All data comes from the New York State Education Department public downloads at data.nysed.gov/downloads.php. Five tables were extracted from the SRC (ESSA accountability) Microsoft Access database, plus one table from the separate Enrollment Database:

• Annual EM ELA — Grade 3–8 English Language Arts proficiency, by district and subgroup
• Annual Regents Exams — Regents exam results; filtered to Common Core English (the HS ELA analogue)
• ACC EM Chronic Absenteeism — Chronic absenteeism for Grades 1–8
• ACC HS Chronic Absenteeism — Chronic absenteeism for Grades 9–12
• BOCES and N/RC — District classification lookup including Need-to-Resource Capacity category
• BEDS Day Enrollment (Enrollment Database) — True K–12 enrollment totals, used as the denominator column throughout the District Explorer

Each release contains two school years of data. Earlier years were stitched together from multiple releases (SRC2019, SRC2022, SRC2024 plus their matching Enrollment releases) to form the time series on the Trends page.

For Grade 3–8 ELA, this dashboard computes proficiency as NUM_PROF / TOTAL_COUNT rather than NYSED’s published PER_PROF (which equals NUM_PROF / NUM_TESTED). The difference is the treatment of students who did not take the test: our denominator includes them, NYSED’s does not.

This is the more conservative choice. Several New York districts — particularly in the Hudson Valley and on Long Island — have historical opt-out rates above 50%. Under PER_PROF, a district where only the most academically secure half of students sat for the exam can appear dramatically more proficient than its peers. Using TOTAL_COUNT as the denominator treats non-testers as non-proficient, which matches NYSED’s “Weighted Average Achievement” methodology in spirit.

As a result, proficiency rates in this dashboard will not match what you see on data.nysed.gov. This is intentional and the gap is largest in low-participation years (notably 2021, post-COVID) and in high-opt-out districts.

The EM 3–8 series and the Regents HS series are displayed separately because they measure different things: different grade-level standards, different test designs, and a different cohort (Regents is essentially 11th-grade test-takers, not the full grade level). EM 3–8 is the headline metric on every page except the District Explorer, both because it has the longest time series and because — unlike the Regents table — it has a TOTAL_COUNT column. The Regents table provides only TESTED, so the HS ELA column in the District Explorer necessarily falls back to PER_PROF and is not opt-out adjusted. Compare HS ELA across districts with this asymmetry in mind.

NYSED defines a chronically absent student as one who is absent for 10% or more of enrolled instructional days. All absences count — excused, unexcused, and out-of-school suspensions. The denominator is students who were enrolled for at least 10 instructional days and present for at least one of those days; students who never attended are excluded. In districts with high student mobility, that exclusion matters.

The Grades 1–8 figure (ACC EM Chronic Absenteeism) and the Grades 9–12 figure (ACC HS Chronic Absenteeism) are published as separate tables. This dashboard combines them into a single K–12 rate using count-based math, not rate averaging:

    combined_rate = (em_absent_count + hs_absent_count) / (em_enrollment + hs_enrollment)

Summing counts rather than averaging rates means the combined figure is correctly weighted by the number of students in each band, and it degrades gracefully when only one side is available (for example, K–8 districts with no high school).

An earlier version of this dashboard reported the Grades 1–8 figure only. That matched NYSED’s ESSA federal indicator scope but produced enrollment numbers in the District Explorer that were systematically 30–40% lower than the district totals published on data.nysed.gov — a confusing discrepancy for anyone cross-checking. The switch to K–12 resolves that.

New York has no statewide assessment of K–2 literacy. Education Law §818 requires every district to attest that it has adopted a literacy curriculum aligned to the “science of reading,” but there is no central data collection on actual outcomes in those grades. This is a limitation of the dataset, not a choice — the earliest grade-level reading proficiency we can observe is Grade 3, and any analysis purporting to measure “early literacy” from this data is by definition measuring the tail end of it.

NYSED suppresses assessment values when the tested group has fewer than five students (the small-N rule). Suppressed cells appear as "s" in the source database and are coerced to missing in the cleaning pipeline.

The small-N rule is a privacy threshold, not the accountability threshold. Accountability determinations under the NY ESSA plan use a stricter n-size: 30 students for performance calculations and 40 for participation. This dashboard is descriptive, not an accountability scorecard, so it reports every cell that survives small-N suppression. Readers should be careful not to read a single-year district value here as a formal accountability result.

All numeric fields in the source databases are stored as 255-character text strings — an Access quirk noted in the NYSED ReadMe. The cleaning pipeline explicitly coerces these to numeric and drops any row where the coercion fails. Reporting year labels follow the NYSED convention: “2024” means school year 2023–24.

NYSED’s Need-to-Resource Capacity index is a measure of a district’s ability to meet its students’ needs with local resources. It is computed from the ratio of estimated poverty percentage to the Combined Wealth Ratio, both expressed in standard-score form. Districts are sorted into six categories:

1. New York City (its own category)
2. Large City Districts — Buffalo, Rochester, Syracuse, Yonkers
3. High Need Urban-Suburban — above the 70th percentile of need, urban or suburban density
4. High Need Rural — above the 70th percentile of need, rural density
5. Average Need — between the 20th and 70th percentile
6. Low Need — below the 20th percentile

This dashboard uses N/RC as its primary framing so that district comparisons are made among peers. Comparing raw outcomes across N/RC categories without context is the single most common way this data gets misread in public conversation.

This is a descriptive dashboard. The correlations shown are not causal. A district with both high absenteeism and low proficiency is not necessarily a district where absenteeism caused the low proficiency — both may share upstream causes (poverty concentration, housing instability, teacher vacancies) that the data does not capture.

The data also does not reflect individual student trajectories. A district-level proficiency rate of 45% tells you nothing about which students those are or whether the same students struggled year over year.