Find What Works: Schools
← All source documentsRaw file
Source document

docs/qa_reports/nysed/NARROWING.md

Served verbatim from the project repository. Internal working document conventions apply: documents may reference file paths, branch names, and findings-ledger anchors from the repo.

2024-25 EM Divergence — Cause Narrowing

Companion to RECONCILIATION.md §6. Date: 2026-06-09. Scope: documents only — what each authoritative source actually says about the 2024-25 chronic-absenteeism (CA) calculations. Downloaded documents saved under nysed-recon/docs/ (PDF + extracted .txt).

Cause narrowing (document review)

1. NYSED SRC ReadMes — SRC2025 vs SRC2024 (local)

Paths: ../nysed-scout/SRC2025ReadMe.txt (PDF SRC2025ReadMe_Group3.pdf, "Last Modified: April 30, 2026") and ../nysed-scout/SRC2024ReadMe.txt (PDF SRC2024ReadMe_Group5.pdf, "Last Modified: August 22, 2025").

SRC2025, Table "ACC EM Chronic Absenteeism":

"(Note: Chronic Absenteeism was not used as an indicator in 2025 for determining accountability statuses and support models. Chronic Absenteeism is still reported in NYS School Report Cards for information and improvement purposes.)"

ENROLLMENT: "Number of students in grades 1-8 (and ungraded age equivalent) enrolled for 10 or more instructional days and present for at least one of those days."

ABSENT_COUNT: "Number of students in grades 1-8 (and ungraded age equivalent) enrolled for 10 or more instructional days, present for at least one of those days, and absent for 10% or more of enrolled days"

ABSENT_RATE: "Chronic absenteeism rate for current reporting year (ABSENT_COUNT / ENROLLMENT) * 100"

SRC2024, same table:

ENROLLMENT: "Number of students in grades 1-8 (and ungraded age equivalent) enrolled for 10 or more instructional days and present for at least one of those days."

ABSENT_COUNT: "Number of students in grades 1-8 (and ungraded age equivalent) enrolled for 10 or more instructional days, present for at least one of those days, and absent for 10% or more of enrolled days"

The substantive definitions are word-for-word identical. Differences between the two ReadMes are limited to: (a) the informational-use note above; (b) SRC2025 drops the LEVEL/OVERRIDE columns; (c) the PARTIAL_DATA_FLAG note encodes the min-N change ("null if (a) subgroup not accountable due to fewer than 30 enrollments in 2024 or (b) subgroup had fewer than 20 enrollments in 2025"). Neither ReadMe says anything about which absence types count.

Establishes/rules out: no NYSED file-level definitional change for the 2024-25 CA tables. Min-N affects suppression only, not rate calculation.

2. NYSED Report Card glossaries, 23-24 vs 24-25

URLs: https://data.nysed.gov/files/essa/24-25/glossary.pdf and https://data.nysed.gov/files/essa/23-24/glossary.pdf (saved as docs/glossary_24-25.pdf/.txt, docs/glossary_23-24.pdf/.txt).

24-25 glossary, "Chronic Absenteeism":

"A Chronic Absenteeism Rate is calculated at the elementary/middle level for students in grades 1-8 (and ungraded age equivalent) and at the secondary level for students in grades 9-12 (and ungraded age equivalent) who were enrolled for a minimum of ten instructional days and in attendance at least one of those days (Students Enrolled) who were absent (excused or unexcused) for at least 10% of enrolled instructional days (Students Chronically Absent)."

23-24 glossary, "EM Chronic Absenteeism" (HS entry parallel):

"Percentage (Chronic Absenteeism Rate) of grades 1-8 students enrolled for a minimum of ten instructional days and in attendance at least one of those days (Students Enrolled) who were absent (excused or unexcused) for at least 10% of enrolled instructional days (Students Chronically Absent)."

Identical in substance: 10-day window, present ≥1 day, "absent (excused or unexcused)". The 24-25 glossary moves CA out of the accountability section and adds the new "Elementary/Middle Attendance" indicator, which is a different measure:

"A Student Attendance Rate is determined for each student in Grades 1-8 and ungraded age equivalent students enrolled 30 or more cumulative instructional days and present (Present, Tardy) for at least one of those days by dividing the number of days the student is present by the number of cumulative instructional days the student is enrolled." … "100 * [((number of students scoring Level 2) + 2(number of students scoring Level 3) + 2.5(number of students scoring Level 4)) ÷ (number of students enrolled ≥ 30 cumulative instructional days and present for at least one of those days)]"

Establishes/rules out: no glossary-level change in the CA definition for 2024-25. The Attendance Index uses a different inclusion window (30 days vs 10) and counts tardies as present, but it is a separate table (ACC EM/HS Attendance) and per the SRC2025 ReadMe does not replace the CA tables' calculation.

3. NYSED SIRS-107 Chronic Absenteeism Accountability Report Guide

URL: https://www.p12.nysed.gov/irs/level2reports/sirs-107-chronic-absenteeism-accountability-report-guide.pdf (saved docs/sirs-107-guide.pdf/.txt; "Last updated: November 19, 2025"; change log entry for that date: "Clarification of which years Chronic Absenteeism were used for accountability status determinations and how the definition of enrollment changed in 2025-26.")

"In 2024-25, chronic absenteeism was replaced by Attendance as an accountability measure. Beginning in 2024-25, chronic absenteeism will be reported in a separate section (not in the Accountability section) of the New York State Report Cards for informational purposes but will not be used to make accountability status determinations for the following year."

"Prior to 2025-26, students are considered chronically absent if they were enrolled for 10 or more instructional days, in attendance for at least one of those days, and absent (excused or unexcused) for 10% or more of their enrolled instructional days. Beginning in 2025-26, students will be considered chronically absent if they were enrolled for 30 or more instructional days, in attendance for at least one of those days, and absent (excused or unexcused for 10% or more of their enrolled instructional days."

"The underlying data used to create this report are in the SIRS-360 Attendance Absence Daily Summary, SIRS-361 Year-to-Date Attendance/Absenteeism, and SIRS-370 Day Calendar Summary reports. These reports are required to be certified."

Establishes/rules out: NYSED's own post-hoc documentation dates the only CA-definition change (10-day → 30-day enrollment window) to 2025-26, not 2024-25. For 2024-25 the definition is explicitly unchanged. Also establishes that NYSED CA is computed from district-certified SIRS attendance submissions.

4. NYSED SIRS Manual 2024-25 (Version 20.11)

URL: https://www.nysed.gov/sites/default/files/programs/information-reporting-services/sirs-manual-2024-2025.pdf (saved docs/sirs-manual-2024-2025.pdf/.txt, 311 pp.)

Attendance codes (p. 145): "State attendance code used to indicate student is present (PRSNT-IN, PRSNT-OUT), excused (E), unexcused (U), tardy (T-IN, T-OUT), in-school suspension (ISS), or out-of-school suspension (OSS)."

Daily attendance rules (p. 48):

"LEAs must report both positive and negative attendance using the Student Daily Attendance template. … There must be an attendance record for every student on all instructional days. Failure to report a record (missing data) will default to an absence on that day."

"On days when a student is suspended, both Level 0 and Level 2 require two attendance records. One record must have the suspension code (ISS … or OSS …). The second record must have one of the following attendance codes: PRSNT-IN, PRSNT-OUT, E (excused absence), U (unexcused absence), T-IN (tardy in), T-OUT (tardy out). Reporting of codes ISS or OSS without an accompanying PRSNT-IN, PRSNT-OUT, E, U, T-IN, or T-OUT will be considered MISSING ATTENDANCE and will default to absent, unexcused absence (U)."

Homebound students (p. 62): attendance prorated by hours of instruction received; shortfall days "recorded using one of the 'absent' student attendance codes."

Revision history (vv. 20.0 Aug 13, 2024 → 20.11): no entry touches student attendance codes, daily-attendance rules, or chronic-absenteeism calculation (entries cover staff templates, CTE, restraint events, course codes, etc.).

Establishes/rules out: no SIRS reporting-rule change for student attendance in 2024-25. Separately documents a concrete mechanism — missing or incompletely paired daily attendance records default to unexcused absence in NYSED's system — by which a district's submission gaps inflate NYSED-side absence counts with no definitional change on either side. Whether this occurred in NYC's 2024-25 EM submission is not established by any document found.

5. NYSED "Attendance Matters!" Board of Regents deck (Jan 13, 2025)

URL: https://www.regents.nysed.gov/sites/regents/files/AUDBF%20-%20Attendance%20Matters.pdf (saved docs/attendance-matters-jan2025.pdf/.txt; linked from https://www.nysed.gov/innovation-school-reform/chronic-absenteeism)

p. 3 (on the OSC "Missing School" audit):

"New York excludes suspensions and extended medical absences from the State's chronic absenteeism calculation."

pp. 12-14: "Proposed Indicators for 2024-25 SY Results … Attendance EM/HS — New indicator"; "a new Attendance indicator will replace the Chronic Absenteeism Indicator … Inclusion criteria increases to 30 instructional days"; index formula as in §2 above.

Establishes: NYSED states, in its own words, that suspensions and extended medical absences are excluded from the State CA calculation — presented as standing policy (context: responding to the Oct-2024 OSC audit), not as a change effective 2024-25. Also confirms the Attendance Index is a new, separate indicator first used for 2024-25 results.

6. NYC DOE StART Educator Guide (DOE's guide to the NYSED data)

URL: https://tools.nycenet.edu/start/guide (fetched 2026-06-09; guide's framework coverage runs through 2022-23)

"Suspensions are not counted as excused or unexcused absences, as suspended students are required to be provided with instruction while they are suspended."

(in its description of the NYSED Chronic Absenteeism metric; inclusion rule quoted there matches NYSED: "enrolled in a school or district for a minimum of ten instructional days that school year and be in attendance at least one of those days").

Establishes: NYC DOE itself documents the suspension exclusion as a feature of the state CA measure. Nothing in the guide describes a 2024-25 change.

7. NYC DOE EOY file NOTES sheet (the DOE side's own definition)

Path: nyc-edu-data/data/reference/sources/nycdoe_attendance/2026-06-01/public-school-attendance-results-2019-2025.xlsx, sheet NOTES (full text reviewed; key paragraph):

"Starting in 2020-21, the NYC DOE transitioned to NYSED's definition of chronic absenteeism. Students are considered chronically absent if they have an attendance of 90 percent or less (i.e. students who are absent 10 percent or more of the total days). In order to be included in chronic absenteeism calculations, students must be enrolled for at least 10 days (regardless of whether present or absent) and must have been present for at least 1 day. The NYSED chronic absenteeism definition is applied to all prior years in the report."

The NOTES sheet contains no 2024-25-specific note, no statement about which absence types count (suspension/medical handling unmentioned), and no methodology-change entries after the 2022-23 universe notes (D75 short-term-stay, District Pre-K Centers).

Establishes/rules out: DOE's published methodology for the EOY file is uniform across 2018-19–2024-25 by its own documentation; if DOE changed anything in 2024-25, it is undocumented in the file.

8. Mayor's Management Report FY2025 (DOE chapter + indicator definitions)

URLs: https://www.nyc.gov/assets/operations/downloads/pdf/mmr2025/doe.pdf and https://www.nyc.gov/assets/operations/downloads/pdf/mmr2025/2025_idf.pdf (403 to plain curl; fetched with a browser UA; saved docs/mmr_doe.pdf/.txt, docs/mmr_2025_idf.pdf/.txt).

Indicator definition (idf, "Students chronically absent (%)"):

"The percentage of students whose attendance rate for the year is less than 90%, as calculated by the number of students with an attendance rate of less than 90% divided by the number of students on register, not including charter school students. Source: Automate the Schools Data System"

DOE chapter:

"the proportion of students chronically absent decreased from 34.8 percent to 33.3 percent" (2023-24 → 2024-25)

and the noteworthy-changes section records only small prior-year revisions ("'Students chronically absent (%)' was changed from 36.1 to 36.2" for FY23; FY22 40.6→40.7) — no methodology note for FY25.

Establishes: DOE's second publication channel (MMR, sourced from ATS) shows the same direction as the EOY file — CA falling in 2024-25 — i.e., the DOE side is internally consistent across its own publications; the divergence is squarely DOE-vs-NYSED, not an artifact of one DOE file.

9. NYC Open Data (the hoped-for third vintage) — not available for 2024-25

Socrata API checks (2026-06-09):

  • sgsi-66kk "2016-17 - 2020-23 Citywide End-of-Year Attendance and Chronic Absenteeism Data": distinct year values 2018-19…2022-23 in API; no DBN column; last data update Feb 2024.
  • gqq2-hgxd "2016-17 - 2020-21 School End-of-Year …": distinct years 2016-17…2020-21.
  • Catalog search (api.us.socrata.com/api/catalog/v1?...q=attendance) surfaces no NYC school-level EOY attendance dataset covering 2023-24 or 2024-25.

Establishes: there is no independent third vintage of 2024-25 school-level EM data on NYC Open Data; the school-level series stops at 2020-21. The spot-check in CHECKS item 4 cannot be run.

10. Other NYC-side policy documents (no 2024-25 change found)

11. NYSED Attendance-table data (Check 5)

Only the CA tables were exported before the mdbs were deleted (nyc-edu-data/data/reference/sources/nysed_src/2026-06-09/extracted/ contains SRC*_EM_CA.csv/SRC*_HS_CA.csv only). Per the SRC2025 ReadMe, the "ACC EM Attendance" table's basis differs from CA by construction (30-day window; LEVELn_COUNT bins of percent of days present; tardy treatment per the glossary). Documentation does not state a different absence-type basis for the Attendance table vs CA; both are silent in the ReadMe on absence types. No data comparison was run (would require re-downloading the SRC2025 zip).

Final assessment (no-speculation)

VERIFIED:

  1. No documented NYSED-side change to the chronic-absenteeism calculation for 2024-25. The SRC2024/SRC2025 ReadMe definitions are word-for-word identical; the 23-24 and 24-25 glossary entries are substantively identical; SIRS-107 explicitly dates the only definition change (10→30-day enrollment window) to 2025-26; the SIRS Manual 2024-25 revision history contains no student-attendance changes. The 2024-25 changes that did occur (CA demoted to informational; Attendance Index added; min-N 30→20) affect accountability use and suppression, not the published CA rate formula.
  2. No documented NYC DOE-side change for 2024-25. The EOY file's NOTES claim one uniform methodology for all years; the FY2025 MMR (separate channel, ATS-sourced) shows the same falling direction with no methodology note; Chancellor's Reg A-210 is unchanged since 2017.
  3. The two measures differ by standing, documented policy — in the wrong direction to explain 2024-25. NYSED (Regents deck, Jan 2025): "New York excludes suspensions and extended medical absences from the State's chronic absenteeism calculation"; DOE's StART guide concurs for the state measure; DOE's own definition is silent on absence types. An exclusion on NYSED's side predicts NYSED ≤ DOE — consistent with the small positive diffs in 2018-19/2023-24, and the opposite of the 2024-25 EM gap (NYSED higher by ~1.8pp median). A standing definitional difference also cannot explain a single-year, single-band divergence.
  4. A documented mechanism exists that can inflate NYSED numerators without any definitional change: under SIRS rules, a missing daily-attendance record — or a suspension record not paired with a present/absent code — "will default to absent, unexcused absence (U)" (SIRS Manual p. 48). This is a mechanism, not a finding that it occurred.
  5. No third public vintage of 2024-25 school-level data exists (NYC Open Data EOY series ends 2020-21 at school level), so publisher-vs-publisher cannot be triangulated from public data.

UNESTABLISHED: which pipeline changed in 2024-25 — NYC's SIRS submission to NYSED (data-quality/coding-completeness, per the default-to-absent mechanism), NYSED's processing of it, or an undocumented DOE EOY-calculation change. No public document records any 2024-25 change on either side, and the EM-vs-HS asymmetry is not addressed by any document found.

Single next best step: email NYSED's Office of Information and Reporting Services via datasupport@nysed.gov (the published contact for data.nysed.gov/report-card data questions; accountability questions: accountinfo@nysed.gov), citing one concrete discrepancy with byte-identical denominators (e.g., Brooklyn Tech 2024-25: NYSED ENROLLMENT 5,840 / ABSENT_COUNT 979 vs DOE-published 14.195% ⇒ 829), and ask: (a) whether the 2024-25 CA tables were computed under the same absence-counting rules as 2023-24, and (b) whether NYC's 2024-25 SIRS daily-attendance submission had completeness issues that defaulted records to absent. (Equivalent DOE-side contact: NYC DOE attendance/research team via the InfoHub EOY data page; NYC DOE can also see the student-level SIRS-107 detail report in L2RPT, which neither publisher's public files expose.)