IREAD-3 & ILEARN · an analytical briefing

What the tests do, and why — before the 2026 numbers land

Indiana's two headline K–8 assessments measure different constructs, are built differently, and carry different stakes. This briefing assembles what is documented about each instrument, the trend data through 2025, and the design changes that bear on how the 2026 release should be compared with prior years.

IREAD-3 2025
87.3%
statewide pass rate · +4.8 pts, largest one-year gain on record
ILEARN ELA 2025
40.6%
grades 3–8 proficient · +0.1 pt, flat five years running
ILEARN math 2025
42.1%
grades 3–8 proficient · +1.4 pts
Both subjects
31.2%
proficient in ELA and math · IDOE no longer headlines this figure
IREAD-3 2026
pending
expected Aug 11, 2026 SBOE meeting
ILEARN ELA 2026
pending
new cut scores required first
ILEARN math 2026
pending
new cut scores required first
2026 retentions
not today
retention counts need the Oct. 1 count day — expect November

IREAD-3 statewide pass rate, 2013–2025

Percent of third graders counted as passing. The reported rate is cumulative — it includes summer retakes and, since 2025, students who passed as second graders. 2020 was cancelled.

The 2025 figure in context

The 2025 rate of 87.3% equals the 2019 pre-pandemic rate and remains 4.1 points below the 2013 peak of 91.4%. It represents a return to 2019 levels, measured under a testing regime that changed in four documented ways during the same window. Those changes are set out under Building the pass rate.

ILEARN statewide proficiency, grades 3–8

Percent At or Above Proficiency. 2020 cancelled; IDOE treats 2021 as the working baseline. Math has climbed steadily since 2021; ELA has not moved.

Comparing the two reading measures

In the same year, on the same children, Indiana reported 87.3% passing IREAD-3 and roughly 40% proficient on ILEARN grade 3 English/language arts. Both are comprehension-weighted reading tests — the state's own alignment study codes 76.9% of IREAD items as comprehension. A 47-point difference between two comprehension-weighted tests of the same population turns primarily on where each cut score was placed rather than on what each test covers. We did not locate a state document addressing the comparison.

Five things worth knowing before the numbers drop

1. "Through-year" ILEARN doesn't work the way it sounds

The three new Checkpoints are formative only. IDOE's own Test Administrator's Manual says they "do not contribute to accountability calculations," and districts tell families they "do not count towards your child's final ILEARN score." The spring summative alone still produces the accountability number. Indiana added up to six state testing sittings per student per year and removed one math performance task.

2. The break in comparability was documented in advance

A 2024 IDOE blueprint document states: "In order to support comparability of test scores across years and avoid multiple years of score delays, IDOE will not engage in standard setting activities until the new through-year design is operational." That deferral has now come due, which places the break in comparability in 2026.

3. The IREAD pass rate is a composite, not a first-attempt rate

IDOE's published methodology counts, in the numerator, students who pass in spring or summer plus second graders who passed the prior spring. Only 65.4% of the students who actually sat the spring 2025 grade-3 test passed it, against a headline of 87.3%.

4. The 446 cut score has no public standard-setting record

Across four posted IREAD technical reports, the year, method, and panel that produced the grade-3 passing score of 446 could not be located. The reports defer to volumes that are not posted. The number that decides whether an eight-year-old repeats a grade is not publicly auditable.

5. Everything changes at once in 2026

A redesigned ILEARN, standards cut ~25%, freshly set cut scores approved the same day results are announced, and the first A–F school grades in five years. Isolating how much of any 2026 movement reflects school performance will be difficult.

How to use this briefing

Evidence on Indiana's literacy strategy points in both directions. The Harvard/Stanford Education Recovery Scorecard ranks Indiana 6th of 35 states in reading recovery; at the same time, several figures the public record would need to evaluate the results have not been published. Both sides of each question are set out in The evidence, with sources, so readers can weigh them directly.

The instruments

What each test actually is: who takes it, how it is built, what it measures, where the line sits, and what happens to a student who falls below it.

IREAD-3
Indiana Reading Evaluation And Determination · first administered spring 2012
Statutory origin
HEA 1367 (2010) / Public Law 109. Retention rule promulgated by the State Board in 2011, effective 2012–13. Current code: IC 20-32-8.5.
Who takes it
All grade 3 students. All grade 2 students since spring 2025 (phased in from 385 schools in 2022 → universal).
Format
Online, adaptive since spring 2024 (fixed-form before that). Paper fixed forms remain for accommodations.
Items
~38 operational items, multiple choice only (MC and multi-part MC). Three segments: discrete phonics/vocabulary, then two passage-linked comprehension segments.
Reporting categories
Reading: Foundations & Vocabulary (25–35%) · Reading: Nonfiction (30–40%) · Reading: Literature (30–40%)
Time
Currently described as untimed; ~90 minutes of working time across at least two days.
Scale & cut
200–650. Pass = 446. Grade 2 adds an "On Track" band at 405–445.
Attempts
Five before grade 4: grade 2 spring + grade 2 summer, grade 3 spring + two grade 3 summer retakes (second summer retake added by HEA 1499, 2025).
Stakes
Mandatory grade 3 retention since 2024–25 (SEA 1, 2024), subject to five statutory good-cause exemptions. Schools under 70% must employ an endorsed instructional coach and choose from an approved reading curriculum list.
Read-aloud
Passages are never read aloud — only directions. Defensible for a reading test; consequential for students with dyslexia who receive that accommodation on every other assessment.
Vendor
Cambium Assessment (formerly AIR). Cambium's own technical report states it "inherited the IREAD-3 item bank … and did not perform any new item development" until 24 new items were field-tested in 2024–25.
ILEARN
Indiana Learning Evaluation and Assessment Readiness Network · first administered spring 2019
Statutory origin
HEA 1003 (2017), replacing ISTEP+. Current code: IC 20-32-5.1.
Who takes it
Grades 3–8 ELA and math; grades 4 & 6 science; grade 5 social studies; Biology ECA. High school moved to the SAT (grade 11) after 2021.
Format
Computer-adaptive (social studies and U.S. Government are fixed-form). Selected response, technology-enhanced, constructed response, performance tasks.
2025–26 redesign
Three Checkpoints (20–25 items, 25–60 min, adaptive) in fall, winter, spring, plus a shortened spring summative. Authorized by HEA 1243 (2024); piloted 2024–25 in 1,350+ schools.
Do Checkpoints count?
No. Per IDOE's Test Administrator's Manual, Checkpoints "do not contribute to accountability calculations." The spring summative alone determines proficiency and accountability.
What got shorter
Mathematics dropped its performance task beginning 2025–26. ELA kept one CAT segment plus a one-day performance task.
Performance levels
Below Proficiency → Approaching Proficiency → At Proficiency → Above Proficiency. "Proficient" = the top two.
Cut scores
Set July 2019 via the Bookmark procedure with 100+ educators across nine grade-level committees plus a policy committee; approved by SBOE July 25, 2019. Panels referenced NAEP and Smarter Balanced benchmarks. New cut scores for the redesigned test are being set in summer 2026.
Stakes
No student-level consequence. Feeds school accountability — and, from 2026, the returning A–F grades.
Vendor
Cambium Assessment (formerly AIR). AIR won the 2017 contract at $43.4M over three years — above the losing $37.5M bid — and built ILEARN substantially from the Smarter Balanced item bank, four years after Indiana withdrew from Common Core on state-sovereignty grounds.

What IREAD-3 measures, versus what the statute says reading is

Indiana legislated a five-component definition of the science of reading at IC 20-18-2-17.5: phonemic awareness, phonics, fluency, vocabulary, comprehension. IDOE commissioned EdMetric, LLC to test whether IREAD-3 aligns to it. EdMetric coded 65 unique items across two forms.

IREAD-3 item composition by science-of-reading component

EdMetric alignment study, September 2023, filed with the Indiana General Assembly, as stated in its executive summary. Two of the five statutory components are measured by zero items.

check this The report is internally inconsistent. Its executive summary and Section 6 state 76.9% comprehension / 12.8% phonics / 10.3% vocabulary. Table 13 of the same report gives, for the combined 65-item sample, 78.5% comprehension / 12.3% vocabulary / 9.2% phonics — reversing the phonics and vocabulary ranks. The figures charted above are the narrative ones. The 0% phonemic awareness and 0% fluency findings hold either way, and they are the load-bearing part.

What the alignment study recommended

EdMetric rated IREAD-3 "strongly aligned" overall, rated Reading: Literature only moderately aligned with "weak alignment for Depth of Knowledge and Balance of Representation," recommended IDOE establish "defensible rationales for why comprehension dominates while fluency remains unmeasured," and suggested a complementary fluency assessment. Separately, IDOE materials and press coverage describe IREAD as assessing "phonics, phonemic awareness, vocabulary, fluency and comprehension" — a description broader than the item composition supports.

A common misreading

The same finding cuts against the common framing that IREAD is a "basic mechanics" test and ILEARN the "real" reading test. IREAD is 77% comprehension items. Two comprehension-weighted tests of the same students returning 87% and 41% proficiency differ mainly in where the cut scores were placed, not in what they measure.

Measurement error against a hard line

IREAD-3's 2024–25 technical report reports a grade-3 standard error of measurement of 28.35 scale points (marginal reliability 0.878). The pass/fail boundary is a single point: 446.

One standard error around the IREAD-3 grade 3 cut score

Scale 200–650. A student scoring 446 passes; a student scoring 420 is retained. On the state's own reported SEM those two scores are not statistically distinguishable.

Document gap

The technical reports contain tables for conditional SEM at the cut and for classification accuracy and consistency — the tables that would quantify how many of the 3,040 retained students are measurement error rather than non-readers. Those values are in volumes that are not posted publicly. IDOE's parent-facing score guide concedes the underlying point ("Scores are an estimation of your child's ability"), but individual student reports display no confidence band. Ask IDOE for IREAD-3 Technical Report Volume 4 (reliability and classification accuracy) and Volume 6 (standard setting).

ILEARN cut scores, grades 3–8

Approved by the State Board July 25, 2019 and unchanged through 2025. New cut scores for the redesigned assessment are being set in summer 2026 — these are the values the 2026 numbers will not be on.

GradeELA · ApproachingELA · AtELA · AboveMath · ApproachingMath · AtMath · Above

ELA scale runs roughly 5060–5920 and math 6080–6950, varying by grade. Science was quietly re-scaled in 2024 from a ~7350–7650 scale to a ~0–600 scale, with no lowest/highest obtainable scale scores published on the new scale — an unflagged science trend break we found no reporting on.

The science of reading policy stack

Indiana did something unusual: it wrote a definition of reading science into statute, then built licensure, coaching, curriculum and retention policy on top of that definition. Understanding the 2026 numbers means understanding what the statute actually says — and what the tests do and do not measure against it.

The goal
95%
of third graders reading by 2027 — set by Gov. Holcomb and Sec. Jenner in the Feb. 2022 priorities document
Where it stands
87.3%
2025 — the gap to goal is 7.7 points with one year left
Committed funding
$170M
Lilly Endowment plus the General Assembly, per Sec. Jenner; a further ~$75M announced Dec. 2025
Schools in the Cadre
564
2025 cohort, up from 41 schools in 2022

The 95%-by-2027 goal against the actual trajectory

The dashed segment is the path required from 2025 to hit 95% in 2027 — not a forecast. It needs +3.85 points per year, twice, against a pre-2025 record of under one point per year.

The 2025 gain of +4.8 points was the largest in the test's history and it still leaves 7.7 points to close. Reaching the goal on the original timeline requires roughly 1.6× that record gain, repeated across both remaining years. Whether IDOE restates, revises, or stops referencing the goal on August 11 is worth noting.

What the statute says

IC 20-18-2-17.5, added in 2023, defines the science of reading in three parts. The distinction between them is doing more work than it appears.

(1) Required components

The statute requires the "explicit, systematic inclusion" of exactly five things: phonemic awareness, phonics, fluency, vocabulary, comprehension. This is the National Reading Panel's list, verbatim, from 2000.

(2) Evidence base

Requires the body of research to inform how proficient reading and writing develop, why some students struggle, and how to assess and teach effectively "to improve outcomes for all students."

(3) Competency outcomes

Adds two things absent from part (1): oral language skills and writing and spelling. So the statute's outcome list is broader than its instructional requirement list — and neither mentions background or content knowledge.

An asymmetry within the definition

The statute names seven competency areas as outcomes but mandates explicit instruction in only five of them. Oral language and writing are things Indiana says the science of reading should produce, but not things it requires schools to systematically teach — and, per the state's own alignment study, not things IREAD-3 measures at all. Nothing in the definition mentions background knowledge or content knowledge, which the Albert Shanker Institute identifies as the genuine gap across all 400+ state literacy laws it reviewed, and which the comprehension research treats as central.

Indiana simultaneously cut grade 3 ELA standards from 62 to 34. If comprehension depends substantially on domain knowledge, narrowing the standards while gating promotion on a 77%-comprehension test works against the state's own goal. That is an inference, not a documented finding — but it is a testable one, and grade 7 ELA falling 3.2 points below its 2021 baseline is the kind of evidence that would show up first.

The statute's five components against the two instruments

Statutory componentShare of IREAD-3 itemsMeasured by ILEARN?Notes
Phonemic awareness0%NoCannot be assessed by multiple-choice items in this format.
Phonics12.8%Grade 3 only, via the Reading Foundations reporting categoryFour phonics items were historically read aloud by the administrator.
Fluency0%NoEdMetric recommended Indiana add a complementary fluency assessment. We found no evidence one was added.
Vocabulary10.3%Yes, within Key Ideas and Textual Support
Comprehension76.9%Yes — the dominant constructSplit across Nonfiction and Literature; Literature rated only "moderately aligned."
Oral language outcome onlySpeaking & Listening, 6–9% of the ELA blueprintNot a required instructional component under part (1).
Writing and spelling outcome only0%Yes — 34–42% of the grade 6 ELA blueprintIREAD has no constructed response of any kind. This is the largest single divergence between the two tests.
Background / content knowledge absent from statuteImplicitly, through passage topicsNot named anywhere in IC 20-18-2-17.5.

IREAD item shares are from the EdMetric executive summary; see the caveat on The instruments tab. ILEARN column is from the published grade-level test blueprints.

The build-out, 2022–2027

February 2022
IDOE publishes Indiana's Priorities for Early Literacy
Three priorities: high-quality professional development, science-of-reading implementation support, and expanded access to literacy interventions and enrichment. Sets the headline target of a 95% IREAD-3 passage rate by 2027, framed against the 2012–13 peak of 91.4%.
2022–2025
Grade 2 IREAD phases in
385 schools (spring 2022) → 771 (2023) → 1,076 (2024) → universal (spring 2025). Grade-2 passers are exempt from retaking in grade 3 and count in the grade-3 pass rate.
2023 · HEA 1558 & the SOR statute
Indiana legislates a definition of reading science and creates the early literacy endorsement
IC 20-18-2-17.5 defines the science of reading; HEA 1558 requires educator preparation programs to embed it and directs SBOE to establish an endorsement. Indiana's framework explicitly prohibits the three-cueing model. $15M appropriated in science-of-reading grants.
June 2023 · HEA 1251 (2022)
Academic standards cut roughly 25%
Grade 3 ELA from 62 standards to 34; grade 3 math from 34 to 24. Only a third of what remains is designated "essential."
2024 · SEA 1
Retention becomes mandatory; the endorsement extends to sitting teachers
Third graders who do not pass after at least two summer retakes must be retained, effective 2024–25. IC 20-28-5-19.7 is amended so current PK–5 and special education licensees must add the endorsement at the first renewal on or after July 1, 2027. Schools below a 70% IREAD pass rate must employ an endorsed instructional coach.
2024 · IC 20-28-5-19.8
A deliberately narrow waiver
A one-year extension is available — but IDOE cannot grant one until it files a report with the Legislative Council by December 1, 2026, and no waiver extends beyond July 1, 2028 absent a hardship finding. That December report is the state's own statutorily required self-assessment of whether the mandate is working.
2025 · HEA 1499
Retakes doubled; EL exemption added and narrowed
A second summer retake brings the total to five attempts before grade 4. An exemption is created for English learners at schools where at least 50% of third graders are ELs — roughly 797 students across 38 schools, sunsetting July 1, 2028. Purdue's authors note this tightened the rules: before 2025 the good-cause exemption was universal for EL students regardless of time in American schools. An administrator literacy endorsement requirement was added for July 1, 2027.
January 2026
IDOE and Notre Dame win a ~$10M federal literacy grant
Part of a national $256M Education Innovation and Research award, with WestEd. Funds high-impact tutoring and science-of-reading training for tutors, aides and parents across all 92 counties, with a rural and underserved focus.
July 1, 2027
The renewal cliff
Every PK–5 and special education licensee renewing on or after this date needs the endorsement: 80 hours of approved professional development plus the Praxis Teaching Reading (5205) exam. Same year the 95% goal comes due.

Two rollouts that shape the numbers

Left: schools in the Indiana Literacy Cadre, the state's embedded coaching program. Right: schools administering IREAD in grade 2, which feeds the grade-3 pass rate.

A limitation in the evaluation design

The Literacy Cadre is the most frequently cited evidence that Indiana's approach is working, and the state's own materials report two different effect sizes for the same year. IDOE's "Transforming K-12" document and Secretary Jenner's op-ed say Cadre schools gained 7 points in 2025, "nearly double" non-Cadre schools. The State Board's own August 2025 IREAD deck says Cadre +2.5, non-Cadre +0.2. The two are not reconcilable as stated, and the difference is large.

More fundamentally: Cadre participation is by cohort application, not random assignment. Schools that volunteer for intensive coaching differ systematically in leadership capacity, staff stability and grant-writing ability. Absent a matched comparison with pre-trend checks, a Cadre-versus-non-Cadre gap is selection, not effect. The program's public impact page cites external evaluators but discloses no sample sizes, methodology, selection criteria or limitations.

Evidence on the other side

No state runs randomized trials on its own coaching programs, and absence of an RCT is not evidence of no effect. The Cadre scaled from 41 schools to 564 across exactly the period in which Indiana posted the largest one-year IREAD gain in the test's history. Independently, the Harvard/Stanford Education Recovery Scorecard — analyzing 35 million students across 2022–2025 with no Indiana involvement — ranked Indiana 6th of 35 states in reading recovery and credited phonics-focused reform in Indiana and seven other states, noting that states rejecting phonics instruction saw no gains.

The same report's caveats cut both ways: Indiana reading is +0.09 grade equivalents above 2022 but still −0.31 below 2019, and Indiana math ranks 29th of 38. Its authors cautioned that evidence-based reading reform "may be a necessary but insufficient path to improvement."

The endorsement mandate, and who it lands on

What it requires

Approved professional development80 hours
Licensure examPraxis 5205
New PK–5 / special education licenseesafter June 30, 2025
Instructional coaches at schools under 70% IREADfrom July 1, 2025
Existing licensees, at renewalon or after July 1, 2027
Administrators (HEA 1499)by July 1, 2027

The 2024 fiscal note assumed roughly 2,700 Indiana teachers already held a literacy endorsement, estimated acquisition cost at about $7,500 per teacher, and projected no increase in total state summer-school expenditure.

The objection from the field

ISTA president Keith Gambill, April 2024: the rollout "has just been kind of the last straw for many of our teachers." The specific complaints were about scope and compensation rather than content — confusion over who is covered (a high school math teacher swept in because he holds a special education license), 80 hours plus a written exam without adequate compensation, and a teacher describing full days spent learning "a skill that he has no plans to ever teach."

Then-candidate Jennifer McCormick said the rollout "creates frustration and confusion." At the time of that reporting no waiver existed in law; IC 20-28-5-19.8 subsequently created a narrow one.

A date to note

December 1, 2026. IDOE must file a report with the Legislative Council covering progress toward the endorsement requirements, student literacy goals, remaining barriers, and whether waiver authority is still necessary. It is the only document in the entire policy stack where the state is statutorily obliged to assess its own mandate — and it lands seven months before the renewal cliff.

The numbers through 2025

Everything here is the last fully published year. Use it as the baseline against which to read whatever is announced on August 11.

ILEARN mathematics proficiency by grade

Every grade is above its 2021 baseline, and the gains are largest in the upper grades. Grade 3 is the only grade that declined in 2025.

ILEARN English/language arts, change since the 2021 baseline, by grade

The signature 2025 finding. Grades 7 and 8 are the only grades below where they started. Grade 7 ELA is 37.8% — the lowest of any grade — down 3.9 points year over year.

An open question about the cause

IDOE and board members attribute the middle-school ELA decline to pandemic disruption for that cohort. Grade 7 math for the same students rose 5.4 points over the same period. A cohort-disruption explanation would need to account for why the effect appears in one subject and not the other.

ILEARN 2024 proficiency by student group

The most recent year with a complete, cross-verified subgroup table. 2025 subgroup levels were largely not published — IDOE released changes only.

A composition effect in the gap trend

Narrowing gaps reflect movement at both ends. On ILEARN ELA, the paid-meals group fell from 54.0% (2021) to 52.6% (2024) while the free/reduced group rose from 25.9% to 28.0%. On math, paid meals peaked at 54.1% in 2023 and fell to 52.8% in 2024. Convergence produced partly by decline at the top describes a different situation than convergence produced entirely by gains at the bottom, and the two are worth distinguishing.

IREAD-3 by student group

Note a reporting practice worth flagging: IDOE's State Board decks report "Black and Hispanic students" as a single combined category against White students. Asian, American Indian, multiracial, and Native Hawaiian/Pacific Islander students do not appear on those slides at all.

Group2024 pass rate2025 change2025 (derived)

"Derived" = 2024 level plus IDOE's published 2025 change. IDOE published subgroup changes for 2025 but not levels. The Black/Hispanic combined 2025 figure of 76.1% and White figure of 92.3% come from Purdue authors citing SBOE 2025 materials, and reconcile exactly with the 2024 baseline plus the published delta.

Indiana against NAEP

The usual critique of state tests — that states set a low bar and call it proficiency — does not fit Indiana well. Indiana's ILEARN proficiency rates sit close to NAEP's, which is unusual.

Percent at or above Proficient: Indiana NAEP 2024 vs. national public

Indiana outperformed the national public average on all four NAEP measures in 2024, and Indiana reading rose while national reading fell.

MeasureIndiana NAEP 2024, proficient+Nearest ILEARN figure
Grade 4 reading34%ELA statewide ~41%
Grade 8 reading33%ELA statewide ~41%
Grade 4 math43%48.0% (grade 4, 2024)
Grade 8 math31%31.4% (grade 8, 2024)

NAEP figures are the published State Snapshot percentages, which NCES reports rounded to whole numbers. The 2026 NAEP was administered January–March 2026; results are not scheduled until early 2027.

Where Indiana's bar actually sits

Grade 8 math is nearly identical: 31% on NAEP, 31.4% on ILEARN. Indiana set its proficiency bar close to NAEP's, which is uncommon among states. The consequence is that roughly 60% of Indiana students are reported as non-proficient in any given year. Whether that is a strength (an honest standard) or a problem (a bar so high the measure stops being actionable, and one the state did not attach consequences to for five years after 2019) is a policy judgment the data alone does not settle.

Federal mapping studies reach the same place from the other direction: NCES maps most states' reading proficiency standards to NAEP Basic, not NAEP Proficient. Whichever framing you prefer, "proficient" is a policy choice, not a natural fact about children.

How the pass rate is built

The reported IREAD pass rate is a cumulative figure, not a first-attempt rate, and it is not measured on a single cohort in a single sitting. This section sets out how IDOE constructs it.

IDOE's published methodology

NumeratorGrade 3 students who pass in spring OR summer administration, plus grade 2 students who passed during the prior spring window.
DenominatorAll grade 3 students eligible in the spring window — including students with good-cause exemptions, undetermined or invalidated scores, and non-participants.
ExcludedGrade 4/5 retesters; third graders with significant cognitive disabilities; students not enrolled during the spring window.

Source: IDOE, "IREAD-3 Results & Pass Percentage Calculations." IDOE has never published a first-attempt-only figure.

Where the 2025 headline of 87.3% actually comes from

Of ~84,163 third graders in the denominator, roughly a third never sat a grade-3 IREAD at all — they had already passed as second graders and are exempt from retaking.

ComponentStudentsBasis
Third graders in the reported denominator84,163SBOE Reading Update, Nov. 5 2025
Actually sat the spring 2025 grade-3 test55,4912024–25 IREAD technical report, Table 7
Never sat a grade-3 test (passed in grade 2)~28,672derived: 84,163 − 55,491
Passed the spring grade-3 test~36,313derived: 55,491 × 65.44%
Passed a grade-3 summer retake~7,567derived: 23,558 × 32.12%
Total counted as passing~73,500IDOE press release, Aug. 13 2025
Headline pass rate
87.3%
as reported, 2025
Spring grade-3 test only
65.4%
of the 55,491 students who actually sat it
Summer retake pass rate
32.1%
of 23,558 grade-3 retakers

The technical report itself publishes a fourth figure: 77.42% for grade 3 once second-grade passers are folded in but before summer retakes. Three defensible "grade 3 pass rates" for the same year — 65.4%, 77.4%, and 87.3% — exist in the state's own documents. Only the largest one is in the press release.

Four design changes in the same window

Each of these took effect during the same period as the 4.8-point gain, and each can raise the measured rate independent of any change in reading ability. All four are defensible policy choices; all four are also confounds for year-over-year comparison.

  1. Grade 2 testing went universal in spring 2025 (from 385 schools in 2022 → 771 → 1,076 → all). Grade-2 passers count in the grade-3 numerator and never retake.
  2. Summer retakes doubled from one window to two under HEA 1499 (2025) — five total attempts before grade 4.
  3. The test went adaptive in spring 2024, having been fixed-form since 2012.
  4. Retention was enforced for the first time, changing the incentive intensity around this specific score.

We did not locate any source quantifying how much of the gain is attributable to these changes versus to instruction. It should be treated as an unquantified confound rather than a measured effect. 2026 is the first year in which all four are fully in place simultaneously.

What happened to the students who did not pass

2025: the first enforced year of the retention law

10,663 third graders finished grade 3 without passing IREAD-3. Roughly 69% of them were promoted anyway under good-cause exemptions.

Where the exemptions went

Special education~75% (~5,200)
English learners with <2 years of services~24% (~1,667)
Grade 3 math proficiency or prior retention~1%

The five statutory good-cause exemptions are at IC 20-32-8.5-2. There is no exemption for test anxiety or poor test-taking, and neither classroom data, progress monitoring, nor parent election can override a failing score.

Context on scale

Retained, 2023 (rule unenforced)412
Retained, 2025 (first enforced year)3,040
Legislative fiscal note projection, FY2026~6,350
Public forecasts at the time7,000–10,000

Between 2012 and 2023, roughly 96% of students who failed IREAD-3 advanced to fourth grade anyway. Secretary Jenner cited the gap between projection and outcome as evidence the law is working. The gap is also arithmetically accounted for by the 6,950 exemptions and the added retake opportunities; the published data does not separate these explanations.

What has not been published

Indiana retained roughly 3,000 third graders in 2025 and has not published the racial or socioeconomic composition of that group. Because about 69% of non-passers were exempted — predominantly students with IEPs or recent EL status — retention applies mainly to general-education students without a disability designation. Given that free/reduced-meal students pass IREAD at rates roughly 12 points below the paid-meal group, that population is likely to be disproportionately low-income, but the composition cannot be confirmed without a published breakdown.

What changed underneath the 2026 numbers

Four things changed at once beneath the ILEARN figure. IDOE documented its decision to defer standard setting until the redesign was operational, which concentrates the resulting break in comparability into a single year. We did not locate any public statement from IDOE or SBOE on how 2026 should be compared to prior years.

IDOE's stated approach, 2024

"In order to support comparability of test scores across years and avoid multiple years of score delays, IDOE will not engage in standard setting activities until the new through-year design is operational."

— IDOE, Spring 2025 Test Blueprint Explanation Document

2022 · House Enrolled Act 1251
Legislature orders the academic standards streamlined
Directs a reduction in the number of statewide standards.
June 2023
Standards cut ~25%; grade 3 ELA goes from 62 standards to 34
Grade 3 math from 34 to 24. Only a third of remaining standards designated "essential." ILEARN realignment deadline set at March 1, 2025.
2024–25
Some content remained on the test after being removed from the standards
IDOE color-coded the streamlined standards: gray = removed from assessment and instruction; yellow = removed from the standards but retained on the test "for comparability and validity reasons." This is a documented trade-off of instructional alignment for trend continuity. We did not locate any news coverage of it.
2024 · House Enrolled Act 1243
Legislature authorizes breaking ILEARN into through-year checkpoints
Author Rep. Bob Behning: "we're going to reduce the amount of time spent on the summative test." Sen. Jean Leising dissented: "I think we have tested kids excessively… we have not improved education."
2024–25
Checkpoint pilot
1,350+ schools, 720,000+ checkpoint administrations, ~60,000 students per grade of ~80,000 enrolled — roughly 75% coverage.
2025–26
First statewide administration of the redesigned ILEARN
Three checkpoints plus a shortened summative. Math loses its performance task. Summative window April 13 – May 8, 2026.
Summer 2026
New cut scores are set
IDOE's own milestones document: "Math, ELA, and Social Studies results will populate following Summer 2026 Standard Setting." Science results populated on the normal schedule; ELA and math did not.
August 11, 2026
Cut scores approved and results released on the same day
Standard-setting practice generally distinguishes between panels that see impact data (how many students land in each level) before recommending cuts and those that do not. Whether and when the board saw impact data is worth establishing.
December 2026
First A–F school grades in five years, calculated from these results
"Year Zero" — grades published with no consequences attached. Consequences to be set in the 2027 legislative session.

The returning A–F model, and the objections filed against it

The State Board approved the new "student-centered" accountability model unanimously on March 4, 2026; Gov. Braun and AG Rokita completed adoption May 1. Its defining feature is that points are assigned student-by-student and then averaged — "N size is 1" — across 18 success indicators in four grade bands. IREAD passage is the K–3 academic-mastery gate; ILEARN across core subjects covers grades 4–8. For grade 12, diploma seals, credentials and work-based learning are 80% of the rating.

Comments filed by AG Todd Rokita
Joint comment — EdTrust, All4Ed, NCLD, National Parents Union, UnidosUS
Only ~33% of grades 3–8 students meet proficiency in both English and math, yet few schools will receive D or F ratings. "Under any system driven by academic performance, these proficiency rates would be expected to produce far more low-rated schools. They do not."
The system has no subgroup accountability — how schools serve specific groups of students "does not influence school letter grades."
The framework elevates "nonacademic, subjective metrics" over academic proficiency; diploma seals are overweighted; SAT is optional rather than primary at grade 12; attendance is "outside schools' control."
In grades 4–8 the design lets student growth replace achievement, and schools earn full credit for a student meeting any of three or more indicators "regardless of their rigor."
Growth can "fully replace proficiency," contradicting statutory language requiring proficiency be primary. No transparent modeling of expected grade distributions was published.
Only graduates factor into high school ratings — a dropout blind spot that "strongly discourages high schools from supporting their most at-risk students."
A shared structural objection

Rokita's "few D/F schools despite 33% dual proficiency" and EdTrust's "pick-any-three menu" arrive from opposite directions but describe the same mechanism: a menu-of-indicators design with N=1 scoring tends toward compressed grade distributions, because most students earn points on some indicator. That is a testable prediction, and when Year Zero grades publish in December 2026 the grade distribution will test it. Also worth tracking: whether ≥25% of schools earn an A, which triggers a 2.5-point ratchet in the scale.

The State Board's rebuttal: for K–8, 11 of 14 metrics are objective, test-based indicators — "nearly 80%" — and the final rule guarantees a school "will never generate more points for a student who has not achieved proficiency compared to a student who has achieved proficiency."

The evidence on both sides

The research on nearly every element of Indiana's approach is mixed, and the two columns below are meant to be read together rather than separately. Left: the strongest documented criticism. Right: the strongest documented response. Sources for both are in the final tab.

Documented criticism
Documented response
A single test should never be the sole determinant of a high-stakes decision. This is the canonical standard — National Research Council, High Stakes (1999), and the AERA position statement. Purdue's Wessel-Powell, Tomkinson & Rippy argue Indiana's automatic-retention design violates it by construction.
Indiana does not use a single administration. Five attempts across two school years, a second-grade early-warning signal, and a good-cause exemption structure that exempted roughly 69% of non-passers. Jenner: "it's not just one test score on one day that defines their child."
Retention research is mixed at best and harmful at worst. Alexander, Entwisle & Dauber found retention weakens attachment to school; a Texas study reports a 19% reduction in earnings at age 26; Chicago middle-grade retention reduces high school graduation, especially among African American girls. Skiba, Scheurich & Murphy cite a synthesis of 84 studies finding an overall academic effect near zero.
The best-identified causal studies are positive. Schwerdt, West & Winters found large achievement gains in Florida persisting through grade 10 (same-grade comparison). Figlio & Özek found third-grade retention halved time to English proficiency for English learners and tripled later college-credit coursework. The negative findings mostly involve middle-grade retention or retention without supports.
The 2025 gain is mechanical. Universal grade-2 testing, doubled summer retakes, a mode change to adaptive, and vastly increased exposure to the test format all landed at once, and all raise the measured rate independent of reading skill.
Extra practice on a 77%-comprehension assessment is not a gimmick — if a student can now pass a test they could not pass before, something changed. And the Harvard/Stanford Education Recovery Scorecard, using entirely different data across 35 million students, independently ranks Indiana 6th of 35 states in reading recovery and credits phonics-focused reform.
Legislating pedagogy oversimplifies the science. NEPC's research guide argues the "science of reading" messaging is misleading; Duke & Cartwright show reading requires executive function and context beyond the simple view; Natalie Wexler argues the reform stops at decoding and neglects knowledge-building.
The Albert Shanker Institute analyzed 400+ literacy laws across 23 states and found the "just phonics" critique largely false — 41 states legislate all five pillars, adoption is bipartisan (70% of both blue and red states), and approved-materials lists are typically developed outside the legislation. Timothy Shanahan — hardly an SOR partisan — states plainly that including phonics "reduces failure and raises average literacy levels."
Indiana's statute omits what comprehension actually needs. IC 20-18-2-17.5 codifies exactly the five National Reading Panel pillars and does not mention background knowledge, content knowledge, or writing. Indiana simultaneously cut third-grade ELA standards from 62 to 34. If comprehension depends on domain knowledge, narrowing standards while gating on a comprehension-heavy test works against the state's own goal.
Even the Shanker authors, who defend SOR laws, name this as the genuine gap: laws "rarely address the role of content and background knowledge in comprehension." The state can fairly say it legislated the consensus and that content-knowledge curriculum is a next step, not a contradiction.
The endorsement mandate is burdening teachers without evidence it works. ISTA president Keith Gambill: the rollout "has just been kind of the last straw for many of our teachers" — 80 hours of PD plus an exam, uncompensated, capturing teachers who will never teach reading (a high school math teacher swept in because he holds a special education license).
Every credible account of SOR implementation identifies teacher knowledge as the binding constraint. The Literacy Cadre coaching model exists precisely because one-off PD is insufficient. IC 20-28-5-19.8 provides a hardship waiver valve, and IDOE must report to the Legislative Council by December 1, 2026 on whether the requirements are working.
The Literacy Cadre evidence does not support the causal claim made for it. Participation is by application, not random assignment. Schools that volunteer for intensive coaching differ systematically. The Cadre's public impact page discloses no sample sizes, methodology, selection criteria, or limitations — and no independent quasi-experimental evaluation exists.
The Cadre scaled from 41 schools (2022) to 564 (2025) alongside the largest one-year IREAD gain on record and a top-10 national reading-recovery ranking. Absence of a randomized trial is not evidence of no effect, and no state runs RCTs on its own coaching programs.
Adding checkpoints increased total testing. Three checkpoints × two subjects = up to six additional sittings per student per year, while the summative lost one math performance task — and the checkpoints feed neither the score nor accountability.
The problem being solved is real: under the old design a student tested in spring and results arrived in late summer, by which point "they likely had a new teacher." Jenner's rationale is actionability, not measurement — beginning, middle, and end-of-year data to drive intervention while the student is still in the room.
Scope limits on the retention research

Every credible causal estimate above is a regression discontinuity at the cut score — it identifies the effect on students who fell just barely below the line. Indiana retained 3,040 students in 2025, many of them nowhere near the margin. The literature has essentially nothing to say about the effect of retention on a student scoring far below the cut, which is exactly the population a mandatory gate captures. Özek's 2026 synthesis names this gap explicitly.

A second caveat: Florida's positive results came bundled with a mandated 90-minute reading block, a highly qualified reading teacher, and mandatory summer reading camps. Indiana's SEA 1 fiscal note states no additional state expenditure for summer school — the existing $18.36M appropriation was reprioritized. Importing Florida's gate without Florida's treatment is not importing Florida's policy.

Release-day checklist

Questions to put to the numbers — and, where relevant, to IDOE — as the August 11 results are presented. Tick them off as they are answered. Progress is saved only in this browser tab.

Documents to request

Sources, caveats and known conflicts

Every figure on this page traces to a public source. These are the places where the public record is thin, internally inconsistent, or simply absent — which is itself part of the finding.

Figures we could not verify

Where sources conflict

ItemThe conflictWhat this page uses
2025 "both subjects"31.2% (Chalkbeat, Indiana Capital Chronicle, reporting off the July 2025 release) vs 33% (WFYI, August 2026)31.2% — consistent with the 30.2 → 30.6 → 30.8 trajectory. Note IDOE's own release contains no combined figure at all.
Literacy Cadre effect, 2025IDOE and Jenner's op-ed: Cadre schools +7 points, "nearly double" non-Cadre. SBOE's own August 2025 deck: Cadre +2.5, non-Cadre +0.2.Both, flagged. These cannot both be the headline effect for the same year. Ask for the definition before treating "Cadre works" as established.
IREAD launch year2012 (technical reports, StateImpact) vs 2013 (IDOE press materials, Purdue)Spring 2012 first administration; 2013 is the first year in IDOE's own longitudinal chart.
IREAD item count40 at launch (2012) · 37 (2020–21 report) · 38 (2021–23 reports)~38, noting the drift.
IREAD testing time"90 minutes of working time" (2023–24 FAQ) vs untimed (2025–26 FAQ and current IDOE page)Untimed, ~90 minutes of work, across at least two days.
Retention arithmetic, 20253,040 retained + 6,950 exempted = 9,990, against ~10,663 who did not pass — roughly 670 students unaccounted for in any source.Both figures shown; ~673 attributed to students who left the state or moved to homeschooling.
Purdue's $105.6M retention cost$105.6M ÷ $13,200 = exactly 8,000 students, but the same article reports 3,040 retained (≈$40.1M).Flagged, not used. The figure appears to use a projected or exemption-inclusive population.
Schools meeting the 95% goal241 (2023) per IDOE vs 261 per Indiana Public Media; 284 + 179 = 463, not the reported 456.IDOE's series: 241 → 284 → 456, with the inconsistency noted.
EdMetric item compositionExecutive summary and Section 6: 76.9% comprehension / 12.8% phonics / 10.3% vocabulary. Table 13 of the same report: 78.5% comprehension / 12.3% vocabulary / 9.2% phonics — phonics and vocabulary ranks reversed.The narrative figures, flagged on the chart. The 0% phonemic awareness and 0% fluency findings are identical in both.
Schools at 95%, 2025The SBOE deck reports 456 schools and describes it as "an increase of 179 from last year," which implies a 2024 base of 277, not the 284 IDOE published.IDOE's press-release series (241 → 284 → 456), with the deck's delta noted as the outlier.
NAEP decimalsNCES publishes state snapshot percentages rounded to whole numbers; unrounded values from the NAEP Data Service differ by up to half a point at the rounding boundary.The published rounded figures.
Number of school corporations291 school corporations (2022) vs 444 operating LEAs including charters (2023–24)Both, as distinct definitions — they are not interchangeable.

Derived figures on this page

Marked as derived wherever they appear. 2021 statewide ILEARN ELA of ~40.5% is not directly published — it is reconstructed two independent ways that agree exactly (IDOE's stated +0.1 from 2021 to 2025's 40.6%, and Chalkbeat's +0.5 from 2021 to 2024's 41.0%). The IREAD 2025 funnel components (36,313 spring passers; 7,567 summer passers; 28,672 never tested in grade 3) are our arithmetic from the technical report's Table 7 and the SBOE totals; they reconstruct the published 73,500 / 87.3% to within about one point. 2025 IREAD subgroup levels are 2024 levels plus IDOE's published deltas.

Primary sources