STAAR Rescoring Requests Triple After More Texas Schools Question Computer Grading

TL;DR

Age 18–24?Offer from Amazon

Prime made for students and young adults

• Fast, free delivery for dorm and study essentials
• Prime Video and Amazon Music included
• Member-only deals

Try Prime for Young Adults →Free trial for eligible 18–24 year olds

As an affiliate, we earn on qualifying purchases.Hide

Texas received 64,489 requests to rescore open-ended responses on reading STAAR exams this spring, nearly three times the prior year’s 21,620. The Texas Education Agency raised more than 13,300 scores after human reviews, though changes affected less than 0.5% of all open-ended reading responses.

Texas school districts and charter networks submitted 64,489 requests this spring to rescore students’ open-ended responses on reading STAAR exams, nearly triple the 21,620 requests last year, according to Texas Education Agency data reported by The 74. Human reviewers raised more than 13,300 reading scores, intensifying scrutiny of the state’s use of computer scoring while showing that score changes remained a small share of all responses.

The request total followed 2,758 appeals in spring 2024, the first year in the comparison reported by The 74. The number of corrections this spring doubled from the previous year, the report said. More than 250 districts and public charter networks, including Cypress-Fairbanks, Dallas, Katy and Northside, submitted requests. In nearly 90% of those systems, at least one student’s reading score increased after review.

The agency’s reviews changed more than 13,300 scores. Of the responses whose scores were raised, 58% had initially been graded by a computer and 42% by a human, according to the report. Human scorers handled every rescore. Across the more than 3.2 million students who took reading assessments this spring, less than 0.5% of open-ended responses had their scores changed.

Texas automatically rescores extended responses when a student is one point short of reaching the next performance level, at no charge. Those levels are did not meet, approaches, meets and masters. Some districts said the automatic reviews let them use their own requests for students farther from the next threshold; one district told The 74 it submitted responses that were two points away.

Score Changes Can Affect Students and Schools

STAAR results can affect students’ next steps as well as school accountability. A revised score may move a student into a different performance category, which, according to accountability director Allison Matney, can affect whether a student must complete remediation or can meet a graduation requirement. The exact consequences depend on the student’s circumstances and the assessment.

Test outcomes also inform campus A–F ratings. Under Texas law, the education commissioner can take action when a campus underperforms for multiple years, including appointing a conservator or replacing a locally elected school board. That makes accurate scoring consequential for districts, although the reported correction rate was below 0.5% of open-ended reading responses overall.

The sharp rise in requests also points to a growing review workload for school systems and raises questions about confidence in the original scoring. University of Texas at Austin professor Rachel White told The 74 that educators’ increasing disagreement with scores suggests a trust problem. That is her interpretation of the request trend, not a finding that the computer scores were systematically wrong.

How Texas Uses Automated Scoring

The Texas Education Agency began using an automated scoring engine in December 2023 for open-ended STAAR responses, including reading and English assessments. The shift followed a redesign that reduced multiple-choice questions and added more written responses. To build the scoring system, the agency used thousands of responses previously scored by people to train a program to assign scores based on patterns in those examples.

The agency says the tool is not generative AI like ChatGPT or Google Gemini: it does not learn from one student response to the next. TEA officials have also described the arrangement as a hybrid system, with human experts involved in scoring and human review of a portion of results. The 74 reported that humans manually reviewed a quarter of exam scores.

The system drew criticism after an early increase in students receiving zeros on writing prompts. Agency officials pointed to other possible factors in those results, including students retaking the exam. The current rescoring figures do not, on their own, establish why districts requested reviews or whether automated scoring was more error-prone than human scoring.

“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.”

— Allison Matney, director of the Texas Center for School Accountability

What the Request Surge Cannot Show

The published totals do not explain why each district sought a rescore, how many requests involved each scoring method, or how the 13,300-plus score increases were distributed among students and performance levels. The increase in requests alone cannot determine whether computer scoring produced more errors, whether districts became more willing to appeal, or whether both factors played a part.

The reported account also does not provide a full breakdown of the cost and staff time districts spent reviewing responses and submitting appeals. TEA’s explanation of its scoring model and the concerns raised by educators are competing perspectives; the data supplied do not resolve the broader question of how closely machine and human scores agree across all responses.

Districts Await Future STAAR Reviews

Districts and families can continue using the state’s rescore process for eligible responses, alongside the automatic review for students one point below the next performance level. Any later assessment cycle will show whether requests remain near this spring’s level and whether the number of score changes follows a similar pattern.

The reported material does not identify a new policy change or a scheduled state review of the scoring system. For now, the next developments to watch are future TEA assessment results and any additional detail the agency or districts provide on appeals, score corrections and the share of responses originally graded by computers.

Key Questions

How many STAAR rescoring requests did Texas receive this spring?

The state received 64,489 requests to rescore open-ended responses on reading exams, up from 21,620 the previous year, according to TEA data reported by The 74.

How many scores increased after human review?

More than 13,300 reading scores were raised after review. Human scorers handled all the rescoring, and the changes affected less than 0.5% of open-ended responses on reading exams.

Were the responses originally graded by computers?

Some were. Among responses whose scores were raised, 58% had initially been computer-graded and 42% had been graded by people, according to the report.

Does a higher STAAR score change a student’s performance level?

It can, depending on the student’s original score and the number of points changed. Texas has four performance levels: did not meet, approaches, meets and masters. A change near a threshold may affect a student’s category and related requirements.

Does the increase prove that computer scoring is inaccurate?

No. The request total shows that more districts sought reviews, and the review process raised some scores. It does not establish the overall accuracy of computer scoring or explain why each district filed an appeal.

Source: rss