Standardized tests are not going away this year. The question worth asking is smaller and more honest: can a school run assessment that teaches thinking, while still meeting the reporting obligations it already has?
This post sketches how gate-oriented assessment sits alongside standardized testing, and where, over time, it can carry some of the load those tests were asked to carry.
What a gate is, in one paragraph
A gate is a short, defined checkpoint in a unit where a learner shows a specific competence before moving on. It is not a quiz score. It is a record: the learner made a prediction, checked it against evidence, and named what would change their mind. The gate captures the reasoning, not just the answer. This framing draws on active-inference pedagogy in the tradition of Parr, Pezzulo, and Friston (2022) (Class E), where perception and action are treated as prediction under uncertainty.
Why not just use the test scores
Standardized tests were built to compare cohorts at scale. They do that job with known trade-offs: narrow item formats, a single moment in time, and a strong pull toward what is easy to score. None of that is scandalous. It is a design choice. The trade-off is that a score alone does not tell a teacher what the learner was thinking, or where a well-formed misconception lives. Two learners can arrive at the same score by different routes, and a gate captures the route.
What a gate record actually contains
In our pilot integrations, a gate record typically holds five fields (Class C, from the rubric config we ship with the workshop template):
- The learner's prediction before evidence.
- The evidence they consulted, with sources.
- The update they made, and why.
- A falsifier: what observation would have changed the conclusion.
- An evidence-class tag on the strongest claim (A empirical-in-session, B code or inspection, C configuration or integration, E expert citation, F falsifier present, U unverified).
None of those five require a proprietary platform. They can be captured on paper, in a shared doc, or in whatever tool the school already uses.
Where this reduces reliance on standardized testing
Three places, cautiously.
First, formative decisions inside a term. If a teacher already has gate records across a unit, mid-term standardized checks add little information the teacher does not already have. The gates are the signal.
Second, portfolio-style reporting to families. A gate log reads as a story of reasoning across a semester. Families tend to find it more legible than a percentile band.
Third, transitions between grades or schools. A short set of gate records travels with the learner and tells the next teacher where the reasoning is strong and where it wobbles. A single score does not do that.
We are not claiming gates replace external accountability testing where a state or district requires it. They sit next to it. Over years, if the gate record proves itself, the conversation about how much standardized testing a school needs can be reopened on better footing.
What this asks of teachers
Less than a new grading system, more than a worksheet change. The lift is in three habits: naming a prediction before evidence, tagging the strongest claim with an evidence class, and writing a one-line falsifier. Teachers who try this for a unit usually report the falsifier line is the hardest and the most useful.
What this asks of leaders
Two things. Time for teachers to design a small number of gates well, rather than many gates poorly. And a public commitment that gate records are not fed into a punitive comparison across classrooms. Gates are for learning, not for ranking teachers.
Where to go next
- Curriculum: gates that teach thinking
- A rubric that rewards falsifiability
- How we measure a pilot honestly
- The workshop, if you want to design gates with us
AI-authorship fence: this post was drafted with LLM assistance, reviewed and edited by a human author, and published under human editorial responsibility. No claim is made that an LLM authored the reasoning. The framing is a working hypothesis on the attainable path toward General Natural Intelligence, natural not artificial, tested in the open.