Arizona lowers English fluency test score, removing 30,000 from language services

Earlier AZELLA version drew federal civil-rights settlement

Outgoing Arizona Superintendent of Public Instruction Tom Horne approved new passing-score thresholds for the state’s English fluency test this spring, after the Arizona Department of Education administered the exam to 600 native English speakers to establish a baseline. ProPublica and Arizona Luminaria, which co-published the reporting, describe Horne as a longtime Republican defender of Arizona’s English-only policies. The change lowered the minimum combined score students need to pass from 1,000 out of a possible 1,405 to between 944 and 962, depending on grade level.

Nearly 113,000 students took the 2026 Arizona English Language Learner Assessment, or AZELLA, and about 30,000 were reclassified as proficient and exited English-language services this school year — raising the statewide reclassification rate to 26% from 12% the prior year. The 12% rate had held steady for three school years. In 2025, more than 118,000 students took the test and just over 14,000 scored high enough to be reclassified as proficient. In the Tucson Unified School District, the percentage of test-taking students who scored high enough to be classified as proficient in English more than doubled, from 10% in 2025 to 24% in 2026.

Arizona is the only state with English-only laws still in effect for its public schools, a status it has held since voters approved Proposition 203 in 2000. Under the resulting structured English immersion program, colloquially called “English-only” because all instruction is required to be in English, students who speak a language other than English at home — nearly 1 in 10 enrolled in Arizona public and charter schools — are placed in daily blocks of language instruction ranging from two to four hours. The students, most of them Spanish-speaking Latino children, remain in the program, separated for part of that time from their English-primary peers, until they pass the AZELLA, which assesses reading, writing, speaking and listening. The English-only instruction limits their time for core subjects, like math and science, that are critical to academic success and college readiness.

Deputy Superintendent Margaret Dugan, who chaired the group that advocated for passage of Proposition 203 and has served with Horne during all three of his terms, said each new version of the test is given to native speakers first and the results are used as a baseline. The department skipped that step in 2020, the last time the test underwent a major change, following a State Board of Education decision granting more flexibility in how schools teach students English. “Because of COVID and all of that that went on in 2020, things were not done in order,” Dugan said. “When we walked in, it was a new test and they had never tested the native speakers, so now they had to.”

Adela Santa Cruz, who leads the Office of English Language Acquisition, which oversees the test, said when the prior version of the AZELLA was belatedly administered to native English speakers, nearly half failed. “With a native not being able to pass that test at a high percentage, there’s something that was just quite awry,” she said. Santa Cruz said the issues were reflected in the pass rate of English learners plateauing at 12% for three years.

Anju Kuriakose, the department’s director of psychometrics, told school district representatives in April that half of the native speakers who had been tested to inform the changes were in kindergarten and first grade — and that students in fifth grade or higher accounted for less than one-fifth of test takers. “My hope is that all of you will see an increase in proficiency,” Kuriakose said upon announcing the changes.

Ida Rose Florez, a psychologist and former professor and researcher at Arizona State University whose 2012 study found the AZELLA “ineffective and obsolete” and its results could be arbitrary, said the 2026 changes repeat problems her earlier work documented. She noted that basing scoring guidelines too heavily on the results of native-speaking kindergartners — without enough data from older children with more advanced language skills — can artificially elevate pass rates. “What I examined, and what I saw, was a test that was absolutely not developed correctly,” Florez said. “I don’t see any basis for validity whatsoever.”

Katie Bernstein, an applied linguist and professor of early childhood education at Arizona State University, said students placed in classrooms where they cannot follow instruction are denied meaningful learning. “If you’re put in a class where you don’t understand what’s going on at all, you’re not learning language or content,” she said. “And so those are the worst outcomes of all, are students who are not given any language services.”

The change echoes findings from a federal investigation more than a decade ago. The U.S. departments of Justice and Education determined that the version of AZELLA administered between 2006 and 2012 — coinciding with Horne’s first two terms — was “not a valid measure of English language proficiency and readiness.” As a result, the state had removed at least 28,000 students from structured English immersion who still lacked proficiency in reading and writing. The Arizona Department of Education was found to be in violation of civil-rights laws barring discrimination on the basis of race and national origin by schools receiving federal funds. The state entered a voluntary settlement requiring intervention services for affected students and a long-term revamp of the test. A second voluntary settlement four years later required the state to raise the threshold for a passing score.

Since Proposition 203 was debated 26 years ago, proponents have argued it helps students learn English more quickly and gets them into mainstream classes faster, though there is limited research to support this. Ample studies have found that students taught under English-only policies have larger achievement gaps than those instructed using bilingual or multilingual education, which reinforces the student’s primary language while accelerating their learning in academic subjects and a second language. A landmark 32-year study spanning 16 states analyzed various language-learning models, including structured English immersion, and found that “the key to accelerated progress is for English learners to receive peer-equivalent, grade-level bilingual schooling,” according to researchers Virginia Collier and Wayne Thomas.

“The fundamental problem with the AZELLA is that it’s trying to measure a system that should never be in place to begin with,” Florez said. Bernstein added: “There is no linguistic, educational, economic, or research-based reason why we shouldn’t have dual-language programs in the state. And so I think that potentially leaves politics as the only answer to why.”

English learners in Arizona already lag far behind their peers on standardized tests. In 2025, third-grade reading tests showed a 31-percentage-point gap between English-language learners and other students, according to the Center for the Future of Arizona’s Education Progress Meter; the eighth-grade math exam showed a 23-percentage-point gap.

In a 2024 interview with Arizona Luminaria, Horne said students’ reclassification rates were highest during his earlier tenure, claiming he had driven the rate “up to 31% in three or four years.” The Arizona auditor general questioned the validity of those results. An audit released in 2011 found that “because data is either unavailable or unreliable, the effect of SEI models is unknown.” Horne lost his bid for reelection in July’s Republican primary and will leave office in December. The Arizona Department of Education said Horne was unavailable for an interview.

At the Tucson Unified School District, Patricia Sandoval-Taylor, the district’s director for language development, said state law requires teachers to track reclassified students’ progress for two years and intervene with tutoring or summer school if needed — but Arizona schools do not receive additional funding for those services. Because part of the district’s funding for English learners is tied to enrollment in structured English immersion, more students passing the 2026 AZELLA will mean less district funding in the 2027 school year. “They’re not going to have the direct services to ensure they become fully English proficient,” Sandoval-Taylor said. The district is preparing to continue supporting students like Arturo within the mandates approved by voters in 2000.

For some families, the change carries concrete consequences. Juana Casas said her fourth-grade son Arturo, who passed the AZELLA this spring after three years in structured English immersion at John B. Wright Elementary School, received after-school tutoring and help from his two older sisters at home. “I couldn’t help my oldest daughter, and she really struggled,” Casas said in Spanish. “Their two experiences are totally different. Arturo doesn’t speak Spanish very well because they’ve always spoken to him just in English so that he wouldn’t have to struggle like my oldest struggled.”

At Wright Elementary, where Samantha Ramos has taught English as a second language for 15 years, she said she continues to celebrate when her students pass the AZELLA. “I think they need to be celebrated because they’re going above and beyond,” she said. “Before they did the change to the cut scores, it was like, ‘Well they passed this really incredibly difficult test that even English speakers couldn’t even pass themselves.’”