Reading comprehension IEP goals: decode first, then count something real
Comprehension goals fail for a reason the other academic areas don’t: comprehension is invisible. You can watch a student compute or hear them decode, but understanding has to be inferred from something they produce — so the goal lives or dies on what you choose to count. Federal law also treats reading as three separate deficit areas — basic reading skill, reading fluency skills, and reading comprehension (34 CFR §300.309(a)(1)(iv)–(vi)) — and writing a comprehension goal for what is actually a decoding deficit is the single most common mistake in this domain. This guide covers that decision, the metrics that survive an audit, and labeled samples. The goal anatomy itself lives in our core goal-writing guide — this page applies it to comprehension.
First decision: is this actually a comprehension deficit?
Run the differential before writing the goal. If the student stumbles on the words — misreads, skips, labors — the bottleneck is decoding or fluency, and the right goal lives in that area (our dyslexia and reading fluency goal guides cover those sides). The cleanest check is listening comprehension: give the same passage and question set read aloud. A student who understands aurally but not in print has a print problem; a student who struggles both ways has a language-comprehension need that a fluency goal will never touch. Students can need goals in both areas — that’s two precise goals, not one blended “will improve reading” goal. Whichever you pick, the PLAAFP must show the evidence in the same metric the goal will use — that’s the job of baseline data.
Metrics that hold up
| Skill | Metric | Why it works |
|---|---|---|
| General comprehension (weekly data) | Maze items correct in 3 minutes on leveled passages | Objective scoring, 3-minute administration, sensitive to weekly growth — the standard comprehension CBM |
| Question answering | Questions correct by type (literal / inferential), out of a fixed set | Naming the mix stops the goal from being met on literal items while inference never moves |
| Retell / summary | Story-grammar elements included (character, setting, problem, events, resolution) out of 5 | Turns 'retells the story' into a count; same rubric works for any narrative passage |
| Main idea (expository text) | Main idea stated + N supporting details identified, per passage, out of a set | Content-area comprehension is a different skill than narrative — measure it separately |
| Vocabulary in context | Unknown-word meanings correctly inferred from context, out of a fixed set | Vocabulary is a top predictor of comprehension; countable when the word set is defined |
| Listening comprehension (differential) | Same question set on passages read aloud vs read independently | Separates a language-comprehension deficit from a decoding bottleneck — informs which goal to write |
The through-line: every metric defines the passage (level, type, read how), the response, and the count. The regulation requires the goal to be measurable and the IEP to say how progress will be measured and when reports go home (34 CFR §300.320(a)(2)–(3)) — and whatever method you name is what your progress-monitoring system has to produce every reporting period. Maze probes exist precisely because “discussion-based checks” can’t be graphed.
Sample comprehension goals (clearly labeled samples)
These are invented, realistic examples for structure-stealing — swap in your student’s baseline, passage levels, and dates. Each has the full anatomy: condition, behavior, criterion, measurement, timeframe, and a baseline in the same metric.
SAMPLE — Maze / weekly CBM (Grade 4)
Given a 3-minute maze passage at the mid-3rd-grade level, Jaylen will select 18 correct replacements, improving from a baseline median of 8, on 3 consecutive weekly probes by April 2027.
SAMPLE — Literal + inferential questions (Grade 5)
After independently reading an instructional-level passage, Sofia will answer 8 of 10 comprehension questions (at least 4 of 5 inferential) across 3 consecutive biweekly probes, from a baseline of 4 of 10 (1 of 5 inferential), by May 2027.
SAMPLE — Retell with story grammar (Grade 2)
After reading an instructional-level narrative passage, Marcus will retell the story including 4 of 5 story-grammar elements (character, setting, problem, key events, resolution) on 4 of 5 weekly opportunities, from a baseline of 1 of 5 elements, by March 2027.
SAMPLE — Expository main idea (Grade 7)
Given a grade-appropriate science or social studies passage and a main-idea organizer, Amara will state the main idea and list 3 supporting details in 4 of 5 passages across 3 consecutive biweekly checks, from a baseline of 1 of 5, by January 2027.
SAMPLE — Inference with text evidence (Grade 9)
After reading an instructional-level passage, Diego will answer 4 of 5 inferential questions and cite the supporting text evidence for each answer on 3 consecutive biweekly probes, from a baseline of 1 of 5 with evidence, by June 2027.
The four ways comprehension goals fail review
- The wrong-deficit goal. A comprehension goal for a student whose probes show a decoding gap. The instruction that follows the goal won’t treat the actual need — and the eligibility areas say these are different deficits.
- The unanchored question set. “80% of comprehension questions” without passage level, question type, or set size. Whose questions? On what text? Met on literal items while inference never moves.
- The grade-level target. “Will comprehend grade-level text” — a moving target with no measurement method. Fix the passage level in the condition; describe the gap in the PLAAFP.
- The unteachable verb. “Will read for meaning,” “will demonstrate understanding” — not observable behaviors. Retell, answer, state, cite, select: verbs a second adult can score. Cross-section consistency between the PLAAFP, the goal, and the service grid is what a full compliance check verifies document-wide.
This page completes the academic trio with written expression and math — the same name-the-deficit-first discipline applies across all three.
FAQ
When should an IEP have a reading comprehension goal instead of a decoding goal?
Only when the data shows the student decodes adequately and still doesn't understand. Federal SLD eligibility treats basic reading skill, reading fluency skills, and reading comprehension as three separate areas (34 CFR §300.309(a)(1)(iv)–(vi)). If the student misreads or labors through the words, the deficit is decoding or fluency — a comprehension goal on top of a decoding deficit measures the wrong thing and the instruction won't transfer. Listening comprehension is the classic differential: if the student understands the same passage read aloud, the problem is getting through the print, not understanding language.
What makes a reading comprehension IEP goal measurable?
The same anatomy as every goal: a condition (passage level and type, read how), an observable behavior, a numeric criterion, a named measurement method, and a timeframe (34 CFR §300.320(a)(2)–(3)). For comprehension the behavior has to be countable — questions answered by type, story-grammar elements retold, main idea stated with supporting details, maze items correct — never 'will improve comprehension' or 'will read for meaning.'
What is a maze probe and why use it for comprehension goals?
A maze is a timed passage where every seventh word (roughly) is replaced with a three-choice item; the student circles the word that belongs. It's the standard curriculum-based measure for reading comprehension: quick to give, scored objectively by any adult, and sensitive enough to show growth week to week. 'Literal and inferential questions' measure deeper understanding but depend on who wrote the questions — many teams use maze for the weekly data and question sets or retells for the richer picture.
Should the goal distinguish literal from inferential comprehension?
Yes, if the data does. Answering 'what happened' (literal) and 'why did the character do it' (inferential) are different skills, and many students pass one while failing the other. A goal that counts '8 of 10 comprehension questions' without naming the mix can be met entirely on literal items while the actual deficit — inference — never moves. Name the split: e.g., 4 literal and 4 inferential of 10.
What's a reasonable criterion for a comprehension goal?
Anchor it to the baseline in the same metric, then set a year's ambitious-but-realistic growth. If the student currently answers 3 of 10 inferential questions on instructional-level text, 8 of 10 with the same passage type and question mix is a defensible annual target. Avoid 'grade level' as the criterion — it moves all year and isn't a measurement method; fix the passage level in the condition and let the PLAAFP describe the gap.