How much of the 2024-25 'recovery' is a test-recalibration artifact?

Thesis

Citywide grade-3 math proficiency jumped from 49.6% (2022-23) to 61.8% (2024-25) — too fast for a real instructional turnaround. NYS revised the math standards and tests in 2023; quantify how much of the gain is the test getting easier vs. actual recovery.

Supporting findings

• Average grade-3 math proficiency by year: 50.9 (2017-18), 52.6 (2018-19), 46.9 (2021-22), 49.6 (2022-23), 49.0 (2023-24), 61.8 (2024-25).
• The 2024-25 jump is 12 points in one year — implausibly fast for normal instructional change.
• NYS revised the math test in 2023 (new 'Next Generation Learning Standards'); the first year under new standards is often a re-baseline.
• Compare proficiency rate change to mean scale score change — if scale scores stayed flat but proficiency jumped, that's a cutscore change.
• Apply test-equating literature (Andrew Ho at Harvard, Sean Reardon at Stanford): convert proficiency rates to standard deviations of student performance.
• Compare NYC's gain to NY state average — if state-wide pattern, it's test design, not NYC-specific recovery.
• Compare to ELA same year (also recalibrated?). ELA went 50.6 to 56.5 in same window.
• What does NAEP say about NYC fourth-graders 2024 vs 2022? NAEP wasn't recalibrated.
• Use the year-over-year reading change in our composite_ms_academic and composite_es_academic for a side-by-side magnitude check.
• The honest framing: 'Some of this is real recovery; some is the test getting easier. Here's the math.'

Reporting directions

• Email NYSED for the technical report on the 2023 standards revision.
• Interview Andrew Ho or Sean Reardon on best-practice test-equating.
• Build a chart with two y-axes: proficiency rate (left), NAEP-equivalent scale score (right).
• If the gap can be quantified, write the methodology piece and link any 'NYC schools recovered!' headlines to it.