Background information

This page provides some background information about the SeeWord project and some information about dyslexia in general. For more dyslexia information, try some of the links on the useful links page.

Contents

Back to main page

Background

History

Previous studies carried out in the Department of Applied Computing at the University of Dundee showed that the ability of dyslexic people to read text on a computer screen can be dramatically improved by changing the way in which the text is presented on the screen. The changes can include the text size, text colour, text font, line spacing, and text spacing. These characteristics can be changed in most word processors, but this involves navigating through long menus. The optimal settings for reading or producing text on the screen are different to the characteristics that the user would want when the text is printed on paper. The user would therefore have to change the text layout back to a black-and-white scheme for printing - a long and tedious process. Appendix 1 discusses the problems experienced by people with dyslexia in more detail and suggests ways in which technology can be of assistance.

Aims and Objectives

The present study proposes an alternative to the universally adopted WYSIWYG  method of word processing – namely, a word processing environment where the user can switch between an easily configurable, specialised “dyslexic view” of the text, and the WYSIWYG view for formatting and printing. Since many computer users are already familiar with Microsoft Word for Windows™, the software developed in the project was developed as an optional extension to this software.
The purpose of the project is to develop and evaluate a high quality software suite, which enables dyslexic computer users to configure Microsoft Word for Windows™ to suit their individual reading preferences. The software would also use a speech synthesiser so that text could be spoken aloud on request.

Acknowledgements

The SeeWord research project was funded by The Viscount Nuffield Auxiliary Fund. The researchers would like to thank Ms Alison MacAffer of the Grove Academy in Broughty Ferry who made the user evaluations possible. We would also like to thank the students from Grove Academy and from the University of Dundee who participated in the user evaluations. In particular, Mrs Di Shaw, Special Needs Advisor at Dundee University provided useful background information on dyslexia and facilitated the contact with dyslexic students.

Back to top
 

Methods

The software development cycle

The exact nature of dyslexia is still debated, and not two dyslexics face exactly the same problems. The development of SeeWord therefore took a pragmatic approach where the focus was to investigate ways in which reading and production of text would be improved through making visual changes to the environment in which the dyslexic is working.
Previous work in the Department of Applied Computing had identified ways in which word processors could be configured to suit dyslexic users (Lodge, 1996). This work provided the basis for the current research.
A user-centred design and development methodology was used to ensure that the software was developed according to the users’ requirements. Users provided formative feedback on prototypes during the development process. The prototypes were presented to different users who evaluated it and provided feedback and suggestions. The feedback was evaluated in order to identify common problems and suggestions, which could be implemented in the following version.
The following describes the development process:
A prototype was developed in Microsoft Word™ version 7 in “word basic,” the built-in macro language. This version did not offer speech synthesis or multiple users, but the user could change the appearance of the text and apply these settings to a new text.

Prototypes

This first prototype was evaluated with dyslexic students from the University. This evaluation showed areas where the software could be improved, and generated new ideas for future work in the field.

The second prototype was developed in C++ and was implemented as a plug-in to Microsoft Word™. It had the facility to catch “reversal” letters and colour them, and it used a speech synthesiser to read a selected passage of text. The users’ profiles were stored in a file, which made it possible (but cumbersome) to allow more than one user to use the system with individual settings.

The evaluation of the second prototype was carried out at Grove Academy with dyslexic pupils. This evaluation showed that the method of implementation was unstable (the communication between the SeeWord program and Microsoft Word™ would sometimes fail).

The final version of the SeeWord program is a set of two programs. One is a stand-alone word processor with the option of setting colours, text size, and fonts. It also includes a speech synthesiser, which can say individual letters, words, or sentences as they are typed. The second program is a pop-up dialogue box in Microsoft Word, which can be activated by the press of a button. The user can use this box to change the text size, colour, font, background colour, and text spacing.

Final version

The final version was published on the World Wide Web in order to gather feedback from a large number of users. The final SeeWord system offers the user an easy way to configure the features of the text. Some features that can be changed are: A speech synthesiser is used for reading out individual letters, words, or passages of the text
Since computers in schools are used by more than one person, the software also has the option to store individual profiles for each user.

Back to top

Evaluation of the software

Evaluation of the usability and functionality of the software was crucial to this project and was carried out in co-operation with the Dundee University Disability Support Centre and Dundee City Council Education Department.
Two methods were used in the evaluation with the users. One was a “think aloud” procedure, where the user were given a specific task in the system, and asked to say whatever came to their mind while using the system. This evaluation method is very useful in the initial stages of the design process, as it revealed usability problems and showed how the system is actually used. The second test was a reading test with standard texts (from a set of texts often used by special needs teachers to assess pupils reading abilities). The users were asked to read the texts one at a time from the screen. For each level of difficulty of text, two texts were read; one with the user’s preferred settings, and one with the normal black and white text. The user’s reading was recorded on tape, and analysed later. Details of the development cycle and evaluations can be found in How we developed the software.

Back to top

Publications

There is a potentially huge user base for a specialised dyslexic word processing aid. Therefore, the software developed in this project has been made available on the World Wide Web, so that users can download it, and the University in return can get valuable feedback. Users can download a packed file, which contains all the necessary files for an installation program. The installation program also contains the option for removing the software again from the user’s machine. See Download and installation for instructions.

The research has resulted in the following publications:

Back to top

Conclusions

User-centred software development

The implementation of a user-centred design process ensured that the final software met the users’ needs, and that possible usability issues were solved in the process. Some of the features of the first prototype were left out in later versions because they were not used by the users (e.g. the underlined text).

Improvements of the users' reading abilities

The evaluations with school pupils showed that there was a significant difference in the number of errors that they made when reading a text off the computer screen with and without SeeWord. All the users found a combination of settings (colours and text features) that they preferred over the black-on-white text. They all described the text with their own settings as “clearer” or “easier on the eyes.”

Difficulties in text production

It was not possible to conclude objectively whether or not users produce text with fewer errors when using SeeWord compared to typing without SeeWord. However, the subjective comments from users indicate that it is easier to read text on the screen using SeeWord, and that it would be less stressful to work with the computer for a longer time, resulting in a better quality of text. Using speech synthesiser capability required some practice, so a user more acquainted with the software would probably use it more often.

Perspectives for non-dyslexic users

As in many other projects concerned with people with special needs, this research showed that an environment that is friendly to disabled  - or in this case dyslexic people – can also be beneficial to everyone else. The idea that the WYSIWYG method of producing text is not necessarily optimal has the implication that it may be a good idea for anyone working with text to produce it first in a format that suits their visual demands, and then do the print formatting at a later stage.
The use of speech synthesis is becoming increasingly widespread, and many applications will include this feature in the future.

Back to top

Future work

Improvements to the SeeWord software

The SeeWord software as it is distributed now relies on Microsoft Word™ for spell checking. Some of the dyslexic users have reported problems with the spell checking in a word processor, since it is mostly tuned to catching typing errors, and not the special problems encountered by dyslexics. Implementation of a spell checker designed at Dundee University for dyslexic users (**reference**) may be useful as an addition to SeeWord.
Currently, the software is tied to English language. An obvious improvement is to make it language-independent. This will require a speech synthesiser that can use other languages than English. Such systems exist, but neither of them is available free, as is Microsoft’s English “Text to Speech Engine.”
The current system imposes some restrictions on the way the user can produce text: the text has to be produced first in a “dyslexic view” and then layout in a “normal view.” It is not possible to switch freely between the two views. This could be improved in a later version.

Large-scale evaluation

With the conclusion of the SeeWord project, the University of Dundee now has software, which can be used in a larger scale evaluation of the concept that changing the visual features of text improves readability. The availability of the software on the World Wide Web means that it can reach a large user base. The University needs to follow up on the feedback that results from this project.

Investigation into the cognitive processes involved in dyslexia

One question, which was not answered in the current study, is: Dyslexic users can read text with fewer errors in the “dyslexic view,” but is the comprehension of the text better? The availability of the SeeWord software makes it possible to investigate reading comprehension.

Collecting experimental results from international associates

Dyslexia is of course not specific to the English speaking areas. A study that involved dyslexic people in other countries would make it clearer which language problems that are specifically English, and if there are problems that are “generically” dyslexic. International co-operation in this field would also generate useful knowledge about localisation in a highly language-specific software like SeeWord.

Back to top

More about dyslexia

The British Dyslexia Association defines dyslexia as follows:

A dyslexic person is one who has a specific language disability affecting spelling, reading, and other language skills characterised by a discrepancy between his or her mental potential and his or her education level despite classroom instruction and despite the absence of any primary emotional trouble or adverse environmental condition.

Dyslexia

Despite the fact that hardly two dyslexics are the same, there are some common problems encountered by many dyslexics. Some of those are: -

1. Number and letter recognition.

One of the fundamental problems faced by dyslexics is the recognition of alphanumeric symbols. This is often seen when letters which are similar in shape, such as ‘n’ and ‘h’, ‘f’ and ‘t’, are confused. The problem is exacerbated with the introduction of uppercase letters. In addition, many dyslexic adults, who are capable of reading printed letters, have difficulty in reading cursive writing.

2. Letter reversals.

Many dyslexics are prone to reversing letters, which results in a particular letter being interpreted as another letter. Examples of these characters would be ‘b’, ‘d’, ‘p’ and ‘q’. This problem can result in poor word recognition, with words that contain reversal characters being substituted for other words such as ‘bad’ for ‘dad’.

3. Word recognition.

As well as the substitution effect caused by letter reversals, words that have similar outline shape can be substituted by dyslexics. A typical example of this problem is the words ‘either’ and ‘enter’. Both words have the same start and finishing characters and this, allied with the fact that both words also have the same overall shape, make them candidates for being substituted for each other when they occur in the text.

4. Number, letter and word recollection

Even once the ability to recognise numbers and letters is mastered, it can still prove difficult for a dyslexic individual to recall the actual form and shape of a character. Many dyslexics have so much difficulty recalling upper and lower case characters that they continue to print later in life. Similarly, poor visual memory means that dyslexics have little ability to distinguish whether or not a word ‘looks right’ or not.

5. Spelling problems

Due to the problems discussed in (1) - (4) above, dyslexics can have great difficulty with spelling and many dyslexics have very poor spelling. Much of the spelling of dyslexics appears to reflect a phonic strategy with words like ‘of’ and ‘all’ being spelt ‘ov’ and ‘olh’.

6. Punctuation recognition

As with characters, dyslexics appear to have difficulty recognising punctuation.

7. Fixation problems

Another problem, which is found in many dyslexics, is their lack of ability to scan text without losing their place. Many find it difficult to move from the end of one line to the beginning of the next and find themselves ‘getting lost’ in the text.

8. Word additions and omissions.

Dyslexics may add or remove words from a passage of text, apparently at random. This is manifested by words being duplicated, extra words being added, or word order being reversed.

9. Poor comprehension

With the variety of errors caused by the factors described above, a dyslexic person may read a totally different passage of text from the one, which is actually in front of them. Dyslexics can thus display poor comprehension skills due to text, which they perceive being significantly different from the actual text.
It is interesting to note that a minor version of this effect can also occur in normal readers, who can completely miss typographic errors. In this case however reading comprehension is usually improved.

Technological and other aids

In an attempt to alleviate some of the problems discussed above, dyslexics, particularly within the education system, are encouraged to use computers for text manipulation. The use of a computer keyboard has the potential to alleviate the problems of character recollection. If a person is capable of recognising the characters as they are on the keyboard, the need for the individual to recall the shape of the required character is no longer necessary. In addition the position of each character on the keyboard can be used to help recognition: as with touch-typing, to print an ‘a’ the user only has to remember to move to the left middle of the keyboard without having to recall the shape or form of an ‘a’. The keyboard, however, only really helps with the recollection of characters, not the recognition of them once they are on the screen.

Another obvious facility already offered by word processors, which can be of benefit to a dyslexic, is the spell checking facilities. As discussed, spelling is an area, which can cause great difficulty with dyslexics. However merely highlighting an incorrect word and offering a replacement may not be enough, since one of the other problems, which some dyslexics face is an inability to tell if a word ‘looks right’. They would therefore be unable to tell what was wrong with the badly spelt word and hence would be unable to correct the errors. There is, however, strong evidence to suggest that the use of lexical and spelling aids can greatly assist with spelling problems exhibited by dyslexics. Many word processors now have grammar checking facilities which can be effective in picking out duplicate words and may also be of use in discovering word additions and omissions but, as with the spell checker, merely picking out these errors may not be enough because the dyslexic may be incapable of selecting the appropriate corrections.

It is also well known that some dyslexics are sensitive to colour, and that lighting conditions can be very important to their ability to read text. Some dyslexics either wear tinted glasses, or read text through a coloured acetate screen, which is placed over the text. This suggests that manipulation of colour might be a valid configuration parameter. Many dyslexics also report interference from peripheral vision, indicating that anything that can be done to reduce screen “clutter” outside the main screen window, such as making the document “page” fill the whole screen may be of benefit.

Back to top

How we developed the software

Preliminary Word Basic based program

The first version of the SeeWord software was a set of dialogue boxes that could be called up within Microsoft Word at the press of a button. The dialogues were divided in three groups; one for spacing of the text, one for text font, size, and other features, and the last for colours of text and background.
Developing a customised dialogue box in Word Basic turned out to be a laborious task, and much time was spent on minor tasks concerned with the display of the box. The software showed some of the possibilities of a specialised word processing environment, and formed the base of the discussions with users.

First evaluation with users

For the evaluation of the first version of the software, an informal “think aloud” method was used. First, the user was introduced to the system and was told how to navigate it. Then the user was given specific tasks, such as “Increase the size of the text,” and “Make the lines of text shorter.” While the user was performing the tasks, he or she was making comments about the functionality of the software, the usability, and inconsistencies that occurred. After the tasks were finished, the user gave their general comments about the software and about the idea of a dyslexic word processor. Comments in this evaluation showed that many users had not previously realised that changing the colours of the text and background could improve the usability of a word processor, either by making the text “more clear” or reducing the perceived stress of reading off a computer screen. Another finding was that it would be necessary to implement a way to change the text (the document in the word processor) back to black-on-white, before printing it out.

Prototype

The second version of the SeeWord software was written partly using Borland C++ Builder and partly in Word Basic. The reason was that C++ gives far more freedom in designing dialogues, while some Word Basic code was still necessary to perform the reformatting of the text. Several methods were investigated to enable the communication between the Word Basic program and the C++ program. Many of these methods proved somewhat unstable, causing the software to malfunction at unpredictable times.
The functionality of the prototype was a slight extension of the first version, with the added functions of colouring of reversal letters, and a speech synthesiser. The speech required the user to highlight the text that should be read. The user interface was improved by adding graphical elements to the dialogue where the user had to choose the various settings for the text. A short text was used as a “preview” so the changes could be studied before they were applied to whatever text the user was working on.

Second evaluation

The second evaluation had the purposes of gathering more information about the usability of the software, and at the same time investigating whether the configurable “dyslexic” view made a measurable difference in users’ ability to read and produce text in the word processor.
Twelve dyslexic pupils from a secondary school volunteered to take part in the experiment. First, each pupil was allowed some time to try the SeeWord system and find a combination of settings (colours, font, text size, etc.) that he or she found easiest to read. After a couple of days the pupils asked to confirm that the settings they had chosen were in fact the settings they would like to work with. A reading test was then carried out where the pupil read texts off the screen.
The test was made both with the pupils preferred settings, and with the settings that the word processor uses by default. The test was repeated with increasingly difficult texts up to a maximum of six texts, or until the pupil’s capability seemed to be reached. The texts that were used were chosen at random from a set of standard texts used by teachers to assess pupils reading abilities. Thus, the texts on each level should be equivalent in difficulty.
The reading test was recorded on tape and later analysed by counting the number of errors. The errors were mispronunciations, substitutions, refusals, additions, omissions, and reversals. For each of the six levels of texts, and for each pupil, the number of errors, as a percentage of the number of words in the text was calculated. The average number of errors per word was calculated per level of difficulty and compared between the reading tests with the pupils’ preferred settings of colours etc, and the standard settings of the word processor.
The results of this test showed a significant improvement in the pupils’ error rate in reading texts off the computer screen. The experiment also showed that when the users set their own preferences, they all found a combination of colours that they preferred over the black on white. There seemed to be two groups of preferences; one with “muddy” low contrast colour combinations, and one with a high-contrast colour combination.
Some functions in the software were not found useful by any of the users, e.g. underlined text, which suggested that these functions could be left out in future versions. A further discussion of the results can be found in the article referenced in Publications.

First release

The release version of SeeWord was developed as a two-part system. One part is a stand-alone word processor with special options for controlling the text appearance, with multiple user facilities, speech synthesis, and the option to integrate a spell checker at a later stage. The text produced in this word processor is automatically transferred to Microsoft Word at the press of a button. The other part is a pop-up dialogue box that is shown in Word and is used for changing the appearance of the text.
The reason for this split is that it proved technically very difficult to integrate the speech synthesiser in the way it was required (being able to read each letter as it was typed), and to catch the reversal letters and colour them “on the fly”. Building the specialised word processor, which integrates with Word by sending the finished text (without colour and other formatting information) to Word, solved those problems.

Back to top

Back to main page


Updated: 29 September 1998
Peter N Andreasen