john Gill technology header image

Text Display Preferences on Self-Service Terminals By Visually Disabled People

Dr. Janet H. Silver, formerly of Moorfields Eye Hospital
Dr. John Gill
James S.W. Wolffsohn
November 1994


Self-service terminals are being used by the general public for an increasing range of applications. The most sophisticated terminal in widespread use is the Automated teller Machine (ATM), but ticket selling machines for public transport now offer a bewildering number of choices to the user. To handle this increased number of choices, the terminal often incorporates a visual display which is typically a mono or colour computer screen. Unfortunately a significant number of potential users have problems in reading these displays, but the technology permits the display of various typesizes and foreground / background colours without significantly increasing the cost of the terminal.

In many countries, anti-discrimination legislation is being considered, which will require service providers to make their services accessible by people with disabilities. In the USA, the Americans with Disabilities Act makes specific mention of ATMs. In the UK, it is likely to become an issue in the next general election, with no party wishing to appear to be against people with disabilities.

Conventional plastic credit/debit cards have a small area of storage in the form of a magnetic stripe which is sufficient to store information about your bank account number which, when read in conjunction with personal identification number (PIN) which is keyed into the cash dispenser, acts a security check. An alternative form of storing information, as used in British telecom and Mercury telephones, uses holographic cards in which a laser beam burns away a small area of the card as each unit is used; these single-purpose cards are simple and cheap to make but are limited in their applications.

Smart cards are credit-card sized pieces of plastic into which an integrated circuit chip has been inserted. Over 900 million smart cards have been produced to date and one European manufacturer is planning to produce over 200 million this year. The majority of these cards contain just a memory chip and are used in a similar way to the conventional magnetic stripe or optical cards. However, an increasing number also incorporate a microprocessor which greatly extends the range of possible applications.

A smart card, unlike the current telephone optical cards, can be recharged with credit at a terminal. In the scheme used by British Gas, the terminals are located in post offices and newsagents. In a pilot scheme in Dublin, the same card can be used in the telephone, on public transport and for paying for car parking. So the smart card becomes a multi-service pre-payment card; in a similar way a bank card can become a form of electronic purse.

Because of the size of the memory that can be included in the smart card, it is possible to have forms of verification other than the PIN which has to be stored in some central computer. For instance, at the Barcelona Olympics, entrance to the stadia for employees was gates controlled by the individual's smart card and verified against their finger print.

Another important aspect is that some cards do not have to be inserted in a slot. Sometimes the card has to be held close to the reader but the card can remain inside a wallet. Other systems permit reading the card at a distance of a few metres for applications such as charging for motorways without requiring the vehicles to slow down.

Smart cards offer the possibility that users could indicate their preferred method for interacting with a terminal (eg large characters on the visual display, or input from a keypad instead of a touch-screen). However, for any of these advances to become reality requires there to be international agreement on the features required to be modified to suit a disabled individual, and a standard form of coding these requirements on a smart card.

With advancing years, the ability of the normal eye to adapt to different distances deteriorates. This process, known as presbyopia, continues throughout life but becomes significant within the fifth decade and is complete within the seventh decade, quite apart from normal corrections for short sight (myopia) or long sight (hypermetropia) and astigmatism. In middle life just about everyone needs a different correction for close work as people read within a range of 30-50 cm, with spectacles occasionally being prescribed vocationally against special needs such as computer screens.

Good data on the incidence and prevalence of ‘blindness’ are hard to find; most depend on individuals registering as ‘blind or partially-sighted’, and these data are acknowledged to be a considerable underestimate. Recent studies (Bruce et al, 1990) suggest that one person in sixty of the UK population has a registerable visual disability, i.e. they are unable to recognise from a distance of 6 metres characters that someone with normal vision would be able to recognise from a distance of 18 metres. The definition does not include reading vision but, at this level, people would not have access to ink-print such as newsprint without special devices.

In Europe and other industrialised countries, visual disability can be viewed as a penalty of ageing. Wormald and Evans (1992) point out that demand for services among the elderly is often limited by perceived availability and ignorance of what is possible; their argument is about cataract surgery, but holds true in other contexts too.

The population is getting older, OPCS (Office of Population and Census Studies) data shows the largest increase in population is in the elderly. Some 75% of all visual disability incurs in people over retirement age. A useful American study (Pizzarello 1987) suggests that by the second decade of the 21st century 17% of the population will be over 65 years of age, and the number of people over the age of 85 will double.

The present population of the European Union is about 345 million, of which pehaps close to 40% are presbyopic and probably close to 2%, mostly from that group, have a significant visual disability. Relatively few people, in developed countries, have no useful vision; over 90% of the people registerable as ‘blind’ can better be described as having ‘low vision’.

In 1992 a World Health Organisation (WHO) group developed a working definition of low vision:-

A person with low vision is one who despite appropriate medical and surgical management has between perception of light and 6/18, or a field of less than 10% but is able to use vision for the planning or execution of a task.

WHO have also produced definitions of standard terms, briefly:-

Impairment is any reduction of function.

Disability is when that impairment makes any task difficult or impossible.

Handicap is when the individual actually wishes to perform that task.

Studies have been undertaken into the legibility of ink-print for people with no impairment of sight (Juola et al, 1993), and the visually handicapped (Shaw, 1969). Maximum legibility is achieved if print is well spaced, sans serif, bold and well contrasted with the background. The process of reading is not a smooth continuous flow across the print. Quite apart from the controlled eye movements which tend to go in bursts with hesitations and pauses, the eye itself moves in a series of tiny horizontal movements (saccades), where the photo-receptors within the eye literally recognise the edge between background and the figure (Tinker 1958). Detailed vision actually takes place at the small central area of the retina called the macula (Moses, 1987).

Most ink-print uses a light coloured background with dark print. This arrangement is so familiar that it is not usually appreciated that the eye sees the background rather than the figure. Screens have the great advantage of being able to provide reverse contrast (ie light figures on a dark background) which, physiologically, is more appropriate. Studies have shown that the majority of people using both hard copy or closed-circuit television reading machines prefer this arrangement (Babalola 1961, Mehr et al 1973, Ehrlich 1987).

Public displays present special and different problems. Unlike print which is hand held and can be moved to the appropriate position, terminals are usually at a fixed height. Other problems, such as glare on screen surfaces are outside the scope of this study but are complicating factors.

In the UK and many other countries, the commonest causes for visual impairment remain untreated cataract which causes a greater or lesser obstruction to the light path (Wormald, personal communication). Because cataract is treatable it is often not quoted as a cause of permanent visual loss. Macular degeneration which extinguishes some or all of the macula function is the leading cause of permanent visual impairment. UK authorities such as Professor Bird of the Institute of Ophthalmology now quotes a prevalence rate of over 40% (personal communication). Pizzarello (1987) reviewing American studies, states it is present and reducing vision to 20/40, i.e. 6/12 in 8-11% of those aged 65-74, and 20% at over 75. Diabetic eye disease and glaucoma are also important causes.


The aim of this study was to determine the text display preferences for three groups of visually disabled people.

A personal computer was supplied with a program that simulated a visual display on an ATM. The program permitted the display of a range of print sizes and various foreground / background colour combinations. The character heights of capital letters were 3, 4, 5, 5.5, 7, 9, 12 and 14 mm. The colour combinations available were:- black on white, bright green on black, black on bright green, black on bright yellow, bright yellow on black, strong blue on white, white on strong blue, bright red on black, and black on bright red.

Patients were recruited from routine attendances at the Visual Assessment Department of Moorfields Eye Hospital; full clinical data were always available.

As well as a control group of people with no visual impairment, patients were selected with macular disease, cataracts, and simple presbyopia; in each case, selection was controlled by the diagnosis in the better eye.

The Department has 5 pre-registered optometrists on the staff, and they see patients under the personal supervision of the Head of the Department. All the display assessments were performed by the pre-registered optometrists after the normal clinical assessment. In every case the patient's consent to data being stored on computer was obtained. The purpose of the study was explained to the patient. Patients were excluded if they did not understand the procedure and its purpose, the diagnosis was complex, there was insufficient vision to recognise the largest character in the display, or clinic demands did not allow time.


A series of 16 local ATMs was inspected prior to the study; the mean height to centre of the screen was found to be 1.24 metres.

1. The patient's (standing) ground to eye level height was measured on a wall mounted scale and the patient's eye to screen distance was calculated.

2. The patient sat in front of the computer screen and his position was adjusted to give the appropriate eye to screen distance.

3. The patient was instructed to wear the glasses that (s)he would usually use to view an ATM. If the patient had never used one before, he were asked to wear the glasses usually worn for shopping.

4. The display then showed the smallest sized white characters on a black background, and the size was increased until the patient could identify them accurately.

5. The sequence of colour combinations between background and foreground colours was then compared to the basic screen, and the patient was asked to decide whether the second screen was ‘better’, ‘worse’ or ‘no difference’.

6. The background and foreground colours of the next screen were varied to determine the patient's preference (background colour was always determined first).

7. The size of the characters was then confirmed.

The procedure was abandoned after ten minutes if no conclusion was reached.


A total of 116 subjects were recruited, 69 [59%] were female, 47 [41%] male. There were 16 normal people who agreed to act as controls, they were mainly hospital clerical staff with no connection with the department and attending for screening under the Display Screen legislation. The average age of these was 24.3 (19 - 33) years, all had 6/9 or better vision for distance and N5 for near with a distance vision correction if needed with both eyes open.

Therefore 100 people with impaired vision were considered. The average age was 73 (33 - 86) years; 21 were presbyopic, 25 had cataract, and 54 had macular disease.

The controls preferred the 4mm characters, and 6/16 selected white on blue, 3/16 selected green on black.

The presbyopes preferred 7mm characters, and 12/21 selected white on black.

The cataract group showed a preference for 4mm characters; the most popular colour combinatioms were white on blue (10/25), and white on black (9/25).

The macular disease group showed a wide choice of size, but only 12/54 selected less than 7mm, with an even distribution across the larger sizes. The colour preference was for white on black (30/54) with 8/54 preferring yellow on black and 8/54 white on blue.


Among the visually disabled subjects, the colour combination of choice (51%) was white characters on a black background, presumably due to maximal colour contrast and reduced disability /discomfort glare effects.

White characters on a blue background was considered to be as good as white on black by the cataract subjects, and yellow on black was also relatively well accepted by the macular degeneration group, but not by the cataract group. A possible explanation for this phenomenon is that senile cataract tends to act as a yellow/brown filter.

The control group could discriminate a smaller font size than cataract subjects, who can discriminate a smaller font size than macular degeneration subjects. Presbyopes' optimal discrimination covers a wide range of character sizes and this may reflect the appropriateness of the spectacles worn during the test, since these may or may not have focussed the display accurately.

The preferences of the control group may have included an aesthetic as well as discrimination characteristics as an important criterion for their preferred colour combination choice, but the colour does not appear to reduce legibility to a great extent. However for the others, the colour and contrast did have considerable significance, 88% of the visually disabled subjects selected one of the four alternatives where there are light figures against a strong dark background.

Of the 100 visually disabled subjects, 88 never used an ATM. It could be that more of them would use an ATM if the displays had dark background with light figures, and there was a choice of character size.


The authors are grateful for the help they received from Colin Peleuvé who provided the data on public displays, Shaheen Asgar, Shaila Chaudry, Michelle Hennelly, and Nitesh Pancholi (who with James Wolffsohn conducted the assessments), the Moorfields consultant ophthalmologists who allowed us access to their patients, AT&T Global Information Solutions who provided the hardware, and Lloyds Bank Charitable Trust for their financial support of this project.


Babalola J (1961) The Facilitation of Reading by Partially-Sighted Persons. British Journal of Physiological Optics, Vol 18, pp 220-234.

Bird AC (1994) Personal Communication.

Bruce I W, McKennell A C and Walker E C (1991) Blind and Partially-Sighted Adults in Britain: The RNIB Survey. HMSO, London, ISBN 0 11 701479 6.

Ehrlich D (1987) A Comparative Study in the Use of Closed Circuit Television Reading Machines and Optical Aids by Patients with Retinitis Pigmentosa and Maculopathy. Ophthalmology and Physiological Optics, Vol 7, No 3, pp 293-302.

Juola J F, Tiritoglu A and Pleunis J M (1993) Reading Text Presented on a Small Display. IPO Annual Progress Report, No 28. pp 37-41.

Mehr E B, Frost A B and Apple L (1973) Experience with Closed Circuit Television in the Blind Rehabilitation Programme of the Veterans Administration. American Journal of Optometry, pp 458-469.

Moses R A (1987) Adler's Physiology of the Eye - Clinical Application. 8th Ed, Butterworth, pp 89-188.

Pizzarello L D (1987) The Dimensions of the Problem of Eye Disease Among the Elderly. American Journal of Ophthalmology. Vol 94:9 pp1191-1198

Shaw A (1969) Print for Partial Sight. The Library Association, ISBN 0 85365 3119.

Tinker M A (1958) Recent Studies of Eye Movements in Reading. Psychological Bulletin, Vol 55, No 4.

Wormald R and Evans J (1992) The Demand for Ophthalmic Services in the UK. British Medical Journal, Vol 305. 1502

Wormald R (1994) Personal Communication.


John Gill Technology Limited Footer
John Gill Technology Limited Footer