, demanding expectations about blacks and women.

The solution cannot be to cushion blacks and women from all challenges (like passing the not-very-tough high school exit exam). [This is a point that has been made about the stereotype threat by John McWhorter both in interviews and in his book Losing the Race.] Life is full of challenges.

Getting rid of the high school exit exam cannot be the solution. The solution has to be preparing low-performing students to pass the exam and telling them that their teachers, parents, ministers, and other community leaders expect them to succeed and will accept no excuses.

California State Supt. Jack O'Connell and others put the exit exam in place in order to have a meaningful high school diploma. It will help the students in life if they reach at least the foundational levels of knowledge that the exit exam measures. The Los Angeles Times describes those foundational levels:

The exit exam, which students can take multiple times beginning in their sophomore year, includes math and English tests, with the math aligned to eighth-grade standards and English to 10th-grade standards.

Abolishing the test and handing out meaningless diplomas to uneducated students will not really help them.

High achievement on tests stems from diligent effort and effective study habits. These need to be persevered in over time, so that knowledge, skills, and understanding can accumulate and be stored in long-term memory.

Blaming all problems on a stereotype threat and then giving up all high-stakes tests (as Reardon and others have suggested) will solve nothing.

[Thanks to Ze'ev Wurman for his advice and counsel on this post.]

UPDATE (4/23): A better place to look for advice is the 2008 report of the Public Policy Institute of California (PPIC) called "Predicting Success, Preventing Failure: An Investigation of the California High School Exit Exam," by Andrew Zau and Julian Betts. This report says that after-school reading programs and the use of peer coaches in professional development -- both instituted in San Diego during the Bersin-Alvarado years -- could help students pass the exam. It also says that the evidence suggests that intensive tutoring focused on academic content is likely to be more effective that less-intensive tutoring focused on test-taking strategies.

The Zau-Betts report also proposes spotting students early who are at risk of failing the exit exam. Then schools should provide them with academic tutoring in elementary and middle school -- rather than waiting until high school when making up the academic deficit is much more difficult.

There is room for improvement on the particulars of what San Diego did, but the Zau-Betts report points out an evidence-based path to better exit exam results: early detection, supplemental reading instruction, and intensive tutoring on academic content.

California Schools Superintendent Wants to Water Down Academic Standards in Name of "21st-Century Skills"

California State Schools Superintendent Jack O'Connell spoke to the annual EdSource Forum in Irvine today (April 17).

O'Connell, who holds a nonpartisan office, began his speech with political partisanship:

President Obama won a mandate for change that has placed him in a position to cause a massive shift in the way our government operates and in the manner in which it serves the needs of its citizens....

In just the first few months of this Administration, I can easily and confidently say that we have seen a dramatic shift in the willingness of this White House to be a partner to states — this is a welcome difference from the previous Administration....

There was more, but you get the general idea.

O'Connell then went on to identify "four key areas" that the Obama administration wants states to concentrate on:

• Adopting internationally benchmarked standards and assessments...;
• Building high-quality data systems that track students' academic careers, making it possible to tell which teachers, programs and schools are effective;
• Recruiting more high-quality educators to underperforming schools, as well as to subjects like math and science; and
• Supporting effective strategies to turn around underperforming schools.

Superintendent O'Connell proceeded to voice views that reflect the standards-and-accountability consensus on most of these matters.

I should point out that he was particularly careful to note that any "creative approaches" and "creative solutions" on teacher pay would be done only "in partnership with all the professionals involved in teaching and learning."

A flag of caution almost went up when the Superintendent said that we need to test California students more "comprehensively" than we do now. Often such phrasing is coded language for statewide portfolios or project-based testing. But then the Superintendent went on to promise that any revised testing would be as "valid" (a technical term in testing) as the current California standards-based tests and the high-school exit exam (of which the Superintendent is rightfully the proud author).

But there was one place where Superintendent O'Connell set off alarm bells: His desire to revamp and "re-up" California's academic-content standards.

The Superintendent acknowledged that this is "an area in which California has served as an established leader for some time now":

I am proud, as all of you should be, that California purposely set the bar high for what our children should know and be able to do, and we have never wavered from that commitment.

But instead of maintaining these internationally-benchmarked high standards, he wants to "rededicate and reinvigorate" them. "Evolving them is healthy," the Superintendent says.

Superintendent O'Connell wants to revamp the California academic standards because, he says, they are “a mile wide, but only an inch deep.” But, with all due respect -- as I have stated in the Los Angeles Times -- I have to disagree. Indeed, the California standards have been judged among the best in the country by the Fordham Foundation and the American Federation of Teachers. California already has official Key Standards in mathematics, for example, to identify teaching priorities and the topics that should be covered in depth.

The Superintendent also wants to revamp the California standards:

...in order to fully engage both students and teachers in the learning process in a way that sees both parties benefit and helps to better prepare students for success in the economy of the 21st century.

Translation from education jargon: He wants to water down California's existing high standards in the name of the wolly concept of "21st-century skills," that is, communicating with each other, working in groups, media literacy, and so forth. He wants to subtract from classroom time spent on solid subject-matter content to teach these supposed stand-alone skills

Unless advocates of solid academics speak up, California will -- in the name of 21st-century skills -- follow Massachusetts down the primrose path of diluted standards.

Should 2nd graders be tested (in Calif.)?

California state Sen. Loni Hancock has once again introduced legislation to end the testing of second graders.

Hancock cites too much "teaching to the test" and wants, she says, to reduce bureaucracy and to "let teachers teach."

But, Sen. Hancock, if I may address you directly, the principal purpose of state accountability tests is to see where schools are not performing well. We need to know this about how schools are doing with 2nd graders.

Class Size

A letter to the editor in today's New York Times (March 1), cites large K-12 classes during World War II as evidence that children can thrive under such conditions. The letter-writer is responding to a Feb. 22 article in the Times, which said:

...[A]s many states and school districts have rushed to reduce class sizes, usually to the low 20s, student achievement has not consistently improved markedly.

In California, a 1996 law provided schools with an extra $1,000 in state money for every student in the earliest grades whose classes had 20 or fewer students. The state quickly hired 28,000 new teachers, but many of them lacked experience or education credentials; a 2002 study by the Public Policy Institute of California found that the best-qualified teachers fled poorer urban schools as the extra funds created jobs in wealthier areas, and that children who were in smaller third-grade classes did not have higher scores on fifth-grade tests.

In New York City, an Education Department comparison over the last two years between school report-card grades and average class size has found little correlation; in many cases, schools with better grades have bigger classes....

Reading First in California

The "Year 6 Report" from Educational Data Systems looks at more than 800 California schools that have received Reading First money since 2003. The report measured student achievement using scores on California's standards-based tests. The Reading First schools had "significantly higher gains" than a control group. The gains persisted into grade 5, even though Reading First funding is only in K-3. English-language learners as well as native speakers had higher achievement in California's Reading First schools. The Report includes a meta-analysis of the effect of Reading First which indicates the overall efficacy of the program.

Here are the first six out of a total of fourteen findings in the Report:

Finding #1: Reading First has had a significant impact on student achievement in California.
Reading First schools have grown significantly since the inception of the program. They have grown significantly more than a statistical control group and significantly more than non-Reading First schools. High implementing schools have grown significantly more than low implementing schools....
Finding #2: The Reading First effect is meaningful....
Finding #3: Growth remains significant....
Finding #4: The Reading First effect generalizes across student performance levels.
Students in all performance levels show a boost from Reading First implementation. The advantage over non-Reading First schools is especially pronounced for students in the “Below or Far Below Basic” categories. On the grades 2-5 California Standards Test (CST) achievement metrics, the migration of students into “Proficient and Above” is matched or exceeded by a migration of students out of “Below or Far Below Basic”. The migration of students out of “Below and Far Below Basic” is more than twice what it is for non-Reading First schools....
Finding #5: Reading First significantly impacts grade 4 and grade 5 performance....
Finding #6: The Reading First effect generalizes to English learners.
English learners in Reading First schools show higher rates of growth than English learners in non-Reading First schools across the state....

Hat tip to "Curriculum Matters" blog.