By the end of this topic, you should be able to...
discuss how the average person correlates to the 50th percentile adult and child, and how it is not always appropriate to design for the average person.
Guiding Question
How do designers design mainstream products and environments that are accessible and attractive to the largest possible number of people?
"If you design for the average, you design for no one."— Lieutenant Gilbert Daniels, US Air Force, 1952
At a Glance
When designers reach for anthropometric data — the measured dimensions of the human body — they are typically presented with percentile distributions.
The 50th percentile figure sits at the statistical centre of these distributions, representing the point at which exactly half of the measured population falls above and half below. It is tempting — seductively so — to treat this figure as representing a real, typical human being: the average person. Design for that person, the reasoning goes, and you design for the majority.
This reasoning is fundamentally flawed. The command term for this topic is discuss — a requirement to present a balanced, evidence-based account that explores both the logic behind percentile-based design approaches and the significant, well-documented reasons why designing for the average person consistently fails to serve real users.
Anthropometrics and the Percentile System
Anthropometrics is the scientific discipline concerned with the measurement and statistical analysis of the physical dimensions of the human body. Its application in design provides the quantitative foundation for decisions about the size, reach, clearance and force requirements of products, environments and systems.
Anthropometric data is collected from large population samples and expressed as a statistical distribution — typically approximating a normal (Gaussian) distribution — in which each measured dimension (stature, reach, hip breadth, hand length, etc.) is described by its mean, standard deviation and percentile values.
The Percentile Scale
A percentile is a value below which a given percentage of observations in a dataset fall.
Percentile | Meaning | Application in Design |
5th | 5% of the population is smaller than this value | Used to set minimum clearance dimensions (e.g. minimum seat depth — must accommodate the smallest likely user) |
50th | 50% of the population is smaller than this value; 50% is larger | Represents the statistical midpoint — often (problematically) used as a proxy for "the average person" |
95th | 95% of the population is smaller than this value | Used to set maximum reach distances, maximum force requirements, minimum opening sizes |
The conventional design wisdom — supported by ergonomic standards including ISO 7250 (Basic Human Body Measurements for Technological Design) — holds that products and environments should typically be designed to accommodate users from the 5th to the 95th percentile range, covering approximately 90% of the intended user population.
The 50th percentile figure does not represent a real person — it represents a statistical abstraction derived from averaging a single measured dimension across an entire population sample. This distinction is critically important.
The Myth of the Average Person
Gilbert Daniels and the US Air Force Cockpit Study (1952)
The most consequential investigation into the fallacy of designing for the average person was conducted not by a designer but by a US Air Force researcher named Lieutenant
Gilbert S. Daniels.
In the late 1940s, the US Air Force was experiencing an alarming pattern: skilled, experienced pilots were losing control of aircraft mid-flight — not due to mechanical failure, not due to enemy action, but for reasons that could not be immediately explained. The aircraft performed correctly. The pilots were well-trained. The accidents continued.
The Air Force's initial assumption was that the pilots needed better training. An alternative hypothesis — eventually investigated by Daniels — was that the cockpit itself was the problem: that the cockpit had been designed in the 1920s around the average dimensions of pilots measured at that time, and that the pilot population had changed sufficiently that the cockpit no longer fitted the men flying in it.
The Study
Daniels measured 4,063 US Air Force pilots across 10 physical dimensions considered critical for cockpit fit:
Height
Chest circumference
Sleeve length
Sitting height
Crotch height
Torso circumference
Hip breadth
Waist circumference
Thigh circumference
Wrist circumference
He then calculated the middle 30% range for each dimension and asked a single question: how many of the 4,063 pilots were within the middle 30% range across all 10 dimensions simultaneously?
The Result:
Zero. Not one single pilot, out of 4,063, was average across all ten dimensions.
When the threshold was reduced to just three dimensions, fewer than 3.5% of pilots fell within the average range across all three simultaneously.
The Conclusion:The average pilot — the person for whom the cockpit was designed — did not exist in the real world. The cockpit fit nobody precisely, and for many pilots, critical controls were just out of reach, or the seat position forced a compromised posture that impaired their control authority in high-stress manoeuvres.
The Air Force's response was transformative: rather than finding the average pilot, they redesigned the cockpit to be adjustable — introducing adjustable seats, adjustable pedals, and adjustable control positions. Pilot performance improved immediately.
Design implication: The solution to the failure of average-based design was not better average data — it was adjustability and adaptability. This principle remains one of the foundations of inclusive design.
Why the Average Person Does Not Exist: The Statistics
The mathematical reason the average person does not exist becomes clear when the probability of being average is calculated across multiple dimensions simultaneously.
If we define "average" as falling within the middle 30% range for a single dimension, then the probability of being average on that dimension is 0.30 (30%).
If body dimensions were statistically independent (which they are not perfectly, but for illustration):
P(average on n dimensions)=0.30nP(\text{average on } n \text{ dimensions}) = 0.30^nP(average on n dimensions)=0.30n
Number of Dimensions | Probability of Being Average on All |
1 | 30.0% |
2 | 9.0% |
3 | 2.7% |
4 | 0.81% |
5 | 0.24% |
10 | 0.0000059% |
The probability of any real person being average across all dimensions relevant to a complex product or environment approaches zero as the number of relevant dimensions increases.
The implication is stark: a product designed for the average person is, in quantitative terms, designed for a person who almost certainly does not exist.
The 50th Percentile Adult and Child: Understanding the Data
Adult Anthropometric Variation
Anthropometric databases such as ANSUR II (US Army, 2012), CAESAR (Civilian American and European Surface Anthropometry Resource) and the UK National Sizing Survey (SizeUK, 2001) reveal the extraordinary range of physical variation within and across human populations.
Selected Adult Anthropometric Data (approximate, combined male/female population):
Dimension | 5th Percentile | 50th Percentile | 95th Percentile | Range (5th–95th) |
Standing Height | 155 cm | 170 cm | 186 cm | 31 cm |
Sitting Height | 80 cm | 89 cm | 97 cm | 17 cm |
Hand Length | 17.0 cm | 19.2 cm | 21.4 cm | 4.4 cm |
Shoulder Breadth | 38 cm | 44 cm | 51 cm | 13 cm |
Eye Height (standing) | 143 cm | 158 cm | 174 cm | 31 cm |
Grip Strength | 20 kg | 38 kg | 62 kg | 42 kg |
Sources: ANSUR II (2012); CAESAR (2002); UK National Sizing Survey (2001). Combined male/female values.
Critical observation: A product designed specifically for the 50th percentile standing height of 170 cm places display screens, controls and safety information at heights that are inaccessible or uncomfortable for the 50% of users who are shorter than this value — including the majority of women, older adults and people from populations with lower average stature than the predominantly North American and European populations from which much historical anthropometric data was collected.
The Problem of Population-Specific Data
Anthropometric data is not universal — it reflects the specific population from which it was collected. Historical anthropometric databases used widely in 20th-century design were overwhelmingly drawn from white, male, North American and European military personnel — a narrow demographic slice applied as a universal human template.
The Consequences:
Population | Issue | Example |
Women | Male-biased data used for product design | Standard tool handles, power drills and steering wheels designed around male hand dimensions — poorer grip, higher fatigue and injury risk for women |
Asian populations | Lower average stature and different body proportions than European populations used in standard data | Car seat positions, helmet sizing and safety equipment designed to European/North American norms perform differently — and sometimes unsafely — on smaller-stature users |
Older adults | Strength, flexibility and anthropometric data skewed toward working-age adults | Packaging requiring grip strength beyond that available to many older adults; pharmaceutical labelling in font sizes unreadable without magnification |
Children | Adult data scaled rather than independently measured for children | Critical safety errors in car seat design, playground equipment scaling and furniture ergonomics when adult proportions are assumed |
The 50th Percentile Child: A Separate and Distinct Challenge
The 50th percentile child is not a scaled-down version of the 50th percentile adult. Children's bodies are proportionally different from adults — and those proportions change continuously and non-linearly through development.
Key Proportional Differences: Child vs. Adult:
Characteristic | Child (age 5–10 approx.) | Adult | Design Implication |
Head-to-body ratio | Head approximately 1:5 to 1:6 of body height | Head approximately 1:7.5 of body height | Child's centre of gravity is higher — affects stability and fall dynamics; critical for car seat head restraint design |
Shoulder width relative to height | Narrower relative to height | Broader relative to height | Standard adult harnesses sit incorrectly on children's narrower shoulders |
Leg length relative to trunk | Shorter relative to trunk | Longer relative to trunk | Seat depth and footrest position critical — feet not reaching ground causes posture problems and discomfort |
Bone density and growth plates | Lower bone density; active growth plates at bone ends | Fully ossified | Injury patterns in children differ from adults — growth plate fractures; products must not apply concentrated force at growth plate locations |
Cognitive development | Developing — limited hazard perception, different information processing | Mature | Safety warnings, instructions and product affordances designed for adult cognition may be entirely ineffective for child users |
Case Study — Child Car Seat Design
The Britax Römer KIDFIX (and equivalent booster seat designs) illustrate the consequences of treating child anthropometrics as scaled adult data. Early booster seat designs — designed around a simplified "average child" height/weight specification — failed to account for:
Variation in torso length relative to overall height across children of the same mass
The higher centre of gravity in younger children increasing head excursion in forward crashes
Shoulder belt routing across the neck rather than the shoulder in shorter-torso children
Progressive revision of UNECE Regulation 129 (i-Size) — the international standard governing child restraint systems — introduced height-based rather than weight-based classification following research demonstrating that weight-based categories masked dangerous variation in torso proportions. This is a direct design response to the failure of average-based classification.
When Designing for the Average is Particularly Problematic
Safety-Critical Products
For products where misfit between product and user can result in injury or death, the consequences of average-based design are most severe.
Personal Protective Equipment (PPE)
PPE designed around the average male worker has a well-documented history of inadequate performance for female workers. Research published by the Trades Union Congress (TUC, 2017) found that:
High-visibility vests designed for male torso proportions left female workers with excess fabric that could catch on machinery
Safety helmets sized for average male head dimensions had inadequate retention on smaller female head sizes — increasing the risk of the helmet dislodging in impact scenarios
Anti-vibration gloves designed for average male hand dimensions reduced tactile sensitivity and grip force in workers with smaller hands, increasing tool handling risk
The consequences are not merely discomfort — ill-fitting PPE may offer substantially reduced protection compared to its specification, effectively leaving users with false assurance of safety.
Crash Test Dummies and Vehicle Safety
For decades, the primary crash test dummy used in vehicle safety testing — the Hybrid III — was modelled on a 50th percentile male: 177 cm tall, 77 kg. Female occupant dummies representing the 5th percentile female were introduced to testing protocols later and, critically, were modelled as scaled-down males rather than anatomically accurate female bodies.
Research published by the University of Virginia Center for Applied Biomechanics (2011) found that female vehicle occupants wearing seatbelts were 47% more likely to be seriously injured in comparable crashes than male occupants. The structural differences — female pelvis geometry, lower average muscle mass, different spinal geometry — mean that restraint systems optimised for the 50th percentile male load path may not transfer forces safely through a female body.
Design implication: Designing safety systems around the 50th percentile male does not merely fail to protect the 50th percentile female — it may actively perform differently on her body in ways that increase injury severity.
Medical and Healthcare Products
Medical devices present a particularly high-stakes context for average-based design failures.
Pulse Oximeters and Skin Tone
Pulse oximeters — devices that clip to the finger to measure blood oxygen saturation by transmitting infrared and red light through the tissue — were designed and clinically validated primarily on lighter-skinned patient populations. Research published in the New England Journal of Medicine (2020, 2021) during the COVID-19 pandemic found that pulse oximeters were significantly less accurate in patients with darker skin pigmentation:
Three times more likely to give falsely normal oxygen readings in Black patients compared to white patients
This discrepancy led to delayed identification of hypoxaemia (dangerously low blood oxygen) in clinical settings
Patients with falsely normal oximeter readings received less supplemental oxygen therapy — a direct patient safety consequence
The oximeter had been designed and clinically validated against a patient population that did not represent the full range of human skin tones. The device met its regulatory standards and its certified accuracy — because those standards and that certification were themselves based on insufficiently diverse clinical validation data.
Design implication: The population used to validate a product's performance is as important as the design itself. Designing and testing against a narrow demographic sample produces a product that may function correctly for that sample while failing others.
Digital Interfaces and Cognitive Diversity
Average-based thinking applies not only to physical dimensions but also to cognitive assumptions. Digital interfaces designed around the "average user" — implicitly assumed to be a literate, neurotypical adult with prior digital experience — systematically exclude:
Users with dyslexia (estimated 10% of the population — standard font choices and layout conventions may be significantly harder to read)
Users with low digital literacy or no prior experience with touchscreen interfaces
Older adults whose processing speed, working memory and visual acuity differ from those of younger users
Users with colour vision deficiency (approximately 8% of males, 0.5% of females have some form of colour blindness — interfaces using colour as the sole indicator of status fail these users entirely)
Web Content Accessibility Guidelines (WCAG 2.1) — the international standard for digital accessibility — exists precisely because designing for the average digital user excludes a substantial proportion of the actual user population.
Is There Any Value in Designing for the Average?
A balanced discussion must acknowledge that the 50th percentile figure is not without utility in design. The question is not whether average data has any value — it is whether it is sufficient, and in which contexts its limitations become dangerous.
Where 50th Percentile Data Retains Utility
Context | Justification |
Initial concept proportioning | 50th percentile data provides a useful starting point for initial product geometry before ergonomic optimisation occurs |
Population-level urban planning | Average stride length, average walking speed and average vision parameters inform pedestrian infrastructure design at scale — where the goal is throughput rather than individual accommodation |
Products with inherent adjustability | When the final product will be adjustable across a wide range, the 50th percentile may define the midpoint of the adjustment range — as in office chair design |
Mass market price-sensitive products | Where designing for extreme percentile ranges would dramatically increase cost or manufacturing complexity, the 50th percentile may represent an informed compromise — provided the excluded population is explicitly identified and the risks of exclusion assessed |
Where 50th Percentile Design is Insufficient or Dangerous
Context | Reason |
Safety-critical products (PPE, vehicle restraints, medical devices) | The 50th percentile user is not the at-risk user — the failure modes occur at the extremes of the distribution |
Products used by children | Children are not scaled adults — their proportions, strength, cognition and injury vulnerability differ fundamentally |
Culturally or demographically diverse global markets | 50th percentile data from one population may represent the 80th or 20th percentile of another |
Products requiring physical interaction across a range of users | Handles, controls, screen heights — any dimension requiring physical interaction should be designed to accommodate the 5th to 95th percentile range, not optimised to the 50th |
Inclusive and universal design contexts | By definition, designing for the 50th percentile excludes the majority — particularly the elderly, disabled users and children |
Designing Beyond the Average: Inclusive Design Strategies
The failure of average-based design has driven the development of inclusive design — an approach that treats human diversity not as a complication to be averaged away but as a fundamental design parameter.
The 7 Principles of Universal Design (Mace, 1988)
Developed at North Carolina State University by architect Ronald Mace, these principles provide a framework for designing products and environments that accommodate the widest possible range of users:
Principle | Description | Example |
1. Equitable Use | Useful to people with diverse abilities | Automatic doors — usable by wheelchair users, people carrying loads and ambulant users equally |
2. Flexibility in Use | Accommodates a wide range of preferences and abilities | Scissors designed for left- or right-hand use |
3. Simple and Intuitive Use | Easy to understand regardless of experience, literacy or cognitive ability | Pictogram-based public signage |
4. Perceptible Information | Communicates necessary information effectively regardless of sensory ability | Tactile paving at road crossings; audio signals at pedestrian crossings |
5. Tolerance for Error | Minimises hazards and adverse consequences of accidental or unintended actions | Undo function in digital interfaces; recessed vs. protruding power socket designs |
6. Low Physical Effort | Used efficiently with minimum fatigue | Lever door handles vs. round knobs — accessible to users with limited grip strength |
7. Size and Space for Approach and Use | Appropriate size and space for approach, reach and use regardless of body size or mobility | Accessible toilet design accommodating wheelchair dimensions; height-adjustable work surfaces |
Design for Adjustability
The lesson of Daniels' Air Force study was that when no single fixed dimension can accommodate all users, adjustability is the solution. The principle of designing for adjustability appears across product categories:
Product | Adjustable Parameter | Range Typically Covered |
Car seat | Fore-aft position; height; lumbar support; headrest height | Designed to accommodate 5th percentile female to 95th percentile male |
Office chair | Seat height; armrest height; lumbar position; seat depth | Designed to accommodate 5th to 95th percentile sitting dimensions |
Bicycle | Seat height; handlebar height; stem length | Adjustable to accommodate substantial stature range |
Computer monitor | Height adjustment; tilt; swivel | Designed to accommodate a range of sitting heights and working postures |
Design principle: Where a product interacts directly with the human body — particularly in prolonged-use or safety-critical contexts — adjustability is not a luxury feature. It is an ergonomic and safety requirement.
Conclusion
The 50th percentile figure represents a statistical average of a single measured dimension across a specific measured population — nothing more.
When designers treat it as representing a real, typical person for whom they can design with confidence, they are committing what Daniels' 1952 study demonstrated to be a mathematical impossibility: no real person is average across all the dimensions that matter simultaneously.
The consequences of this error range from inconvenience — a door handle that requires uncomfortable grip force for users with smaller hands — to genuine harm: PPE that leaves female workers less protected than they believe, pulse oximeters that underestimate hypoxaemia in patients with darker skin, vehicle restraint systems that perform differently on female anatomies than the crash tests that certified them.
The most elegant and practical design responses to this challenge are adjustability — demonstrated by the Air Force cockpit redesign and carried forward into every well-designed adjustable workstation today — and inclusive design methodology — which treats the full range of human diversity, including children, older adults, disabled users and diverse ethnic populations, as the design brief, not an afterthought.
In conclusion: the 50th percentile adult or child provides a useful initial design reference point and retains value in limited, non-safety-critical contexts. However, it is never sufficient as the sole basis for design decisions, and in safety-critical applications — PPE, vehicle restraints, medical devices, children's products — reliance on average-based design can produce outcomes that are not merely exclusive but actively dangerous.
Key Vocabulary
Term | Definition | Context |
Anthropometrics | The scientific measurement and statistical analysis of the physical dimensions of the human body | ISO 7250 provides international standards for human body measurements for technological design |
Percentile | A statistical value below which a given percentage of observations fall | The 5th percentile stature means 5% of the measured population is shorter than this value |
50th Percentile | The statistical midpoint of a distribution — 50% above, 50% below | Commonly (problematically) used as a proxy for "the average person" in design |
Normal Distribution | A symmetrical bell-shaped statistical distribution in which most values cluster around the mean | Anthropometric data for most human dimensions approximates a normal distribution |
Inclusive Design | A design approach that considers the full range of human diversity — including age, ability, culture and body size — as a core design parameter | OXO Good Grips kitchen tools — originally designed for users with arthritis, they proved superior for all users |
Universal Design | A design philosophy seeking to create products and environments usable by all people, to the greatest extent possible, without adaptation | The 7 Principles of Universal Design (Mace, 1988) |
Adjustability | A design strategy in which key dimensions of a product can be modified by the user to suit their individual body dimensions | Adjustable car seats accommodating 5th–95th percentile stature range |
5th–95th Percentile Range | The design convention of targeting 90% of the intended user population by accommodating dimensions between the 5th and 95th percentile | Office chair height adjustment range |
ANSUR II | A 2012 US Army anthropometric database measuring 93 dimensions of 6,000 soldiers — a key source of contemporary anthropometric data | Used to establish design envelopes for military equipment and uniform sizing |
WCAG 2.1 | Web Content Accessibility Guidelines — the international standard for digital interface accessibility | Specifies minimum contrast ratios, text sizing and navigation requirements for accessible digital products |
Hybrid III Dummy | The standard crash test dummy used in vehicle safety testing — modelled on the 50th percentile male | Used to certify seatbelt and airbag performance — does not accurately replicate female injury biomechanics |
Growth Plates | Areas of developing cartilage tissue near the ends of long bones in children — more vulnerable to injury than adult ossified bone | PPE and restraint systems must avoid applying concentrated compressive force at growth plate locations in children |
Cognitive Diversity | The range of variation in how people perceive, process and respond to information | Interface design must accommodate dyslexia, varying digital literacy and age-related cognitive changes |
Design for the Extremes | A design strategy that accommodates users at the extremes of the anthropometric distribution — the very small and the very large | A cockpit designed to accommodate the 5th percentile female pilot and the 95th percentile male pilot serves the entire pilot population |
Practice Questions
Question 1
Discuss how the 50th percentile figure is used in design and why it is not always appropriate to design for the average person. [8 marks]
Examiner Hint: The command term discuss demands a balanced, evidence-supported argument. Begin by explaining what the 50th percentile represents and acknowledging the contexts in which it has utility. Then use Daniels' Air Force cockpit study as a pivotal piece of evidence — it is the most compelling demonstration of the mathematical fallacy of the average person. Extend your discussion to at least two specific product design contexts where average-based design produces harm (PPE, medical devices, child products). Conclude by arguing for adjustability and inclusive design as the appropriate design response. Avoid producing a one-sided critique — the examiner rewards nuance.
Mark Band | Descriptor |
1–3 | General statements about average design; limited or no reference to percentile data; one-sided or undeveloped argument; no conclusion |
4–6 | Clear explanation of the 50th percentile; at least one specific product example; acknowledgement of both utility and limitations of average-based design; partial conclusion |
7–8 | Sophisticated, balanced discussion; Daniels' study or equivalent evidence cited; at least two product contexts explored with specific consequences; clear argument for adjustability or inclusive design as the solution; well-reasoned conclusion |
Question 2
Discuss how anthropometric variation between adults and children presents specific challenges for product designers. [8 marks]
Examiner Hint: Do not simply state that children are smaller than adults — this will score in the lowest mark band. The key argument is that children are proportionally different from adults, not merely scaled versions of them. Use specific proportional differences (head-to-body ratio, growth plates, shoulder width relative to height) and connect each difference to a specific design challenge or documented design failure. Reference UNECE Regulation 129 i-Size and the shift from weight-based to height-based child seat classification as evidence that the industry has had to respond to the failure of average-based child product design.
Mark Band | Descriptor |
1–3 | Children described as smaller than adults; general statements; no specific proportional differences or design consequences identified |
4–6 | At least two specific proportional differences between children and adults identified; connected to design implications; limited product reference |
7–8 | Sophisticated discussion of proportional differences with specific design consequences; named product examples or standards; clear argument that children cannot be treated as scaled adults; well-developed conclusion |
Question 3
Outline how the principle of adjustability addresses the limitations of designing for the 50th percentile user. [4 marks]
Examiner Hint: The command term outline requires a clear, structured account rather than extended discussion. Identify the limitation of 50th percentile design (1 mark), explain how adjustability addresses this (1 mark), and support with two specific product examples where adjustability enables a product to accommodate a range of users rather than a single average dimension (2 marks).
Mark Band | Descriptor |
1–2 | Adjustability mentioned with a general explanation; limited or no product reference; limitation of 50th percentile design vaguely stated |
3–4 | Clear identification of the 50th percentile limitation; adjustability clearly explained as the design response; two specific product examples with the adjustable parameter and the range of users accommodated |
Sources
Daniels, G.S. The Average Man? Wright Air Development Center Technical Note 53-7, US Air Force, 1952.
Lamb, M.R. et al. Racial Bias in Pulse Oximetry. New England Journal of Medicine, 2020 and 2021.
Mace, R.L. et al. The Principles of Universal Design. North Carolina State University, 1997.
Mohan, D. et al. Injury Biomechanics and the Female Occupant. University of Virginia Center for Applied Biomechanics, 2011.
Rose, T. The End of Average: How We Succeed in a World That Values Sameness. HarperCollins, 2016.
Trades Union Congress. Risks at Work: The Problem with PPE. TUC, 2017. tuc.org.uk
UNECE. Regulation No.129 — Uniform Provisions Concerning Enhanced Child Restraint Systems. UNECE, 2013 (revised 2023).
United States Army. ANSUR II: Anthropometric Survey of US Army Personnel. Natick Laboratories, 2012.
Web Accessibility Initiative. Web Content Accessibility Guidelines (WCAG) 2.1. W3C, 2018. w3.org/WAI
World Health Organisation. World Report on Ageing and Health. WHO, 2015.
Linking Questions
To what extent is a deep understanding of ergonomics important when engaging with inclusive design? (A1.1)
To what extent can designers remove personal bias when using user-centred research methods? (A2.1)
How can products integrate mechanical systems to improve accessibility and usability in an inclusive design approach? (A3.3, B3.3)
To what extent can the inclusion of electronic systems in products enhance accessibility and usability for all end-users? (A3.4, B3.4)
Which aspects of inclusive design benefit from the designer going beyond usability when designing products? (C1.3)
How important is accessibility and usability when conducting product analysis and evaluation? (C3.1)