STAAR computer grading rescoring requests explained for families and schools

STAAR computer grading rescoring requests explained for families and schools

STAAR computer grading rescoring requests nearly tripled in Texas this year, rising from 21,620 in 2025 to 64,489 in 2026. The reviews involve written responses on reading exams, including responses that were initially scored by computers. More than 250 Texas school districts and charter networks requested reviews, and nearly 90% had at least one score increase, according to the Texas Tribune.

The increase affects schools deciding whether a score should be reviewed and families trying to understand what a STAAR result means. It also brings attention to the Texas STAAR automated scoring system, which the Texas Education Agency began using in December 2023 for open-ended responses. The available research explains how the system works and what happened during the 2026 review cycle, but it does not establish a single statewide procedure for families to request a current-year rescore.

Why are STAAR rescoring requests increasing?

The number of requests has risen each year. Texas schools made 2,758 requests in spring 2024, followed by 21,620 last year and 64,489 this spring, the Texas Tribune reported last week.

Allison Matney, director of the Texas Center for School Accountability, said district administrators historically requested rescoring only rarely. She attributed the increase to greater awareness of the automated scoring and review process.

“That traction of superintendents talking to each other, and media continuing to talk about it, made more districts aware that not only can they do this process, but actually they should, and they need to,” Matney said, according to the Texas Tribune.

The rising number of requests does not, by itself, show that every computer score was wrong or that the statewide system has failed. It does show that more districts are examining scores and using the review process when local educators believe a response deserves another look.

Rachel White, an associate educational leadership and policy professor at the University of Texas at Austin, offered a more critical interpretation. She said the tripling of requests suggests educators who read students’ essays are increasingly disagreeing with assigned scores, whether those scores came from a person or a machine.

“That to me just indicates that there is a bit of a trust problem in how these are being scored,” White said, as reported by the Texas Tribune.

How STAAR computer grading and rescoring requests work

The TEA’s automated scoring engine uses natural language processing, a form of artificial intelligence technology, to evaluate open-ended questions on reading, writing, science, and social studies assessments. The system is not described as an autonomous evaluator or chatbot. It was developed using 3,000 responses that had already received two rounds of human scoring, allowing the computer to identify response characteristics associated with those scores, the Texas Tribune reported in 2024.

The original process included several human checks. Computers scored constructed responses first, while a portion of responses was rescored by people. Responses received additional human attention when the computer had low confidence or encountered a type of answer its programming did not recognize, such as one containing substantial slang or words in another language, according to the 2024 report.

For the 2026 scores that were reviewed, 42% had initially been assigned by humans and 58% by computers. Those percentages apply to the reviewed group, not to every STAAR response taken in Texas. Every rescore in that group was conducted by a human, the Texas Tribune reported last week.

TEA spokesperson Jake Kobersky defended the hybrid approach. He said grading every response by hand would require four to five times as many human scorers, add between $15 million and $20 million in annual costs, and delay results for districts, according to the Texas Tribune.

“TEA uses a hybrid model where human experts drive the scoring,” Kobersky said. He described the model as a way to keep human judgment involved while returning results faster than hand-scoring every response would allow.

That distinction matters when families hear about “AI grading of STAAR written responses.” A computer may assign the initial score, but the review process described in the reporting sends requested responses to human scorers.

What changed after STAAR reading test score reviews?

Statewide, fewer than 0.5% of open-ended reading responses changed after review. More than 3.2 million Texas students took the reading STAAR exams this spring, including students in grades 3 through 8 and students taking the English I and English II end-of-course assessments, the Texas Tribune reported last week.

That statewide percentage does not describe every district’s experience. Dallas ISD submitted more than 5,100 reading tests for review, and 34% returned with a higher score, according to the Texas Tribune.

Dallas officials said they requested the reviews because they believed students could benefit from human scorers examining their answers. The district estimated that automatic and district-requested reviews together helped five campuses improve their state accountability ratings before the public release. One campus moved from D to C, while four moved from C to B, the Texas Tribune reported.

Public campuses and districts receive A–F accountability ratings based partly on how many students reach specified STAAR performance levels. The review results therefore can matter beyond an individual score, particularly when a small score change affects a campus rating.

The stakes can also be personal. Matney said a difference of a few points could move a student from failing to passing. For an individual high school student, she said, that change could affect whether the student can graduate or must complete remediation hours, according to the Texas Tribune.

Who can request a review, and what is confirmed?

The reported 2026 process confirms that school districts could request reviews. Districts paid $50 per test when the score did not change, while the fee was waived when the score increased, the Texas Tribune reported last week.

TEA also automatically rescored extended constructed responses at no cost when the result was one point below the next performance level. The four levels are did not meet, approaches, meets, and masters, according to the same report.

The research does not confirm whether a parent or student could independently request a 2026 review, whether the district fee policy applies to families this year, the current deadline, or the expected turnaround time. A 2024 report said students and parents could request a rescore for $50, with the fee waived if the new score was higher, but that earlier description should not be treated as current 2026 guidance without confirmation from TEA or the district, the Texas Tribune reported in 2024.

District practices also differ. North East ISD spokesperson Aubrey Chancellor said the district submitted no requests after reviewing multiple students’ scores and deciding that another review would not benefit scores or campus ratings, according to the Texas Tribune.

For students in grades 3 through 8, the available reporting says a low score does not affect a student’s grades or academic progress. For high school students, meeting a minimum STAAR score is one common way to fulfill a state graduation requirement, but it is not the only way, the Texas Tribune reported in 2024. Families should confirm how a particular score affects a student with the campus or district because the research does not provide individualized graduation or remediation determinations.

What families and educators should do next

A parent or student with a question about a STAAR score should first contact the campus administrator or district testing coordinator. Ask whether the district reviewed the response, whether a district-requested rescore is available for that assessment and response type, who would submit it, what fee applies, and what deadline controls the request.

Also ask what happens if a score changes. The available reporting confirms that the 2026 rescoring process occurred before the release of A–F accountability ratings in August, but it does not establish a general deadline for every future testing cycle or confirm how every higher score will update an individual student record.

The Texas Assessment Research Portal can help readers examine aggregate STAAR data by state, region, district, or campus. It does not necessarily provide a student’s individual response or rescore record, TEA explains. For an individual student’s result, the appropriate source is the school or district testing office.

The 2026 numbers show a sharp increase in review activity, while the statewide share of open-ended reading responses that changed remained below 0.5%. Those facts leave room for different interpretations, but they give families and educators a practical starting point: verify the student’s result locally, ask which current review rules apply, and get the answer from the district before paying a fee or assuming that a score can be changed.