In 2004, to meet a No Child Left Behind Act mandate that states must have tests that closely match what children learn in the classroom, the Stanford test was selected as the state's only test. Up to that time, the state allowed schools to pick from four tests.
Last year was the first time the Stanford test was used, and the number of students who passed and were no longer English-learners doubled to 31,118 from 15,268 the year before.
The sudden surge alarmed teachers. They say that the big gains are
the result of a new test that is easier for students to pass and that
some students who pass are not ready for the regular classroom.
Arizona Department of Education officials say the test is not easier. They say the numbers are up because the Stanford test is more in line with what students learn in class than previous tests. Schools also are held more accountable for their numbers than in past years. They are no longer simply allowed to self-report the total number of students who pass the test each year. Each child's test score is now in the Education Department's computer system.
The test has far-reaching consequences for students, teachers, schools and taxpayers. Arizona uses the results to determine whether a student is in the English-learner program and qualifies for special services. At stake are thousands of dollars in state money because schools get an extra $358 for every student who is an English-learner, and the test alone determines whether schools get this money. Schools spend the money on everything from supplementary books to specially trained teachers.
Education Department officials say the Stanford test is valid, reliable and reflects the language skills students learn in the classrooms. The Stanford tests students in listening, reading, writing, speaking and comprehension. Students who pass are considered to have enough language skills to be ready for the regular classroom.
Department officials say a single statewide test provides a better measure of proficiency than the multiple tests used in the past. But instead of providing a better measure, the switch to a single test has led to more questions. The central questions that swirl around the test are whether the new test is easier to pass than previous ones and whether one test should be the only factor to decide if a student is no longer an English-learner.
Is the test easier?
To show how dramatic the change was in the number of students passing
the test, consider what happened in the Phoenix Union High School
Two years ago, 638 students, or 14 percent, of the students who had been in English-learners programs passed the test.
Last year, taking the new Stanford test, 1,143, or 26 percent, of English-learners passed.
"That's a big leap," said Joan Mason, who directs English-learner programs. "And although we'd love to take credit for having improved our program so much, we don't think it would have inched up that dramatically."
Instead, Mason and other school officials have a different theory. They say the new test is easier to pass.
The test uses a composite score or total score. So it's possible for a student to score "proficient" overall without scoring proficient on every skill. One example: A student can score "proficient" in speaking but score below proficient in reading and writing and still pass the test, because the three scores are added up for a single composite score. The previous test used in the district, the Idea Proficiency Test, required students to pass all sections - reading, writing and oral - to pass the entire test.
In the Phoenix Union district, 56 percent of students who passed weren't proficient on one or more parts of the test, Mason said.
Overall, the district's total number of English-learners is down 765 students, which means the district also stands to lose the $358 per student the state pays for English-learner programs, or about $273,870.
Once students pass, they are no longer eligible for the English-learner program. So, for example, a student who passed with a "proficient" score in speaking, but who scored below proficient in reading and writing, cannot work on those deficiencies in an English-learning program.
Passing the test, even with the composite scoring system, means that student is considered able to keep up in a regular classroom.
Those students are still eligible for some services such as after-school tutoring and summer school.
In the Phoenix Union district, English-learners take up to three 55-minute classes a day designed to develop their English. In math and science classes, teachers use strategies such as special textbooks to help them understand the material.
District officials are still analyzing how these students are doing in regular classes based on test scores and grades, Mason said, but anecdotal comments from teachers indicate some are falling behind.
This concerns Arizona attorney Tim Hogan, who originally filed the Flores vs. Arizona lawsuit in 1992, in which a judge found that state funding was inadequate to help students overcome language barriers.
Hogan contends the state Department of Education violates federal law because students are taken out of the English-learner program when they are not proficient in all three areas of speaking, reading and writing.
"I think it's inevitable those kids are not going to succeed in the school system," he said.
Hogan filed a complaint in August with the Arizona State Board of Education and asked members to stop the test. When they didn't, he filed a complaint in November with the U.S. Department of Education Office for Civil Rights. The office recently launched an investigation.
State Department of Education officials say the main reason that more students pass the test is because the Stanford is better tailored to measure the language skills taught in Arizona classrooms than other tests used in the past.
They compare it with a written test an adult takes at the Motor Vehicle Division. If the test asked questions about car parts and how an engine works, most people would fail because they wouldn't study for that. But if the test is on proper lane use, parallel parking and speeding laws, they will do much better because that is what they studied.
State Superintendent of Public Instruction Tom Horne said children shouldn't be in the English-learner program more than one or two years. The goal is to get them to learn English as soon as possible.
Some school districts have kept students as English-learners for years, he said, because there is a financial incentive to do so. The students are referred to by school officials as "lifers."
"We are putting a stop to that, and some of them are complaining," Horne said.
He called Hogan's complaint "completely lacking in merit."
Horne said the U.S. Department of Education required Arizona to adopt a language test that was in line with what students learn in Arizona classrooms.
The state had neither the time nor the money to develop its own test, so it took bids from companies, Horne said, and the Stanford test was the closest match.
Used in other states
The Stanford test is published by Harcourt Assessment, the same company
that offers the well-known Stanford Achievement Test Series or Stanford
10. The Stanford language test is used in seven states besides Arizona.
When it comes to what scores indicate proficiency, the test company does not determine the pass-fail scale. States set the pass-fail scales for the test. Students earn points in eight categories on the Arizona test. Within each category, they receive one of five performance levels based on the number of correct answers. "Pre-emergent" is the lowest level, and "proficient" is the highest. The pass-fail score differs depending on the grade.
Harcourt spokesman Rick Blake said the test was developed using educators who are experts on English-learners.
"We stand behind the test, and we stand behind the accuracy and appropriateness of the different levels," he said. "The test does, in fact, validly and reliably measure proficiency."
Pennsylvania and Virginia are two states that use the Stanford test.
Unlike Arizona, Pennsylvania doesn't attach an extra dollar amount of state funds to English-learners. And Pennsylvania doesn't rely on the test alone to determine whether a student leaves the English-learner program. School officials also look at the student's grades and scores on statewide academic tests. Schools officials, not the state, decide whether a student is ready for the regular classroom.
"That is a decision made by the people who best know the student in the local education agency," said Barbara Mowrey, who oversees English-learners for the Pennsylvania Department of Education.
Mowrey said the state has not seen a dip in the number of English-learners since they began the Stanford test in the 2002-03 school year. Last year, the state had 42,542 English-learners. The most common language is Spanish. She was unable to provide comparisons on the passage rates from previous language tests because before 2002, districts used their own tests.
Virginia allows school districts to use the Stanford or their own test provided they get the test approved by the state. All but one school district has chosen the Stanford test, said Val Gooss, who is in charge of the English-learner program for the Virginia Department of Education.
Virginia doesn't attach extra state money to every English-learner.
Virginia also requires schools to look at a student's other standardized test scores in addition to the Stanford score to determine whether they can leave the program.
Students who pass the test are still required to take the reading and writing parts of the Stanford once a year for two years as a way to monitor their progress. Last year, 31 percent of English-learners were proficient enough to join the regular classroom, she said.
Unlike Virginia, Arizona students aren't required to take the Stanford test again once they pass.
Margaret Garcia Dugan, Arizona's deputy state superintendent of public instruction, says it would be a waste of time and money to retest students. Students don't lose their English proficiency, she said, unless they move back to their native country and stop speaking English.
Garcia Dugan also said Arizona's policy of using a single test gives more consistent results than if schools were allowed to use multiple tests as they did in past years. Before the Stanford, each district usually selected one of four other tests.
If the state allowed multiple tests, they would have to do expensive correlation studies each year to compare student progress on each set of tests.
In Arizona, there is help available to students who are no longer English-learners. Their progress is monitored for two years, and they are eligible for other services such as after-school tutors and summer school.
Test is subject of study
As part of their academic research, two educators from Arizona State
University and one from State University of New York Fredonia are
conducting a validity study to see whether the Stanford test is an
appropriate way to determine whether students should leave the
English-learner program. Because the test is so new, published for the
first time in spring 2003, they haven't found any external studies.
Their results are expected in April, but so far, their preliminary findings from one Phoenix-area school district show that 40 percent of students who passed the Stanford failed AIMS math, 36 percent failed AIMS reading and 30 percent flunked writing. Arizona's Instrument to Measure Standards is the state-mandated test required for high school graduation.
Kate Mahoney, an assistant professor at SUNY Fredonia, said the evidence suggests that some students leave the program too early and aren't ready for the regular classroom.
Mahoney said states should not use a single test to make such important educational decisions because no test is perfect and test scores are subject to error.
It's better, she said, to look at multiple factors such as grades and other test scores.
Garcia Dugan, Arizona's deputy schools chief, said it's incorrect to assume that all students who pass the Stanford test will pass AIMS. Stanford measures language proficiency while AIMS gauges academics.
State officials have not determined whether students who pass the Stanford test fail AIMS to a greater degree, but they expect to look into this.
Using a test to determine whether a student is an English-learner is a more equitable way of making that decision, she said, than if teachers were allowed to consider multiple factors such as grades and other test scores.
Test will change
The Arizona Education Department is changing the Stanford for the next
school year, although officials say the changes aren't in response to
criticisms or the civil rights investigation.
They instead are part of a directive from the U.S. Department of Education that the language test must be more in line with the language standards taught in Arizona classrooms. Right now, Arizona's test is 67 to 69 percent in line.
The publisher will test a new version in a few select districts and rename the test the Arizona English Language Learner Assessment. Other students will continue to take the Stanford for the rest of the school year.
Because the test will be revised, the State Board of Education will have to set new pass-fail scores for the latest version, perhaps as early as May. The state board approved the original Stanford test and the pass-fail scores in 2004. Whether the new scores will make the test easier or harder to pass has yet to be determined.
State board President JoAnne Hilde was concerned at first when she saw large numbers of students pass the test.
Hilde said it's too soon to tell whether a significant number of those students are falling behind.
The board should find out more later this spring. When that happens, if large numbers of students still need extra help, she said "we need to take that into consideration."
Reach the reporter at