Did you know that a school corridor can seem quieter on a graph even when nobody has changed how loudly they are talking? A group might take one reading next to the entrance and another at the far end of the corridor. The device displays two convincing numbers. The trouble is that the measuring position changed along with whatever the group hoped to study.
A sound-level reading is a measurement of acoustic level at a particular microphone location, with a specified instrument, setting and time interval. To compare noise fairly, the immediate answer is to keep the microphone position, orientation, settings and observation window comparable, or document changes and explain why a direct before-and-after inference is not justified. A different microphone location is not a harmless detail. Distance, nearby surfaces, people and competing sound sources can all affect what the device detects.
A reading belongs to a position, not to a whole corridor
Imagine recording 68 dB(A) near a talking group and 62 dB(A) several metres farther away. Those values might be real readings of the sound arriving at those two points. They do not by themselves show that the group lowered its voice by six decibels between the measurements. The microphone did not remain in the same place. A wall, doorway, ventilation system or approaching footsteps could also have changed the measurement.
Sound is a physical pressure fluctuation, and sound pressure at a listener’s location is affected by the surrounding acoustic environment. The same speaker may produce different levels at a microphone depending on where it is placed. Corridors often include reflecting surfaces and multiple sources, so a simple ‘twice as far means precisely this many decibels’ shortcut is not safe as an explanation for every indoor result.
First define the classroom question
There are at least three legitimate questions hidden inside ‘measure the noise’. One is, ‘What level is measured at the library entrance during this specified time?’ Another is, ‘How does level change with distance from an identified source?’ A third is, ‘Did a noise-reduction intervention change the level at the same point under comparable conditions?’ Each is a different experiment. The microphone positions and conclusions must match the question.
- Location survey: multiple labelled microphone positions are the object of study; do not call position differences a change over time at one point.
- Distance study: hold the source and measuring method steady as far as practical, then vary distance deliberately and record the geometry.
- Before-and-after test: keep the microphone at one fixed marked position and use comparable time windows, settings and environmental conditions.
- Exposure or regulatory assessment: requires appropriate instruments, protocols and trained interpretation; a casual school survey is not a compliance measurement.
A useful diagnosis is to ask the pupil to finish this sentence without guessing: ‘Our reading describes the sound level at … during …’. If they cannot say where the microphone was, their evidence record is incomplete, even if the numerical reading looks impressively precise.
Worked example: a fictional corridor project with a misleading comparison
The following values are invented for instruction; they are not measurements at an actual Singapore school. During interval A, a student holds an app-enabled phone 1 metre from a fixed sound source and notes a displayed 68 dB(A). During interval B, another student moves the phone to 3 metres from the source and records 62 dB(A). Both numbers are shown to whole decibels, and the app, room reflections and source activity were not fully controlled.
The numerical subtraction is 68 − 62 = 6 dB. The arithmetic is straightforward. The inference ‘our new quiet-hours sign reduced the noise by 6 dB’ is not established. The apparent reduction could relate to the new position, changing background activity, differences in timing or several factors at once. A distance change is a rival explanation that the design has not removed.
The class then repairs its design. It marks a single microphone position with removable tape, uses the same device and settings, and makes two comparable observations at that position with the same stated duration. Suppose the display shows 61.0 dB(A) in one interval and 60.5 dB(A) in another. The displayed difference is 0.5 dB. Without the measurement uncertainty, stability of sound sources, observation duration and a suitable acoustic summary, the group should not claim that a small real improvement has been demonstrated. ‘Our readings were close under the revised method’ is a defensible preliminary statement.
Why decibels require careful language
The decibel is a logarithmic way to express a level ratio. A shift from 60 to 65 dB(A) is not simply ‘five more units of noise’ in the way 60 to 65 grams would mean five more grams. A-weighted decibels, written dB(A) or dBA, apply a frequency weighting intended to reflect aspects of human hearing sensitivity. Another weighting setting can give a different number for the same sound.
For a simple educational chart, pupils may report individual observed sound-level values against labelled times and positions. But they should not casually add readings or claim that the ordinary arithmetic average of displayed decibel levels is the same as a properly calculated energy-equivalent average, often labelled LAeq. The relevant device and method must determine the metric. Likewise, a maximum display reading and an interval average answer different questions; do not compare one with the other.
The fixed-position checklist
- Position: choose and mark one suitable point, record microphone height and approximate distance to relevant sources.
- Orientation: use consistent microphone orientation and avoid blocking the microphone with hands, clothing or a case.
- Instrument: note the actual device or meter model, software, calibration status where appropriate, and its limitations.
- Settings: record weighting, response or averaging metric and interval; do not mix instant peaks with interval averages.
- Timing: specify duration and what the corridor was being used for during each interval.
- Environment: note doors, windows, ventilation, rain, traffic, crowds and nearby surfaces that may affect the reading.
- Privacy: collect numerical acoustic measurements only when appropriate and permitted; do not make covert voice recordings.
For a pupil exercise, the protocol should be modest. It does not need to reproduce a professional noise survey to teach honest comparisons. It needs to say which variables are controlled, which changes are being studied and what the instrument actually reports. Safety and school permission outrank curiosity, especially in corridors where movement must not be blocked.
Three mistakes that sound plausible in a class presentation
Mistake one — ‘we measured the same place’ because both readings were indoors
The same building is not the same microphone position. A reading next to a wall and another in the centre of a room may be influenced differently by reflections. A reading beside a fan is not interchangeable with one across the room. Ask for a simple sketch showing microphone location, source and prominent surfaces.
Mistake two — ‘the reading fell, so the sign worked’
An intervention may coincide with quieter readings without causing the change. If the first observation was during a busy recess and the second during a quiet period, time and corridor activity differ. A stronger classroom investigation uses more comparable windows, repeats observations and records what else changed. Even then, a claim about causation should be proportionate to the design.
Mistake three — ‘an app gave a precise number, so we have a certified measurement’
A digital display can show decimal places without guaranteeing that degree of accuracy. Phone microphones and software vary. The US National Institute for Occupational Safety and Health describes validation and limitations of its own specialised sound-level meter app, including the role of calibrated external microphones. A casual phone reading should be described as exploratory when the device and calibration are unsuitable for formal conclusions.
Progressive practice: from identifying the flaw to redesigning the test
Stage 1 — identify the confounder
Show two fictional field notes: ‘Monday 12:20, beside the classroom door, 67 dB(A)’ and ‘Tuesday 12:20, ten metres away, 61 dB(A)’. Ask the pupil to state what has changed besides the day. The expected answer includes position and potentially source activity. Then ask, ‘What can we safely say?’ We can say that two different observations were recorded at two stated locations. We cannot conclude that Monday’s corridor became exactly six decibels quieter on Tuesday.
Stage 2 — draw the setup and choose a fair comparison
On paper, mark a microphone position two metres from a nominated source and label the wall, doorway and fan. Design observations before and after a proposed change using the same location, equipment and metric. The pupil should be able to describe a repeat procedure another student can follow. If the proposed position is unsafe, intrusive or disallowed, redesign it rather than collecting data anyway.
Stage 3 — work out a claim with appropriate caution
Give the pupil two fictional readings made at the same marked point, 64 dB(A) and 60 dB(A), each identified as a momentary display value. The arithmetic difference is four decibels. Ask whether that alone proves a sustained four-decibel reduction. The answer is no: single readings cannot establish stability across a meaningful period. Repeated observations with a consistent interval metric would be more relevant, subject to device limitations.
Stage 4 — independent audit
Hand the learner an anonymised table with columns for time, location, height, device, metric, measured value, conditions and note. Deliberately remove two locations or mix one maximum with several interval averages. Ask them to highlight which rows cannot be compared as a single before-and-after series. Strong performance means rejecting an attractive but unsupported comparison and explaining exactly how to repair it.
What parents and teachers can observe
A parent may ask, ‘What did your sound meter actually measure and where was it?’ rather than celebrating the smallest number. A teacher can use a sketch map and mock data before any live investigation. A tutor can extend the reasoning to temperature measurements, light meters or traffic counts: each number belongs to a method, location and time.
Watch for genuine learning progress. Can pupils label the point, distinguish reading types, separate arithmetic from causation, identify at least one confounder, and write a cautious conclusion? Those behaviours are visible in their notebooks and explanations. Nothing here guarantees a particular assessment score, and a small class trial should not be portrayed as professional acoustic testing.
Limitations and respectful school fieldwork
School spaces contain people. Any measurement plan should avoid identifying or recording private conversations. Work under school rules and staff guidance; keep a clear route for pupils to move safely. Do not use an exploratory student survey to diagnose hearing injury, declare a building legally compliant, or judge an individual student. Specialist workplace or environmental measurements have their own validated equipment, time intervals and procedures.
The proposed fixed-location method controls one important variable, but not all of them. Sound sources can vary unpredictably and reflected sound complicates indoor comparisons. Even a carefully managed exercise may conclude ‘the observations were too variable to support a clear effect’. That is not a failed investigation; it is a successful distinction between evidence and overclaim.
Frequently asked questions
Can we compare readings taken at different points?
Yes, if the question concerns spatial differences and the positions are carefully mapped. No, if you present those different-position values as a direct before-and-after test of one unchanged location.
Can we subtract two sound-level readings?
You can compute their numerical difference in decibels. But explain the metric, measurement conditions and whether such a difference supports the inference you intend. Do not interpret decibels as an ordinary linear quantity.
Is a phone app sufficient for a school project?
It may support an exploratory demonstration with caveats, but accuracy depends on device, microphone, settings and calibration. It is not automatically a substitute for a professional, calibrated sound level meter.
Should we record a video to prove who made the noise?
No such recording is needed for an educational sound-level comparison. Use appropriate, approved non-identifying observations and numerical readings instead. Protect students’ privacy.
The deeper educational aim
The important result is not the loudest or quietest number. It is learning when two numbers can be compared and when the setup has changed too much to justify a conclusion. Pupils who can reject their own tempting story, redraw the measuring setup and try a better protocol are practising independent scientific judgement.
Connected routes and authoritative context
For a related privacy-respecting investigation, read the Core Aim school corridor noise guide. The field-notes measurement guide explains how to label evidence. The eduKate Education learning hub and Clementi diagnosis-first learning reference connect method with independent checking. For primary technical context, see NIOSH information on the Sound Level Meter app and its limitations and NIOSH research on sound-measurement conditions and microphone use. This school exercise does not claim to follow a formal environmental noise standard.