A machine scores the longer written answers on English STAAR first
The Texas Education Agency scores the constructed responses on English language STAAR tests with an automated scoring engine first. At least 25 percent of those answers then go to people. Where a person scores an answer, that score is the one that counts. A March 2024 agency presentation puts the engine's score as the score of record on about 75 percent of responses. Constructed responses on Spanish language tests are scored by people alone. The agency began the hybrid model in December 2023, and its presentation says full human scoring would have cost $15 million to $20 million more per year. On September 4th, 2026 the agency said about 27,200 students received additional credit after content experts rescored short typed answers. It said none of the regraded responses had been graded by the engine. A district can ask for a constructed response to be rescored by a person for a fee, and the fee is waived if the score changes.
How to take part
A parent or guardian who has concerns about a score on a constructed response can raise them with the district's testing staff. The district files the rescore request in the state's testing system within a set window. A rescore is done by a human scorer. The request carries a fee, and the fee is waived when the score changes.
Where
Statewide. This decision applies across Texas rather than to a named county.
Timeline
- decided
The agency said about 27,200 students received additional credit after content experts rescored text responses
- Today
How this decision moved
One dated line per check, oldest first. A line that says nothing changed means somebody looked and it had not.
- 2026-09-30
Admitted on the agency's own scoring documents and the agency's account of its September rescore.
The evidence
Every fact above rests on one of these. The words are the source's own.
This means that the responses are initially scored by an automated scoring engine (ASE) and at least 25 percent of student responses are then routed to human scorers.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
Student responses to SCR and ECR questions for STAAR Spanish assessments are 100 percent human scored and do not get scored by the ASE.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
Any student responses that are routed for human scoring maintain the score assigned by the human scorer as the score of record.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
The ASE is trained on student responses and human scores from the field-test data.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
Rescore requests are sent for human scoring.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
If the score on the constructed-response question changes, the fee is waived.Texas Education Agency, Scoring Process for STAAR Constructed Responses, December 2023 Primary source, official · tea.texas.gov
Scenario 1: Auto scoring engine assigns score of record Approx. 75% responsesTexas Education Agency, Hybrid Scoring Key Questions, March 2024 Primary source, official · tea.texas.gov
maintaining full human scoring would have cost $15-20M more per yearTexas Education Agency, Hybrid Scoring Key Questions, March 2024 Primary source, official · tea.texas.gov
The engine uses a sample of ~3,000 human scored responses from the field test for programming.Texas Education Agency, Hybrid Scoring Key Questions, March 2024 Primary source, official · tea.texas.gov
The hybrid scoring transition started in December 2023, and student responses went from fully human scoring to hybrid scoring.Texas Education Agency, Hybrid Scoring Key Questions, March 2024 Primary source, official · tea.texas.gov
Texas Education Agency officials said Friday about 27,200 students received additional credit for their text responses to some questions on the reading portion of the State of Texas Assessments of Academic Readiness.The Texas Tribune, After rescoring STAAR exams, 11 Texas districts seeing A-F grades improve, September 4th, 2026 Journalism · texastribune.org
However, Kobersky said none of the regraded responses were graded by its “automated scoring engine.”The Texas Tribune, After rescoring STAAR exams, 11 Texas districts seeing A-F grades improve, September 4th, 2026 Journalism · texastribune.org
Content experts rescored “text entry” responses, where students type brief text — such as a number, word, or phrase.The Texas Tribune, After rescoring STAAR exams, 11 Texas districts seeing A-F grades improve, September 4th, 2026 Journalism · texastribune.org
For example: If the correct answer to a question was ‘first,’ content experts may decide to award credit for a student that entered the response ‘ferst.’The Texas Tribune, After rescoring STAAR exams, 11 Texas districts seeing A-F grades improve, September 4th, 2026 Journalism · texastribune.org
No students’ scores will be lowered as a result of the review.The Texas Tribune, After rescoring STAAR exams, 11 Texas districts seeing A-F grades improve, September 4th, 2026 Journalism · texastribune.org
Questions about this decision
Answered from the record itself. Every answer is assembled from stored fields, so an answer the record has no basis for is left out rather than guessed.
What is this decision?
The Texas Education Agency scores the constructed responses on English language STAAR tests with an automated scoring engine first. At least 25 percent of those answers then go to people. Where a person scores an answer, that score is the one that counts. A March 2024 agency presentation puts the engine's score as the score of record on about 75 percent of responses. Constructed responses on Spanish language tests are scored by people alone. The agency began the hybrid model in December 2023, and its presentation says full human scoring would have cost $15 million to $20 million more per year. On September 4th, 2026 the agency said about 27,200 students received additional credit after content experts rescored short typed answers. It said none of the regraded responses had been graded by the engine. A district can ask for a constructed response to be rescored by a person for a fee, and the fee is waived if the score changes.
Who decides it?
Texas Education Agency decides. The record names the deciding body for every entry it carries.
Can the public take part?
A parent or guardian who has concerns about a score on a constructed response can raise them with the district's testing staff. The district files the rescore request in the state's testing system within a set window. A rescore is done by a human scorer. The request carries a fee, and the fee is waived when the score changes. No dated public window is on the record. The deciding body is named and reachable.
Where in Texas does it apply?
It applies statewide.
Has it been decided?
It has been decided. The dates on the item page carry when.
What happens next?
No future date is on the record. The last dated step on it was decided on September 4th.
When did it start?
The earliest date on its record is September 4th, 2026.
What kind of decision is it?
It is filed under health and education.
What sources back it?
Three sources back it. Two of them are primary.
Is it on the ERCOT grid?
No. It sits outside the ERCOT interconnection.
When was it last checked?
Every fact on it was last verified against its source on September 30th, 2026.
Cite this
Texas AI Docket, A machine scores the longer written answers on English STAAR first. Tracked since September 4th, 2026. Last verified September 30th, 2026. https://texasaidocket.com/item/tx-2026-0194/. Reuse permitted under CC BY 4.0 with attribution. The same entry is in the docket JSON as item tx-2026-0194.
Beat
Filed under Health and education, with every other decision on that beat.