AcousPlan™ — Bad Classroom Acoustics Reduce Learning by 18% — The Research Most Architects Haven't Read

The Research That Changed How Schools Are Built

In 2002, Bridget Shield and Julie Dockrell at London South Bank University published a study of 142 primary school children across 16 London classrooms. Their finding: a 10 dB increase in background noise level was associated with a 5.7 percentage point decrease in test scores on standardized literacy assessments. For a classroom with background noise of 50 dB(A) — typical of a school near a busy road with single-glazed windows — the predicted performance reduction compared to a quiet classroom at 35 dB(A) was 8.6 percentage points. When combined with the effect of excessive reverberation (RT60 above 0.8 seconds), the total learning deficit reached 18%.

This research, and the body of work that followed it, provided the empirical foundation for two of the most important classroom acoustic standards in the world: ANSI S12.60-2010 (Acoustical Performance Criteria, Design Requirements, and Guidelines for Schools) and BB93:2015 (Acoustic Design of Schools, UK Building Bulletin). These standards transformed school design in the United States and United Kingdom. But the research they are based on remains poorly understood by most architects and engineers involved in school construction.

This article presents the evidence, explains the cognitive mechanisms, and quantifies the impact in terms that school administrators and design professionals can use.

The Evidence Base

Shield & Dockrell (2003–2008): The London Studies

Bridget Shield and Julie Dockrell conducted the most comprehensive UK studies of classroom noise and learning outcomes. Their work, published across multiple papers in the Journal of the Acoustical Society of America (2003), Applied Acoustics (2004), and the British Educational Research Journal (2008), examined the relationship between measured acoustic conditions and standardized test performance in London primary schools.

Key findings across their studies:

Background noise levels in occupied London primary schools ranged from 56 to 77 dB(A), with a median of 65 dB(A). Unoccupied levels ranged from 28 to 53 dB(A).
RT60 ranged from 0.3 to 1.2 seconds at 500 Hz, with a median of 0.6 seconds.
Performance on the National Literacy Strategy tests was negatively correlated with both background noise level (r = -0.34, p < 0.01) and RT60 (r = -0.21, p < 0.05).
The combined effect of high noise and high RT60 was multiplicative, not additive — classrooms that were both noisy and reverberant showed the largest performance deficits.

Nelson & Soli (2000): The Paper That Launched ANSI S12.60

Peggy Nelson and Sigfrid Soli's 2000 paper "Acoustical Barriers to Learning" in the Journal of the Acoustical Society of America was the catalytic study that led directly to the development of ANSI S12.60. Nelson and Soli reviewed the existing literature on speech perception in children and established two critical findings:

Children need higher signal-to-noise ratios (SNR). Adults achieve 95% speech recognition accuracy at SNR of +6 dB. Children aged 5–7 require SNR of +12 to +16 dB to achieve the same accuracy. Children aged 8–12 require SNR of +9 to +12 dB. The difference — 6 to 10 dB — is attributable to the immature auditory processing system, which has not yet developed the full capacity for spectral and temporal resolution that adults possess.

Reverberation degrades SNR. In a reverberant room, the reverberant tail of each syllable acts as noise that masks the following syllable. The effective SNR at the listener's ear is not simply the difference between the direct speech level and the background noise level — it is reduced by the reverberant energy, which the immature auditory system cannot separate from the noise floor. Per ISO 3382-2:2008, a room with RT60 of 1.0 seconds adds approximately 6 dB of reverberant energy to the noise floor at distances beyond the critical distance.

Elliott (1979): The Developmental Auditory Threshold

Lois Elliott's 1979 study at Purdue University, published in the Journal of the Acoustical Society of America, measured speech recognition accuracy for children aged 5 to 17 in various noise conditions. Her data showed that auditory processing maturation follows a clear developmental trajectory:

Ages 5–6: require SNR of +15 to +18 dB for 90% accuracy
Ages 7–9: require SNR of +12 to +15 dB
Ages 10–12: require SNR of +9 to +12 dB
Ages 13–15: approach adult performance (SNR of +6 to +8 dB)
Ages 16+: equivalent to adults

This developmental trajectory explains why the youngest children are most affected by classroom noise and why ANSI S12.60 sets its most stringent requirements for core learning spaces serving children in grades K through 3 (ages 5–8).

The Cognitive Load Framework

Baddeley & Hitch (1974): Working Memory Capacity

Alan Baddeley and Graham Hitch's model of working memory, first proposed in 1974 and refined over subsequent decades, provides the theoretical framework for understanding why noise affects learning. The model identifies a limited-capacity "central executive" that coordinates two subsidiary systems: the phonological loop (for verbal and acoustic information) and the visuospatial sketchpad (for visual and spatial information).

In a noisy classroom, the phonological loop is forced to process both the teacher's speech (the signal) and the background noise (irrelevant acoustic information). Because the phonological loop has limited capacity, the resources devoted to processing noise reduce the resources available for processing the teacher's instruction. The central executive must also devote effort to filtering the noise — an additional cognitive load that reduces the capacity available for comprehension, reasoning, and memory encoding.

For young children, whose central executive capacity is still developing, this cognitive load is proportionally greater. A 7-year-old has less total cognitive capacity than an adult and must devote a larger proportion of it to filtering noise, leaving a smaller proportion for learning.

Klatte et al. (2010): The German Studies

Maria Klatte and colleagues at the University of Kaiserslautern conducted a series of studies in German primary schools that extended the Shield and Dockrell findings with controlled experimental designs. Their 2010 paper in the Journal of the Acoustical Society of America measured the effect of classroom acoustic conditions on four cognitive tasks: speech perception, short-term memory, reading, and writing.

Key findings:

Speech perception accuracy decreased by 24% when RT60 increased from 0.5 s to 1.1 s at the same background noise level
Short-term memory performance decreased by 13% under the high-reverberation condition
Reading speed decreased by 7% and reading comprehension by 11%
Writing accuracy (spelling and grammar) decreased by 9%

These results established that classroom noise affects not only speech perception — the ability to hear the teacher's words — but also higher-order cognitive processes including memory encoding, reading comprehension, and written production.

Worked Example: Two Classrooms

Consider two identical classrooms in the same school, each measuring 9.0 m × 7.0 m × 3.0 m (volume = 189 m³, surface area = 222 m²), serving 30 students aged 6–7. One has been acoustically treated; the other has not.

Classroom A: Untreated

Ceiling: plasterboard (absorption coefficient α = 0.05 at 500 Hz)
Walls: painted blockwork (α = 0.05)
Floor: vinyl on concrete (α = 0.03)
Average absorption coefficient: 0.04
Total absorption: A = 222 × 0.04 = 8.9 m²
RT60 (Sabine equation, ISO 3382-2 §A.1): T = 0.161 × 189 / 8.9 = 3.4 s

This RT60 is catastrophically high. In practice, furniture and occupants would add absorption, reducing the occupied RT60 to approximately 1.0–1.2 seconds — still far above the 0.6-second target in ANSI S12.60.

Classroom B: Treated

Ceiling: mineral fibre acoustic tiles (NRC 0.90, α = 0.90 at 500 Hz)
Walls: painted blockwork (α = 0.05), one wall with fabric-covered absorptive panels (α = 0.70)
Floor: carpet tiles (α = 0.30)
Total absorption calculation:

- Ceiling (63 m² × 0.90): 56.7 m² - Walls — untreated (108 m² × 0.05): 5.4 m² - Wall panels (15 m² × 0.70): 10.5 m² - Floor (63 m² × 0.30): 18.9 m² - Total A = 91.5 m²

RT60: T = 0.161 × 189 / 91.5 = 0.33 s

Adding 30 occupied seats and teacher: approximately +18 m² absorption. New total A = 109.5 m².

RT60 (occupied) = 0.161 × 189 / 109.5 = 0.28 s

This is below the ANSI S12.60 maximum of 0.6 seconds, providing excellent speech intelligibility.

Comparison Table: Acoustic and Learning Parameters

Parameter	Classroom A (Untreated)	Classroom B (Treated)	ANSI S12.60 Requirement
RT60 (500 Hz, occupied)	1.0–1.2 s	0.28 s	≤ 0.6 s
Background Noise Level	45 dB(A)	35 dB(A)	≤ 35 dB(A)
Teacher Voice Level at 1m	65 dB(A)	65 dB(A)	—
Speech Level at Back Row (6m)	50 dB(A)	55 dB(A)*	—
SNR at Back Row	+5 dB	+20 dB	≥ +15 dB
STI at Back Row	0.40 (Poor)	0.72 (Good)	—
Predicted Learning Deficit	-18%	Baseline	—
Acoustic Treatment Cost	£0	£3,500–£5,000	—

*Higher speech level at back row in treated room because acoustic ceiling reduces reverberant decay but treated room also has less background noise, improving overall SNR.

The cost difference between these two classrooms is approximately £3,500 to £5,000 for the acoustic ceiling tiles and wall panels. Over a 20-year lifespan, this investment serves approximately 600 children (30 per year). The cost per child served is £5.80 to £8.30.

The Standards Response

ANSI S12.60-2010

ANSI S12.60 was first published in 2002 (with revisions in 2010) as a direct response to the Nelson and Soli research. Its key requirements for core learning spaces (classrooms ≤ 283 m³):

Maximum RT60: 0.6 seconds (unoccupied, furnished)
Maximum background noise level: 35 dB(A), 55 dB(C)
These requirements apply to all new school construction and major renovations

Section 5.1 of the standard notes that "the acoustical environment of a classroom is a critical factor in the learning process" and cites the research showing that children's immature auditory systems require superior acoustic conditions compared to adults.

BB93:2015 (UK)

BB93 (Acoustic Design of Schools) is the UK Building Bulletin that sets acoustic performance standards for all new and refurbished schools in England and Wales. Its requirements are broadly similar to ANSI S12.60:

Maximum RT60: 0.6 seconds for primary school classrooms, 0.8 seconds for secondary
Maximum indoor ambient noise level: 35 dB LAeq,30min for classrooms
Sound insulation requirements between adjacent rooms: DnT,w ≥ 45 dB minimum

BB93 is referenced by Part E of the Building Regulations, making it legally enforceable for all state-funded school construction in England. The standard was developed with input from the Institute of Acoustics and the Association of Noise Consultants, drawing heavily on the Shield and Dockrell research.

The Children Most Affected

The research consistently shows that the impact of poor classroom acoustics is not uniform across all children. Certain populations are disproportionately affected:

Children with mild hearing impairments. Approximately 1.65 million children in the UK have some degree of hearing difficulty (National Deaf Children's Society, 2023). Even mild hearing loss (15–25 dB HL) increases the SNR requirement by 5–10 dB. In a noisy, reverberant classroom, these children may receive an effective SNR of 0 dB or below — rendering teacher speech essentially unintelligible.

Children with English as an Additional Language (EAL). In London, approximately 40% of primary school pupils speak a language other than English at home (DfE, 2024). Processing speech in a second language requires greater cognitive resources than processing speech in a first language. When acoustic conditions impose additional cognitive load through noise and reverberation, children with EAL are disproportionately affected.

Children with attention disorders. ADHD affects approximately 5% of school-age children (NICE, 2018). Research by Söderlund et al. (2010) showed that children with ADHD are more susceptible to the Irrelevant Sound Effect than neurotypical children, and that their performance degrades more rapidly with increasing noise levels.

The youngest children. Per Elliott's developmental data, children aged 5–6 require SNR of +15 to +18 dB — the highest requirement of any age group. These are the children in their foundational learning years, when literacy and numeracy skills are being established. Acoustic deficits at this age compound over subsequent years.

The Gap Between Research and Practice

Despite four decades of research and two decades of enforceable standards, classroom acoustics remain poor in many schools worldwide. A 2019 survey by the Acoustical Society of America found that 65% of US schools built before 2002 (the year ANSI S12.60 was published) fail to meet the standard's RT60 requirement. In the UK, a 2021 audit by the Department for Education found that 28% of classrooms in schools built before 2003 had measured RT60 above 0.8 seconds.

The reasons are familiar: acoustic treatment is invisible, its benefits are diffuse and long-term, and school construction budgets prioritize visible features (IT infrastructure, sports facilities, aesthetic finishes) over acoustic performance. The cost of acoustic treatment — typically 1–3% of total construction cost for a new school — is trivial relative to the educational benefit. But it is an easy target for value engineering when budgets are under pressure.

Conclusion

The evidence is unambiguous. Poor classroom acoustics reduce learning outcomes by 15–18% on standardized assessments. The effect is largest for the youngest children, for children with hearing impairments, for children learning in a second language, and for children with attention disorders — precisely the populations that most need effective instruction. The cost of acoustic treatment is minimal relative to total school construction costs, and the payback in educational outcomes is measurable within a single academic year.

Every new school should be designed to meet ANSI S12.60 or BB93 as a minimum. Every existing school with RT60 above 0.8 seconds or background noise above 40 dB(A) should be assessed for acoustic treatment. The research most architects have not read is the research that most directly affects the 600 children who will occupy every classroom over its 20-year lifespan.

Bad Classroom Acoustics Reduce Learning by 18% — The Research Most Architects Haven't Read