How is the Writing Extended Response scored on California’s CAASPP ELA test?

What is the Writing Extended Response on California's ELA test?

California’s statewide testing system is the California Assessment of Student Performance and Progress, or CAASPP, established on January 1, 2014 in place of the STAR Program. Its English language arts test is the Smarter Balanced Summative Assessment for English language arts/literacy, taken in grades 3 through 8 and grade 11, computer adaptive with a performance task attached.

The performance task is where the writing happens. The California Department of Education’s explainer describes a task with two questions: first a multiple-choice or short-answer item assessing research skills, then the Writing Extended Response, which CDE also calls the full-write or essay question. Students read multiple sources and respond with one extended written response. Performance tasks are not computer adaptive and may take one to two test sessions.

Each student is randomly assigned one WER question in one of three purposes: narrative, opinion/argumentative, or informational/explanatory. Narrative appears only in grades 3 through 8, so a grade 11 student never sits a narrative full write. Grades 3 to 5 get informational and opinion. Grades 6 to 11 get explanatory and argument, with counterclaims required from grade 7 up. Each task carries two to five grade-appropriate source texts, though a narrative may be text-inspired rather than text-based. A student cannot choose the genre, and the purpose that lands on their screen is the purpose they are scored against.

Which traits does the California full-write rubric actually score?

Three traits, with asymmetric point ranges. On the argument, explanatory, informational and opinion rubrics the traits are Organization/Purpose at 1 to 4 points, Evidence/Elaboration at 1 to 4 points, and Conventions at 0 to 2. The narrative rubric used in grades 3 through 8 keeps Organization/Purpose and Conventions but replaces Evidence/Elaboration with Development/Elaboration. There is no voice or style trait, and the traits are not all worth four points.

Conventions is scored holistically against grade-level Common Core expectations on three named guidelines: Variety, Severity and Density. Severity is worth explaining to students, because basic errors are weighted more heavily than higher-level errors. Density is the ratio of errors to length. CDE draws the boundary plainly: grammar and spelling are assessed in Conventions, worth a maximum of two points, and are not taken into account in the other traits, worth a maximum of four points each.

The two four-point traits are then combined. The current scoring specification says evidence/elaboration and organization/purpose are averaged and the average rounded up. Any trait can also carry the non-score code NS, covering insufficient responses including copied text, responses in a language other than English, off-topic and off-purpose.

One rule deserves its own lesson. Since 2022-2023, if an essay is flagged off-purpose the Conventions trait is scored and evidence/elaboration and organization/purpose are not, regardless of the writing purpose assigned. A student who writes a fluent, well-evidenced narrative in response to an argument prompt walks away with a grammar score and nothing else.

Why does the grade 11 writing score follow students to college?

California is the state where the grade 11 ELA result doubles as a college placement instrument. Through the Early Assessment Program, the California State University system assigns each student an academic preparation status in mathematics and in written communication from that score.

Level 4, Standard Exceeded, means ready for college-level work without support. Level 3, Standard Met, means ready without support once the student completes a senior-year English course, and the approved options include AP or IB English, approved honors English, and the CSU Expository Reading and Writing Curriculum. Level 2, Standard Nearly Met, means demonstrating preparation another way through ACT, SAT, AP or IB results, or enrolling in supported coursework at CSU. Level 1 is Standard Not Met. Results are used for placement, not admission.

For an English department that raises the stakes on senior-year course selection as much as on the test itself.

What changed in California's score reporting this year?

The State Board of Education adopted new achievement level labels for CAASPP, launching with reporting for the 2025-26 summative assessment. Standard Exceeded becomes Advanced, Standard Met becomes Proficient, Standard Nearly Met becomes Developing, and Standard Not Met becomes Minimal. CDE’s stated reason is that the new labels make the results easier to understand. The criteria and cut scores for each level have not changed, so results can still be compared from year to year. The board approved the change on a split vote on March 5, 2025.

Watch the denominator when someone quotes a figure at you. The state’s October 2025 results announcement gave students meeting grade level expectations as those scoring at levels 2, 3 or 4, which came to 70.6 percent in ELA, with 29.4 percent at the minimal level. That is not the conventional Level 3 and above measure. Two people can quote California ELA results honestly and be far apart.

A second quirk hits departments directly. From the 2020-21 administration California adopted adjusted-form blueprints, which reduced the number of questions on each test. Individual area results are now reported only for groups of 30 or more, and composite area results for groups of 11 or more from 2023-24. For ELA the Writing and Research claims are combined into a single composite area, with achievement given as Above Standard, Near Standard or Below Standard. Teach one section of 28 and there is no clean writing-only signal in the report.

Who actually scores student writing in California?

More often than teachers expect, they do. On the interim assessments, educators hand score the constructed-response and performance task items themselves and enter the scores in the Teacher Hand Scoring System. CAASPP publishes an on-demand learning module specifically on hand scoring ELA performance tasks, worth finding before a department tries it cold.

The same pattern holds on California’s English learner test. The state runs its own instrument, the English Language Proficiency Assessments for California, which assesses four domains, Listening, Speaking, Reading and Writing, and is aligned to the 2012 California English Language Development Standards. EL students take it every year from kindergarten through grade twelve. Writing is locally scored: in grades K through 2 the answer books are scored locally and the scores keyed into the Data Entry Interface, and in grades 3 through 12 the test examiner enters scores in the Teacher Hand Scoring System.

One California rule also shapes who is in the writing cohort at all. A recently arrived English learner, one whose first date of entry into a US school as shown in CALPADS falls after April 15 of the year prior to the assessment, does not need to take the ELA assessment.

How can a department run practice full writes that match the state task?

Build the practice task the way the real one is built. Assign the purpose rather than letting students pick it. Supply two to five source texts. Put a short research question in front of the essay. Require counterclaims from grade 7. Then score it with the real weights, four, four and two, and keep grammar out of the two four-point traits, because scoring conventions at four teaches students that a clean sentence is worth as much as a well-chosen quotation.

Calibration is where the disagreement lives, and Evidence/Elaboration is usually the trait two teachers read differently. Before a department grading meeting, putting a small sample of scripts through expert essay feedback gives you an independent read to argue against, which beats arguing from memory of last year’s anchor papers.

Whatever you use for the second read, the rubric you load into it has to be the real one. Jeddle lets a department set its own criteria and comment banks, so the four-four-two split and the off-purpose rule survive into the feedback students see. Every JeddAI draft is reviewed and edited by the teacher before a student reads it, which is what keeps a grade defensible when a parent asks how it was reached.

Source selection quietly decides how hard the task is. Pulling passages from shared study guides for the subject keeps the reading load comparable across sections, so a low Evidence/Elaboration score tells you something about the writer rather than about the texts they were handed.