Higher CO₂ levels (2115 ppm vs 1045 ppm) significantly increased student error rates by 1.65

A cross-over cluster-randomised study across 20 mechanically ventilated classrooms compared two CO₂ conditions: 'worse' averaging 2115 ppm and 'better' averaging 1045 ppm, on separate days each week. Using the d2 concentration test with 2366 test results from 417 third- and fourth-grade students, hierarchical regression found no significant change in overall concentration performance or characters processed, but the total number of errors rose significantly by 1.65 (95% CI 0.42-2.87) under the worse condition.

For those responsible for classroom ventilation and indoor air quality, this suggests that even when CO₂ levels do not visibly slow down how much work students complete, accuracy may still suffer at higher concentrations. This points toward CO₂ control in schools as a factor worth weighing when aiming to reduce careless errors, though the pathway from CO₂ to error-making itself is not established here.

The confidence interval, 0.42 to 2.87, is fairly wide and close to zero at its lower bound, and the same study found no significant effect on overall performance or output volume, so the error result stands somewhat apart from the rest of the data. The design is a controlled crossover trial in real classrooms with children of a specific age group, so it does not generalise beyond that setting without care.