School at Home Is a Data Problem Before It Is a Teaching Problem

A family pulls their daughter out of school in March, planning to finish sixth grade at home and re-enroll her in the fall. In September the district registrar asks for an account of the spring: days of instruction, subjects covered, evidence of progress. What the parents have is a stack of finished workbooks, a library card with a long history, and a phone full of undated photos. They spend a weekend reconstructing five months from memory.

Versions of this happen constantly, and not because anyone was careless. Home education in the United States is governed by state law, and the states disagree about almost everything, including what must be written down. But set the legal question aside and simply inventory what a full year of home learning has to keep track of. Attendance, counted as days in some states and hours in others. The subjects taught and the materials used to teach them. Samples of the child's work from early and late in the year, so that progress is visible rather than asserted. Reading lists. Test scores or evaluation results where those are required. For a high school student, course content and credits that will eventually need to become a transcript. It is not a large dataset. It is a small one with a multi-year retention need and, in most households, no one formally responsible for it.

The shapes state law takes

There is no national rule, only a spectrum. At one end sit states such as Texas and Illinois, where, according to state-by-state summaries from the Home School Legal Defense Association, families are required to keep essentially no records. In the middle are notice-and-attendance states, where the obligation is a filed letter and a running count of days. At the far end are the portfolio states. Florida's home education statute requires parents to keep a log of educational activities along with samples of the student's writings and worksheets, preserved for two years and open to district review on notice. Pennsylvania requires a portfolio and an annual evaluation of it. New York's regulations go further still: an instruction plan filed at the start of the year, four quarterly updates to the district, an annual assessment, and instructional targets of 900 hours in the elementary grades and 990 in grades seven through twelve.

The trap is reading that spectrum as a statement about effort. A family in a no-records state ends up needing most of the same artifacts as a family in Pennsylvania, because the records a family needs, catalogued in guides from Numa School, are set less by statute than by the requests that arrive later. The difference is that nobody says so in advance.

Where the data gets lost

Almost never in a disaster. The losses are ordinary. The finished workbook, which doubles as the only evidence of a year of math, goes into the recycling in June. Project photos live on a phone, undated and unlabeled by subject. An attendance log starts cleanly in September and stops in February. The co-op instructor who kept the only gradebook for a lab science moves away over the summer. A curriculum gets swapped in November, so the table of contents that was meant to serve as a course description covers half the year.

Two structural problems sit underneath these. The first is fragmentation: a paper planner, a reading log in one app, photos in another, test results in an envelope from the testing company. Each piece survives on its own. The year they jointly describe does not. Anyone who runs systems for a living will recognize the condition: data that was technically captured but cannot be produced on demand. The second problem is timing. Records are cheap to capture in the moment and expensive to reconstruct later. A parent who spends ten minutes on Friday noting what actually happened that week has done something a full weekend in September cannot replicate, because by then the information no longer exists anywhere except, partially, in memory.

The moment of request

What makes the paper trail matter is that requests arrive long after the fact, on someone else's form, with a deadline. Re-enrollment is the common case. A district deciding grade placement can ask what was studied and to what standard, and a family with dated work samples has a short conversation while a family without them may face placement testing. Families who move mid-year from a light-regulation state to a heavier one find the new state's expectations apply on arrival, including expectations about the months already past. For older students the requests grow more formal. Most colleges accept a parent-issued transcript, but admissions offices commonly ask for the course descriptions and grading criteria behind it, and the NCAA eligibility process asks home-educated athletes for documentation of coursework that must reach back through all four high school years.

None of these requesters care how the year felt. They ask narrow questions about dates, subjects, and evidence, and ask them of the only adult who was in the room.

A plan is not a record

There is one more failure mode, and it is subtle because it looks like diligence. Most home-educating parents plan. August produces a schedule, a curriculum list, goals for each subject. The plan is written in optimism, and by March the real year has diverged from it: a math program abandoned, a history unit that ran eight weeks long, a semester of biology that turned into marine science because the child fell in love with an aquarium. When a request comes, families often submit the plan, because it is the document they have. But a registrar or evaluator is asking what happened, not what was intended. Those are different artifacts, and only one can be written after the fact.

Families who come through requests easily tend to do three unglamorous things. They date material as it is produced. They keep everything in one place. And they compile the year in June, while it is fresh, rather than in whatever future September someone comes asking.

The durable part and the fragile part

The instinct is to treat teaching as the hard part and record-keeping as clerical overhead. The teaching is the hard part. But a year of learning is remarkably durable where it matters most: the reading, the projects, and the problem sets stay with the child who did them, whether or not anyone wrote anything down. The written account of that year is different. It is the only version of the year that can be lost, and it is the only version any institution will ever ask to see.