Chronic absence and attendance tiers

How the attendance rate is worked out, why the denominator is days enrolled, what each tier means — and why the card shows no rate at all until you upload a full roster instead of an absentee-only export.

At the bottom of the Attendance tab is the chronic-absence card. It is the most consequential number in the product, because a tier decides whether a family gets a letter, a meeting, or a home visit — and you have to be able to defend that to a parent and to an authorizer.

Why there is no risk score

There isn't one, on purpose. A score cannot be argued with. What the card gives you instead is a numerator, a denominator, a threshold and a sentence: open any student's row and you see the arithmetic that produced their tier, line by line, in numbers you can check against your own system in a minute.

The line

New York (and ESSA reporting) counts a student chronically absent when they are absent 10% or more of the days they were enrolled. So the line is 90% attendance, and three rules follow from it — all three implemented literally, because getting any of them wrong makes the number look better than the truth:

• Excused and unexcused absences count the same. Chronic absenteeism measures instructional time lost, not whether a parent followed the rules. A doctor's note does not put the child back in the classroom. Your discipline and truancy processes care about the excuse; this measure does not.
• The denominator is days ENROLLED, never days in the school year. A student who joined in February is measured against their own enrollment. They are never marked chronic for days before they existed at your school.
• A tardy is a present day. Late is not absent. A half day counts as half a present day. An out-of-school suspension is an enrolled day that is not a present day, which surprises people, so it is spelled out in the arithmetic.

The tiers

• Tier 1 — universal (95% and above). No individual outreach. The daily absence notice you already send *is* the intervention, and an automated "your child has 96% attendance" text is noise that teaches families to ignore the channel you need for the real ones.
• Tier 2 — at risk (90–94.9%). A check-in, and the student goes on the weekly attendance review.
• Tier 3 — chronic (below 90%). A call, a letter, and an attendance team meeting with the family. Below 80% is additionally flagged as severe.

Everything a tier recommends at Tier 3 is work a person does. The card recommends; it never sends anything by itself.

Approaching the line

Students still above 90% who are a handful of absences from crossing, or whose recent absence rate puts them below it by term end, are flagged separately. The projection deliberately uses the last few weeks rather than the year to date: a student who was fine in September and has missed four of the last ten days is not a 96% student any more.

A student already below the line is never in this list. They are chronic, which is a tier, not a warning.

"Provisional"

A student with only a handful of enrolled days gets a rate but no action. Two absences in a three-day enrollment is 33% and means nothing. The card shows the rate and says it is provisional rather than hiding it.

Why your card may show no rate at all — and how to turn it on

Because your uploads contain absences only. An absentee export is a list of who was out. It carries no record of who was present, and therefore no record of which days each student was enrolled.

That means there is no denominator. We could count absences against days in the school year, but that is the one thing the state definition forbids and it destroys any student who transferred in mid-year. We could count against "days you happened to upload", but then a Friday you skipped makes every student's attendance look better.

So while that is all we have, the card shows absence counts against the number of days you uploaded, labelled as exactly that, and no rate and no tier.

Send your full daily roster instead and every figure above turns on. Export every student with an attendance status column — present, tardy and absent alike — and upload it exactly as you upload the absentee list today. Nothing else about your day changes: only the absences still call, text or email a family, and a student marked Present or Tardy is simply filed (unless you have switched on late-arrival texts). A skipped day costs you nothing, because a date with no row for a student sits in neither side of the fraction, which is the enrolled-days rule rather than a workaround. See Attendance absence alerts for the exact columns and status words we read.

What we hold, and what we never do

One attendance category per student per school day, from a fixed list of eight: present, tardy, half day, excused absence, unexcused absence, out-of-school suspension, not enrolled, no session. Whatever your export writes in its status column is matched to one of those eight when the file is uploaded, and the text itself is thrown away — so if your SIS writes "Absent — Medical" or "Excused – Illness", those words never reach our database, and the reason for an absence, a doctor's note, and any medical detail are not stored, not asked for, and not kept. The excused/unexcused split exists only because your export carries it, and both score identically here.

A very large history shows no rate rather than a partial one. The card reads your attendance records in one pass, and a school big enough to pass roughly 50,000 of them in that window sees a sentence explaining so instead of a percentage. That is deliberate: measuring part of the year would shrink the days-enrolled denominator alongside the absences, so the rate would look ordinary and be wrong. Nothing is deleted and nothing is lost — narrow the date range, or tell us and we will raise the limit for your school.

How far back the numbers go. The card measures every day you have uploaded that we still hold, and prints those dates above the table. We delete attendance records 90 days after they go inactive — the retention your DPA commits to — so on a long school year the window is what remains rather than September to today. Days older than that count on neither side of the fraction. That is why the dates are shown rather than assumed: the number you defend to an authorizer should be the number for the window you actually gave us.