A classroom with thirty children and a closed door is one of the most reliably under-ventilated spaces most of us ever spend time in. The occupants cannot feel the problem developing, the teacher has thirty other things to manage, and the only symptom is that the room becomes progressively harder to concentrate in.
Of everything in this category, measuring carbon dioxide in classrooms has the clearest evidence behind it and the lowest cost. It is where we would start in any school.
What the number means
Carbon dioxide is not itself harmful at indoor concentrations. It matters because it is a direct proxy for ventilation: people exhale it continuously, so the level in a room tells you how much of the air has already been breathed. Outdoor air currently sits at roughly 420 parts per million, and that is the floor.
In a classroom the arithmetic is unforgiving. Thirty children in a sealed room of ordinary size will drive the concentration up steadily through a lesson, and by the end of a double period with the door and windows shut it is common to see readings well above 2,000 ppm.
Guidance varies by country, but the broad consensus in school building standards is that classrooms should stay below roughly 1,500 ppm as a daily average with a lower target, often around 1,000 ppm, treated as good practice. Check your own national guidance, because in several countries this is a regulatory requirement for school buildings rather than a suggestion.
On the effects: there is a reasonable body of research linking higher classroom CO2 to reduced attention and poorer performance on cognitive tasks, and separate work linking better ventilation to fewer absences. The effect sizes differ between studies and we would not quote a single dramatic figure as settled. The defensible position is stronger than a contested statistic anyway: a high reading is direct evidence that a room is under-ventilated, and under-ventilation is worth fixing on its own terms.
Measuring it credibly
Insist on NDIR
Buy sensors that measure CO2 directly using non-dispersive infrared. Cheaper devices advertise “eCO2” or “equivalent CO2”, which is estimated from a volatile organic compound reading. In a classroom full of marker pens, hand sanitiser and thirty children, those estimates are worse than useless because they will move for the wrong reasons and undermine trust in the whole exercise.
Placement
Mount at seated head height on an internal wall, away from the door, windows, ventilation grilles and any individual child's breathing zone. Avoid direct sunlight. In a large classroom, one sensor is adequate for the pattern; two tell you whether one end of the room is being ventilated and the other is not, which in rooms with single-sided windows is a common and useful finding.
Calibration
Most affordable sensors use automatic baseline correction, assuming the room reaches outdoor levels at some point each week. In a school that empties every evening and weekend, this works well. It is worth knowing about because it is the reason a sensor that seemed accurate in September may drift in a room that is occupied continuously.
Turning readings into ventilation
A dashboard changes nothing on its own. What changes the air is one of three things.
Mechanical ventilation, where it exists
The best case. Drive the ventilation rate from CO2 rather than a timetable, using two thresholds — boost above roughly 1,000 ppm, relax below 800 — so the fan does not cycle constantly around a single set point. Demand-controlled ventilation of this kind usually costs less to run than a fixed schedule, because it stops over-ventilating empty and half-full rooms.
Windows, which is most schools
Most classrooms are ventilated by opening a window, and the decision sits with a teacher who is teaching. The intervention that works here is not automation but a visible indicator: a traffic-light display showing green, amber and red. It requires no training, it is legible from across the room, and it converts an invisible problem into an obvious prompt.
This is genuinely effective and very cheap, and it has a second benefit — it gives the teacher a reason to open a window that is not simply their own judgement, which matters in a cold classroom in February when children are complaining.
Window actuators
Motorised openers driven by CO2, with rain and low-temperature conditions. Excellent where budget allows, and worth prioritising in rooms with no openable window within reach or where the windows are high-level.
What to do with the evidence
The data is valuable beyond the automation, in three ways schools tend to underuse.
Prioritising works. If four of thirty classrooms are consistently above 2,000 ppm, you know exactly where a ventilation upgrade should go first. That is a far better basis for spending than the age of the windows.
Funding applications. Grant and capital funding bodies respond to measurement. A term of graphs from the worst rooms is materially stronger than a description of stuffiness.
Teaching. Classroom CO2 data is a genuinely good dataset for science and mathematics — real, local, and about the students themselves. We look at this in using the building as a teaching resource.
Cautions worth knowing
Do not use CO2 as an occupancy sensor for attendance. It is a poor proxy and using environmental data to monitor people invites entirely reasonable objections from staff and unions. Measure air quality to improve air quality.
Expect to explain it. A red light in a classroom with no explanation will be read as a fault or, worse, as a judgement on the teacher. Brief staff first, frame it as information about the building rather than about them, and say plainly what the numbers mean.
Watch for gaming. If a display is treated as a performance measure, someone will put the sensor by an open window. Keep it as information, never as an assessment metric.
Do not over-instrument. One sensor per classroom is plenty. Thirty sensors and no ventilation changes is a data collection exercise, not an improvement.
A term-long plan
- Buy six good NDIR sensors, not thirty cheap ones, and place them in the rooms staff complain about most.
- Log a full half-term without changing anything, so you see the daily and weekly shape including Monday mornings and Friday afternoons.
- Add traffic-light displays in those rooms and brief the teachers.
- Re-measure and quantify what the prompt alone achieved. This is often most of the available improvement, for almost no money.
- Use the remaining evidence to prioritise mechanical ventilation or actuators in the rooms that windows cannot fix.
- Then extend one sensor per classroom, with an exceptions dashboard rather than thirty live readouts.
The order matters because step four frequently delivers the majority of the benefit, and it costs almost nothing. Automation is the answer where behaviour cannot reach, not the first thing to buy.