In August 2026 MARS changed how some screening results are classified for research reporting. A new formula is only half the job: every school year already on the dashboards has to be recalculated with it.
The dashboards read precomputed analytics, so we recomputed them for every past screening. Then we compared the totals for the 2025-26 school year with counts the research team had prepared on its own. They matched exactly.
We also keep one check that must always read zero. Any other value means a stale result has crept back into the data.