← All proof trees · AMY NORTHWEST

1. 1
   What you seeWithheld by the publisher

   The publisher suppressed this cell (too few students to report). The site stores it as suppressed and never as zero.

   vs. peers

   —

   vs. citywide

   —

   unit

   percent

   direction

   higher is better

2. 2
   How it was computedvalue = the state's published PercentPersistentAttendance (2021-22+ workbooks) or 'Percent Regular Attendance' (2018-19, 2019-20, 2020-21 workbooks), as-is; served year = workbook year − 1

   1. Read the state's wide-format datafile: one row per Pennsylvania school, one column per measure × subgroup (e.g. PercentPersistentAttendance_Black).
   2. Keep Philadelphia rows by mapping the state's AUN + Schl codes to the district's ULCS code through the Master School List.
   3. Unpivot each recognized measure × subgroup column into one served cell; state suppression codes ('IS', '--', '*') are stored as suppressed.
   4. For the 2018-19, 2019-20, and 2020-21 workbooks the older long-format files are read instead: one row per school × data element; the 'Percent Regular Attendance (<subgroup>)' elements are kept.
   5. Stored under the attendance year (workbook year − 1), per PDE's Future Ready glossary: 'a lagging indicator indicating data is from the year prior to the reporting year'.

   Note: stored percentiles are computed by a batch script after loads; a difference from the live reconstruction means a newer load changed the comparison set and the batch has not been re-run.

   Evidence: metric tree: Regular attendance rate (state)

3. 3
   The stored cellThe stored cell

   table

   philly_school_year_metrics

   value

   null

   denominator

   null (not published)

   suppressed

   true

   citywide_percentile

   null

   comparison_group_percentile

   null

   source_load_id

   7ef7c64a-acd5-4eed-b836-38f328274a2d

   Evidence: data model overview

4. 4
   The load that wrote itWritten on 2026-09-08 by futurereadypa_longformat_attendance.ts

   source id

   futurereadypa_performance_2018-19

   file URL

   https://futurereadypa.org/home/getdatafile?id=20

   SHA-256 of the file read

   e338ebd1985196c003a934718fadb610d5ddb0563d2a4bc4ea004931d01cb95bCopy hash

   bytes

   37,114,681

   year covered

   2018-19

   year of this cell

   2017-18 — this file's attendance indicator is a lagging indicator (the state's glossary: data from the year prior to the reporting year), so the 2018-19 workbook's value is stored under 2017-18

   code commit

   c19ea6ec845c65bcc40eb49644be1c7e67b05f25

   notes

   long-format attendance only ('Percent Regular Attendance (<subgroup>)' → attendance_persistence_rate); lagging indicator stored under 2017-18

   To confirm you have the same file: download it from the URL above, then run shasum -a 256 <downloaded file> Copy command — the output should equal the hash above. If the publisher has since replaced the file, the hashes will differ, which is itself useful to know.

   Evidence: loader script · first-fetch findings for this file · source registry

5. 5
   The publisher's fileFuture Ready PA Index datafile, 2018-19 (older long format)

   Pennsylvania Department of Education (Future Ready PA Index). Older-format state datafile; only the attendance element is loaded. The pre-pandemic anchor for the state's attendance measure.

   format

   XLSX (long)

   join key

   AUN + Schl → ULCS

   landing page

   https://futurereadypa.org/Home/DataFiles

   • Indicator named 'Percent Regular Attendance' — same 90%+ definition (documented assumption).
6. 6
   How it was checkedChecks: 1 passed, 1 flagged

   • ✓ passedBase case: every served cell re-read from the seven state workbooks under the aligned year and matched
   • ⚠ flaggedCross-publisher vs the district's 90%+ figure for the same schools and the same attendance year: r ≥ 0.99 every year, stable offset, NOT interchangeable under the pre-committed splice threshold

   Evidence: Base case · Cross-publisher vs the district's 90%+ figure for the same schools and the same attendance year

7. 7
   Assumptions & caveatsAssumptions and caveats

   • The state's AUN + Schl → ULCS mapping from the 2024-25 Master School List is correct. State rows for schools not in that directory — district schools that closed before 2024-25 — are not loaded, so every year of state data is a survivor-only panel; cyber charters have their own AUNs and are outside the district's directory.
   • Where two directory schools share a state code (a school and its Continuation Academy), the state's single row belongs to the main school (deterministic rule; the full absenteeism QA lists every shared code).
   • 'Percent Regular Attendance' (older workbooks) and 'PercentPersistentAttendance' (newer workbooks) are the same indicator: PDE's glossary defines Regular Attendance with the 90-day enrollment and 90% attendance rules, and the aligned cross-publisher sweep shows the state's figure tracks the district's for the same schools in every year from 2020-21 on (r ≥ 0.99).
   • PDE's enrollment rule (students enrolled 90+ school days) differs from the district's (10+ days at the school). The state's figure runs a median ~0.8 to 1.7 points higher than the district's for the same schools in the same year; the eligibility difference is a plausible contributor but has not been measured, and local attendance coding (PDE lets each LEA decide how partial days count) is another candidate.

   • No student counts published, so no denominators are served.
   • Latest attendance year is 2023-24 (from the 2024-25 workbook); the district's own file runs a year later.
   • Before 2026-09-07 this metric was stored under the workbook year — one year too late — and the site wrongly said the 2019-20 workbook was never published. Both corrected after an external audit.

   Algorithms involved: Reading the state's wide-format datafile, Chronic absenteeism as 100 minus the 90%-attendance share, Citywide and peer-group percentiles, Peer groups: 40 nearest neighbors on demographics