"Publicly binding" conflates cite-against-the-city with a duty to answer the trend. The interview shows the duty-to-answer was already dead while the count was still in the report — the number rose in public. Withdrawal kills reviewability. It does not remove a constraint that was working, so "would have most strongly constrained the institution" overstates the published years.

The recommended press repair is a piece asking why the assessment-report format changed during the growth. That question is the specimen: the journalist asked it, and Nurmi answered with a format change plus internal tracking. Repeating it is not a new procedural object.

"The report now covers different items" can be a local choice to drop an embarrassing count or an inherited template change that no longer has a slot for it. Timing-with-growth is a coincidence test and does not tell those apart.

Even if the template was handed down, the leftover is whether Espoo could still publish the count in another public slot. If it could, dropping it from the report is a citation-channel choice. If the report was the only public slot, the template change is the whole mechanism and Espoo is the site, not the mover.

PPE used on a small scale in remaining demanding-support special classes is evidence about the leftover segregated rooms, not a before/after on ordinary classes after special-needs classes were closed. The same line can cut the other way: the remaining special classes are the high-hazard tail.

A workplace-safety complaint about PPE in special-class rooms investigates teacher injury risk in those rooms. It does not restore the public annual count and does not fill the contribution table. It is a real route for a different object than the structural finding.

The PPE admission can support a last-resort-control claim about those rooms (elimination and isolation already failed) without supporting inclusion as the cause of the citywide trend. Those are two jobs stacked on one line.

The teachers' union also has a paid conclusion (restore special-needs classes). Applying the same cause-name test yields four paid explanations, not three talismans plus a privileged practitioner candidate. Section III's convergent displacement needs the protected object to be independently the thing being avoided; the union saying so does not do that.

Computing the special-class row first can survive as a measurement order (Espoo has placement dates and incident rates) without treating the union as a truth-teller. What cannot survive on union speech alone is the claim that the three public explanations are protecting the apparatus from inclusion-policy.

Section I bounds the premature-baby share (very preterm around 0.8 percent; medical profile pointing away from aggression). Section V then reports a uniform "unknown" in every share field as the table-level finding. If the table is what computation would produce, computation already produced a small bound for one row. That is not a uniform unknown.

The table as printed is a picture of the public argument (nobody published a share), not of the computation. The 0.8 percent bound is the essay filling one share field. "What computation would produce" is the wrong caption for a table whose result is empty shares. Computation is what would fill the blanks.

The cheapest-row story needs school-level implementation timelines that actually differ. If they do not, Espoo's before/after is one city-year in which welfare cuts, cohort effects, and the placement-model change move together, and "Espoo's responsibility" to compute the inclusion contribution is too small.

Hypothetical: if school-level dates are basically the same, the comparison that would tell inclusion apart is a national panel of municipalities that kept small-group placements versus ones that did not. Espoo does not own that panel. Its remaining job shrinks to publishing the raw incident series.