Readers might also be interested in a more recent complementary paper, namely:
Hartley, J. (1998 in press). What does it say? Text design, medical
information and older readers. In D. C. Park, R. W. Morrell & K. Shifrin
(Eds.) Processing of medical information in ageing patients. Mahwah,
N. J.: Erlbaum.
Key words: text design, gerontology; text comprehension,
writing, individual differences, evaluation, written communication, readability
Abstract
This paper addresses the question of whether or not we need to design text differently for older readers. Research with relatively simple text indicates that there is little firm evidence to suggest that making design changes for older readers (except for increasing type-sizes) facilitates their performance. However, research with more complex text settings suggests that changes made to facilitate the understanding of complex text appear to hinder older readers. These findings suggest that older readers should be asked to play a part in the user-testing stage of the development of instructional text.
Life expectancy at birth in the UK has increased by 50% in this century and 4 in every 10 British adults are now over 50. In the United States currently 12% of the population is 65 years of age or older and the number of Americans over the age of 65 years is expected to double to 65 million by 2030. Thus people are living longer and the number of elderly people in the community is getting larger. Consequently there are more older people reading the texts that are currently available, and more texts are being produced especially for older readers.
The research literature on the effects of ageing on learning and memory is voluminous. Following Bond, Coleman and Peace (1993) we can distinguish between three overlapping areas: physiological, cognitive, and social. Physiological research looks at the biology of ageing and its physiological correlates. Most people, for example, experience with age a sharp decline in eyesight. Cognitive research on ageing focuses on changes in memory, learning and judgement. Such effects have implications for work on text design. Social research on ageing examines how, for example, societies expect their older members to function. Studies of ´ageism', for instance, focus on how commonly held attitudes and beliefs about what old people should and should not do determine to a considerable extent what, in fact, they do do.
It is difficult to summarise in a few lines the main findings of these multi-faceted studies of ageing and their implications for text design. The picture is complex, and research in the field is expanding rapidly. Nonetheless, there are two main points that is helpful to bear in mind when thinking about text design for older readers. These are:
1. Working memory capacity (i.e. information held and used in ongoing tasks) declines as people get older.
2. The more difficult the task and the older the person, the more disproportionately difficult that task becomes.
Thus, for example, older people might recall text relatively well but find summarising it disproportionately difficult compared to younger adults (Byrd, 1985).
Meyer, Young and Bartlett (1989) suggest that it is important to consider three overlapping variables in studies of older people learning from text. These are:
o Reader variables - such as verbal ability and prior knowledge.
o Text variables - such as text structure, difficulty, etc.; and
o Task variables - such as remembering, following instructions etc.
Thus one might not expect differences between older and younger readers when the verbal ability of the readers is high, when they have good prior knowledge, when the texts are well presented, and when the tasks are relatively straightforward. Differences, however, might well be expected to emerge with less-able readers, less familiar materials, poorly designed text and more complex tasks.
Improving typographically simple layouts
Generally speaking, the studies on the effects of ageing outlined above suggest that text will be easier for older people to use if their perceptual and memory processing loads are reduced. I would want to suggest that this can be achieved by, for example:
o Using larger type-sizes;
o Using more readable text;
o Using clearer layouts; and
o Clarifying the structure of the text by, for example, using summaries, headings and ´signals'.
In this first section of this paper I shall outline some of the results from studies specifically carried out with older users in these respects. These studies have mainly used what I call typographically simple text - that is, continuous expository prose. I shall discuss studies with typographically complex text in the next section.
Large print
There are undoubtedly many situations in which text is printed in too small a type-size for the elderly or visually impaired reader: one only has to think of the small print in holiday booking forms, financial agreements, prescription labels, etc. Yet despite this well known fact I have been unable to trace many studies of type-size with older readers.
Type-sizes are traditionally measured in points, of which there are 72 to an inch.
Generally speaking type-sizes of less than 8 point are too small to be read easily, and type-sizes of 10-12 point are normally used in academic text of the kind you are reading now. (This text was set in 10 in 12 point in the printed version) Larger type-sizes are used for headings, titles, and display purposes.
Two studies were conducted by Poulton in the 1960s with very small type-sizes, using housewives as participants. These housewives ranged in age from their twenties to their seventies but Poulton did not report the results for different age groupings. However, he did report that the participants found lists of ingredients on food labels were easier to read in 6 point than 4 point type, and that newsprint in 9 point was scanned seven per cent. faster than newsprint in 8 point.
Type-faces also come traditionally in two main kinds - with or without serifs. Serifs are the small finishing strokes at the ends of the letters. Faces without these finishing strokes are called sans serif faces. The following text illustrates the difference:
The quick brown fox jumped over the lazy dog (serif)
The quick brown fox jumped over the lazy dog (sans serif).
There is much - unresolved - debate about how helpful or not serifs are to readers (see Hartley, 1994a).
Shaw (1969) used short passages of meaningless text such as the following:
Noble ways sing other bread. Long stores perform second teeth. Religious fashions compose wide factories. Excellent officials appear usual towns. Sorry coals walk five defences. Numerous flowers speak wrongly.
set in two type-faces (Gill sans serif and Plantin), two weights (medium and bold) and in different pairs of type-sizes according to the participant's visual acuity (e.g. 12 and 14 point, 14 and 16 point, etc.). The participants in her study were 288 visually handicapped adults of varying ages. Shaw did not report specific comparisons between her older and younger participants, but she did make occasional comments in this respect.
Shaw found that an increase in type-size achieved a 16% improvement in reading performance, an increase in weight 9%, and a change from Plantin (a serif face) to Gill sans (a sans serif face) a 4% improvement. Shaw reported that this type-face change was particularly helpful for participants over 50 years of age.
In another study carried out in the 1960s, Prince (1967) compared short texts (210 words long) printed in both 14 on 18 point and 18 on 23 point. The participants were various groups of older readers over 60 years old and a number of visually handicapped children. Prince concluded that the larger type-sizes helped ´a great number of readers'.
Vanderplas and Vanderplas (1980) compared several type-faces in 10, 12, 14, 16 and 18 point settings with proportional line lengths and spacing. The participants were aged between 60 and 83 years of age. The results suggested that more text was read per unit time as the type-sizes increased, but there were interactions between different sizes and different type-faces. In one experiment, where 12 and 14 point type-sizes, line-length, and interline-spacing were all systematically varied, the 14 point type was read more rapidly than 12 point when the inter-line space was 2 or 3 points, but not when the spacing was 4 points. Vanderplas and Vanderplas recommend (with numerous caveats) that 12 or 14 point type-size is suitable for older readers.
This suggestion is in line with many guidelines offered by practitioners. The British Royal National Institute for the Blind (the RNIB), for instance, recommends the use of 14 point type for partially sighted readers (and, incidentally, the use of 12 point for sighted readers, since many still have difficulty with small print). The RNIB also suggests the use of a sans serif type-face for small blocks of instructional text. Table 1 lists some guidelines suggested by the RNIB in this respect.
__________________________________________________________________
o Contrast
There needs to be good contrast between the type and the paper on which it is printed or photocopied. Contrast is affected by paper colour, print colour, type-size and weight.
Black type on white or yellow paper gives a very good contrast. Pale coloured papers provide better contrast than dark ones. Black, or very dark coloured print can be used if the paper is very pale.
The print should not run across photographs or illustrations.
o Type-sizes
14 or 16 point is acceptable when printing for the partially sighted (see the text).
o Type-weights
Light type-faces should be avoided, especially in small sizes. Medium and bold type weights are more appropriate in this context.
o Type-faces
Most type-faces in common use are suitable. Bizarre or indistinct type-faces should be avoided. Numbers need to be printed clearly: blind and partially sighted people can easily misread 3, 5, and 8 in some faces, and even O and 6.
o Capital letters
Avoid presenting long strings of text in capital letters as they are harder to read than lower case ones.
o Line-lengths
These, ideally, should be in the range of 5O-65 characters. Blind and partially sighted people may prefer shorter lines than this. Avoid hyphenating words at the ends of lines. Do not fit text around illustrations.
o Spacing
Keep to regular word spacing: do not stretch or condense lines of type, that is, avoid justified type-settings. Allow the line-spacing to be equivalent to the type-size plus the word-spacing. Use a line-space between paragraphs, and space to show the underlying structure of the text. Additional lines or ´rules' may help keep separate unrelated sections. It is also worth noting that blind and partially- sighted people often need more generous space on forms for hand-written responses as their handwriting tends to be larger than average.
o Paper
Print on glossy paper can be difficult to read. Very thin papers also cause problems because text can show through from the reverse.
_____________________________________________________________________
Experiments on type-sizes and old age are difficult to carry out, and it is never fully clear whether one is investigating the effects of visual impairment or old age, or both, or indeed some other factors. Presumably it is not difficult to set text in 12 (or 14) point in many situations but larger type-sizes lead to less words per given line-length, and this might cause problems with more complex typography. In addition, lengthy texts become bulkier and heavier, factors that might cause problems for the elderly.
Unjustified text
If the reader examines the text of the printed version of this article he or she will find that it has a straight right- and left-hand edge. Technically this is called ´justified' composition. The straight edges are achieved by varying the spaces between the words on each line and sometimes even between the letters within a word. An alternative way of printing text is to use ´unjustified' composition, where the spacing between the words is equal and this leads to a ragged right-hand edge (as in typescript).
The general conclusion reached by many reviewers after examining studies which have compared justified and unjustified text settings is that it makes little difference how the text is set for most readers in most situations and often it is a matter of preference. (See Hartley (1994a) for references.) Difficulties arise for both settings with narrow column widths.
There have been some suggestions in the literature, however, that unjustified text may be more advantageous for older readers and for young children learning to read. Zachrisson (1965), for instance, reported that low-ability adult readers (of unspecified age) read unjustified settings faster than justified ones when the line-lengths contained eight or nine words.
In a more detailed paper Gregory and Poulton (1970) reported the results from a study of justified and unjustified typesettings with participants ranging in age from their mid-twenties to their eighties. These participants were not divided into age groups, but they were classified in terms of their reading ability into three main groups - A, B and C. The results indicated that there were no significant differences in the performances of groups A and B, but that readers in Group C did significantly better with the unjustified settings when the line-lengths were about seven words. Gregory and Poulton replicated these findings in a second study using only Group C readers. In a third study, however, when the line-length was increased to accommodate twelve words, the difference between the settings was no longer significant.
Underlining
Taub (1984) examined the performance of 24 elderly participants (average age 72 years) reading ten passages (average length 700 words) which had selected portions underlined. After reading the passages the participants were required to answer comprehension questions and then to perform a free recall memory test. Taub found that there was no effect of underlining on comprehension, but that underlining helped the recall scores of participants with high vocabulary scores. In a second study (Taub, Sturr and Monty, 1985) the investigators removed the comprehension test, arguing that the results from the free recall test were possibly confounded by this procedure. In this second experiment underlining improved the performance of participants with both high and low vocabulary scores. Taub concluded that underlining can thus help elderly participants to recall more from text.
Improving readability
A number of studies have examined the effects of making text easier to read (see Britton, Gulgoz and Glynn, 1993). A contentious issue in such studies is how to measure the difficulty or readability of a piece of text. There are several measures available, including `readability formulae', each with advantages and limitations (see Hartley, 1994b).
Two studies with older readers on the effects of improving the readability scores of text have taken place in a medical context where, it is generally acknowledged, much material is difficult for the general population. Taub, Baker and Sturr (1986) investigated the effects of re-writing a patient's informed consent form. The original document (283 words long) had a Flesch readability score of 45, which according to the formula, requires 13-18 year old reading skills. A revised version was produced by shortening sentences and simplifying the technical vocabulary. This version had a Flesch score of 76 - i.e. it was now deemed suitable for 12 year old readers.
These documents were read by two groups of readers aged 27-49 years and 50-59 years old. Each group was further subdivided into three groups in terms of their years of education. The participants completed a 10-item multiple choice test on the passage (with the passage freely available to them). The results showed that the older readers had greater difficulties than the younger ones, and that the less well-educated readers had greater difficulties than the better educated ones. However, changing the readability level of the passage had no significant effects.
In a similar study, Walmsley, Scott and Lehrer (1981) assessed how good and poor readers over 60 years of age fared on three versions of four documents which outlined US medical insurance policies. The average length of the four original texts was 1635 words and their average estimated reading age level was 15-16 years. Two revised versions of each document were produced. The first revision simplified the texts by routinely shortening sentences and substituting easier words. The second revision was made by skilled writers aiming to clarify the text for older readers. The average estimated reading age level of the texts after the first revision was 14-15 years and after the second it was 12-13 years.
Each reader was asked to read one version of each of three out of the four passages for five minutes, and then to answer comprehension questions (without the passages being available to them). The results indicated that the good readers did significantly better than the poor readers, but that for only one of the four documents (the longest and the most highly simplified) was comprehension improved by using the skilled writers' version.
These two studies show that difficult text may be a critical problem for elderly readers with low education, but that it is not easy to revise text to make it easier for older people to read.
Questions in text
There has been considerable discussion concerning the effects of questions in text with younger readers (e.g. see Britton et al, 1993). Much debate has focused on the effects of (i) placing questions before or after paragraphs containing pertinent information on the recall of key and incidental material, and (ii) providing different kinds of questions (e.g. factual versus higher-order). However, few elderly readers have taken part in these experiments.
One exception is provided by Woods and Bernard (1987). Their participants were two groups of Canadian university students over 60 years of age who read an 18-paragraph long passage on weather forecasting. Two higher-order (conceptual) questions were inserted for the experimental group following each of two consecutive paragraphs. There were no additional questions for the control group. When the participants had read their passages they were then required to write out as much as they could remember, and these recalls were scored for key and incidental material. The results showed that the experimental group recalled significantly more of the key concepts than did the control group, but that there were no differences between them on their scores for the incidental information. Such findings are typical of many studies carried out with younger participants and they suggest that questions in text can help elderly people to recall key information.
Advance organisers
Advance organisers are devices, somewhat like extended summaries, which set a piece of text in context (Ausubel, 1978). Expository organisers are used when the ideas in the following text are completely new to the reader. Comparative organisers are used when the ideas in the following text are not completely unknown to the reader. Comparative organisers point to the ways in which the old (known) and the new (unknown) material are similar and different.
There has been considerable debate about the effectiveness of such devices (and similar ones, such as outlines and summaries) in helping children and young adults to comprehend text, but there are few studies in this context with older participants. As with questions, two of the key issues discussed on research with organisers and younger participants are, (i) whether or not such organisers help readers to remember better specific or incidental material (or both); and (ii) whether or not they are more useful for readers with different levels of ability.
Such issues were studied with older readers in an experiment by Thompson and Diefenderfer (1986). Their participants (28 young adults aged 18-32 years and 42 older people aged 55-82 years) were asked to read a 1,600 word passage about developing a wildflower garden with, or without an advance organiser. The participants were then asked to complete (i) a 20-item multiple choice test designed to measure their recall and comprehension, and (ii) a test of verbal ability.
The results showed no significant main effects for age, but there was an interaction between verbal ability and experimental condition. Thus it was found that, irrespective of age, participants with low verbal ability performed significantly better if they had the advance organiser than if they did not.
Another smaller study with advance organisers was carried out by Charness, Schumann and Boritz (1992). Sixteen young adults (average age 33 years) and sixteen older ones (average age 60 years) took part in two training sessions on learning how to use a word-processor. Half of the group in each case were presented with an advance organiser, which was designed to help orient the participants to the sessions, and half were not. The organiser consisted of a graphical representation of the keyboard, a listing of several keys used often during the sessions, and general tips on the session formats and contents. The participants completed a 21-item word-processing test at the end of the second session.
The results showed a significant effect for age, with the young group significantly outperforming the old group (70.6% vs 55.4%), but no effect for the advance organiser (67.8% with the advance organiser vs. 63.1% without it). There was no interaction between the age groups and the presence or absence of the organiser.
The mixed results of these two small-scale studies are not unlike the mixed results reported for similar studies of advanced organisers with younger participants (e.g. see Jonassen (1982) for a review).
Signals
´Signals' have been defined as devices used to enhance the structure of text. For example, comparisons can be signalled by phrases such as ´however', ´on the other hand', ´but', and ´in contrast'. Causal relationships can be signalled by words such as ´therefore', ´it follows that', and ´as a result'.
Some researchers have been interested to see if adding such signals to text would help the recall of older learners. Meyer and Rice (1989), for example, described the recall of main points and details from passages with different kinds of signals by groups of young (18-32), middle aged (40-54) and older readers (62 and over). The findings suggested that when the main ideas were signalled, the respondents in all three age groups recalled the main ideas more successfully than when they were not signalled. However, verbal ability was important. The more-able participants of all ages did significantly better than the less- able ones. Curiously enough, when the details were signalled, the older participants with higher verbal ability tended to record more details at the expense of the main ideas, but this was not the case for the younger ones.
Clarifying text structure
Researchers such as myself have long maintained that text can be spaced effectively in order to show its underlying structure (see Hartley, 1994a for illustrations). However, I know of no research which has looked at the effects on older learners of different ways of spacing text in order to display its underlying structure. There are, however, some interesting experiments that do involve manipulating the sequence of information in the text.
Rice, Meyer and Miller (1989) argue that ´more than a decade of research in the area of prose recall has clearly established the "levels effect": readers recall information that is high in the context hierarchy of a text better than information that is located at the subordinate levels'. Rice et al describe one particular experiment in which seventy older people (over 65) read and recalled two passages containing medical information about hypertension and arthritis. Half of the participants read the original passages and the other half read the passages in which the text structure had been revised so that the target ideas identified as important by medical consultants were located at the highest levels of the content structure.
After the participants had read their passages they were then asked to write out what they could remember of them, before completing a 14 item cloze-type test. (This involves filling in missing words: see Hartley, 1994b.) The results showed that the participants reading the revised passages did significantly better on the free recall measures than did the participants reading the original passages, but that this effect was not found with the cloze-type test. It was also found that the participants with the revised passages recalled more main points but less details than did the participants with the original passages. Again, as in other studies, educational level (as measured by years of education and vocabulary knowledge) was important, with better educated participants doing better in both conditions. Rice et al concluded that older adults' memory for important information could be improved by the use of principles of text design developed with younger participants.
Conclusions
Some of the conclusions that can be drawn from the above review are as follows:
(1) There are few studies of any one variable.
(2) Most studies have an older group alone - so direct comparisons with younger participants are not possible.
(3) The results indicate that larger type-sizes are helpful. However, the findings for each of remaining variables are mixed.
(4) Half of the studies show ability effects: high-ability respondents did better than low-ability respondents, irrespective of age.
(5) One third of the studies showed interactions between conditions and ability. (Three of these studies showed that the conditions helped the more-able participants, and three of them showed that the conditions helped the less-able ones.)
What is not revealed in Table 4 is that very few of these studies actually used materials specifically designed to take into account the visual problems of their older readers. Thus, one might argue, many of the older participants in these studies were possibly reading with an additional handicap.
Typographically complex texts and the older reader
So far I have discussed research with texts which have had relatively simple typographic settings. I now turn to studies of older people using more complex materials. Such materials include, for example, bus and train schedules, labels on medicine bottles, food packaging, and government forms. There is considerable evidence that many older adults perform poorly on everyday tasks involving printed materials such as these (Meyer, Marsiske and Willis, 1992).
Income tax forms
In the UK, unlike most other countries, tax is deducted at source from earnings throughout the tax year. It is expected that the right amount will be deducted from pay packets each week or month, and that this will continue throughout the tax year. Thus, at the end of the tax year, it is assumed that the correct amount of tax has been paid and that there is no need for further action. Accordingly, most people in the U.K. do not complete annual tax returns. Checks are, however, made. These are typically done on a three-yearly basis, using a very complex form. This system is different from that of many other countries including, for example, the USA Here most people assess their own income tax liability and file a simpler return by a particular date every year.
James, Lewis and Allison (1987) report that many senior citizens and widows in the U.K. have very complex financial circumstances, and that they find completing income tax forms especially trying. These authors provide a comprehensive report on the difficulties of both ordinary and senior citizens in completing such forms in the U.K., and they present data from several comparison studies. Here I shall restrict my account to one or two of their findings from their work with senior citizens.
In one such study one particular form was redesigned: the introductory letter was made clearer (see Figure 1). Questions on the form were asked in a more consistent manner and the text layout was revised.
As your pension is being increased from 22 November 1982 the allocation of allowances has been changed as shown overleaf. As your net allowances are less than your National Insurance pension, an ordinary code will not deduct sufficient tax so I am giving you a special code. After 22 November your employer or paying officer will use this code to deduct tax from your pay or pension from a previous employment. The effect of the new code is that you will pay more tax. In general the extra tax on the pension increase will be 30%. If your pay or occupational pension is increased the code will also collect tax on that income.
About this Coding Notice
From 22 November the National Insurance
pension is being increased. Since this pension is
taxable, your code number has been changed so
that the right amount of tax is deducted from
your other pension or pay.
This means that the deductions on your other
pension or pay include the tax payable on your
National Insurance pension. If this were not
done, you would have to pay the tax on your
National Insurance pension separately.
Figure 1. Top: an extract from the original text. (Actually this had a line-length of 165mm, and was thus only six lines long.) Bottom: an extract from the revised text used in the experiment reported by James et al., 1987.
In another study with another form the authors assessed the effectiveness of the changes made (i) by interviewing individuals and (ii) by carrying out a postal survey. In the study with individuals, 71% of the respondents reported that they would have difficulty completing the original form, but this figure was reduced to 53% for the revised one. James et al also revised the introductory notes to accompany this new form: now 81% of the respondents found the revised notes helpful, compared with 64% for the original notes.
Diagrams
Lipman and Caplan (1992) compared the ability of older adults (average age 67 years) with that of younger ones (average age 32) to remember a tape slide presentation depicting a particular journey or route. Half of the participants in each group were given a diagram to aid their recall (see figure in the printed version of this paper) but the other half were not. In addition there were three experimental conditions for each group: one third of the participants were instructed to remember the information as a route; one third to remember it as a set of unrelated pictures; and one third to remember the set of slides.
The detailed findings of this particular study were complex but the main findings were that the younger participants recalled more than the older ones, and that, when the diagram had an effect, it improved the performance of the younger participants and worsened the performance of the older ones.
Models
Caplan and Schooler (1990) investigated how useful an analogical model might be to older (average age 45 years) and younger (average age 28 years) participants learning to use a commercial drawing package on a Macintosh computer. The participants were presented with an instructional booklet but, for half of them, this contained an extra introductory page. This page depicted, in a schematic format, the ´tools' that the participants were about to learn to use in the form of an illustrated ´desktop', containing a T-square, a pencil-holder, stencils, paint, scissors, etc. The sentence, ´You can think of the tools you will learn on the Macintosh as tools you might find on your desk' appeared below this illustration.
After fifteen practice trials all the participants were tested on their ability to use the program. They were asked to reproduce each of three pictures in eight different ways. (These ways were defined by providing the participants with sets of particular tools, each one of which had to be used.) The results showed:
that the younger participants did better overall than the older ones; that, without the model, both groups performed equally as well; but
that, with the model, the performance of the older group was significantly worse than that of the younger group.
Prescription information
Morrell, Park and Poon (1990) studied the effects of presenting prescription information in two different formats to an older group (average age 71 years) and a younger one (average age 19 years). The participants studied the labels provided on medicine bottles in one of two formats: verbal labels only and verbal labels with pictorial information. (These labels are illustrated in the printed version of this paper.)
The results indicated that:
overall, the younger participants recalled this information better than the older ones;
the younger participants did better with the illustrated format than they did with the plain text; and
the older participants did worse with the illustrated format than they did with the plain text.
Procedural instructions
Morrell and Park (1993) surveyed the literature on the effectiveness of procedural instructions for assembly tasks. In this survey they noted that, although the findings suggested that realistic, text-relevant illustrations did help people to carry out such tasks, none of the studies involved had included older participants. The authors anticipated that elderly respondents would find such procedural tasks difficult to do because of their reliance on working memory.
Morrell and Park asked a group of young adults (aged 18 to 30) and a group of older adults (aged 60 to 75) to take part in an appropriate investigation. Each participant was provided with an instruction booklet for building two practice figures and nine experimental figures out of Lego blocks and pieces of foamboard. These figures varied in difficulty. There were three different kinds of instructional booklets: one had text only, one had pictures only, and one had both text and pictures. The written instructions were printed in 18 point bold type.
The results indicated that performance was affected by age, instructional format, and level of difficulty. The younger adults consistently out-performed the older ones with each of the three instructional formats. The text and picture together condition reduced the error rates for both of the age groups compared to the other formats. The difficulty of the task had a greater effect on the elderly than it did on the young adults.
Flow charts
Many learners find prose text a barrier, particularly when it is complex. Some people have concluded that continuous text is probably not the best vehicle for expressing complex inter-related rules, and they have turn to alternative modes of expression, particular the algorithm or flow-chart. An algorithm may be defined as an exact prescription or recipe leading to a specific outcome (Michael and Hartley, 1991).
Although previous research has shown that algorithms can often be easier to understand than their conventional prose counterparts, it seems that many adults, especially older ones, might be confused by their unconventional appearance. Indeed, this appears to have been one of the findings reported by Michael and Hartley (1991).
This study looked at the performance of secondary school children (aged 11-12 and 14-15 years) and senior citizens (average age 71 years) solving problems with either a flow chart or a set of contingency statements to help them. (These are illustrated in the printed version of this text.)
In this study, for example, a typical problem might read:
Fred is going to the bank. It will soon be time for the bank to close. Fred is carrying a big bag of ten pound notes. He has to travel by bus because he can't drive a car. The bus journey is 18 miles long.
What colour bus does Fred catch?
After a practice session the participants had to use either the flow chart or the contingency statements to solve eight such problems, and their speed of responding and error-rates were recorded over a series of trials.
The results indicated that the school children solved the problems more quickly with the flow chart than with the contingency statements, and that there were clear practice effects. The participants got faster as the trials progressed. However, there were no significant differences in error-rates. Here the flow chart and the contingency statements were equally effective.
The picture for the senior citizens was rather different. This time there were no significant differences in speed of responding with the two formats, or in the error-rates. However, the speed of responding was much slower for the older participants (double that of the younger ones on the first trial) and the error-rates were much higher. The initial error-rate for the senior citizens was 75% on the first trial and 58% on the second one (compared with 39% and 29% for the 14-15 year old school children).
Conclusions
It is difficult to draw firm conclusions from these six very different studies. Simplifying the tax forms does appear to help the older user. However, some of the other ways of improving the texts outlined in the remaining examples seem to have helped the younger respondents and to have hindered the older ones. The results of these experiments thus suggest that the producers of complex text for the elderly will need to give careful thought to the design of such materials. Morrell and Park (1993) point out that the old people can do the tasks, but that they need additional help. They suggested in the context of their experiment that additional subsets of instructions needed to be prepared for their older participants. One way forward might well be to include elderly participants in any on-going evaluation studies that may be carried out with such texts prior to publication.
References
Ausubel, D P (1978) In defence of advance organisers: a reply to critics Review of Educational Research 48 2 251-258
Bond, J, Coleman, P and Peace, S (eds) (1993) Ageing in Society: An Introduction to Social Gerontology 2nd edition Sage London
Britton, B K, Gulgoz, S and Glynn, S (1993) Impact of good and poor writing on learners: research and theory. IN B K Britton et al (eds) Learning from Textbooks: Theory and Practice Erlbaum Hillsdale New Jersey
Byrd, M (1985) Age differences in the ability to recall and summarize textual information Experimental Aging Research 11 2 87-91
Caplan, L J and Schooler, C. (1990) The effects of analogical training models and age on problem-solving in a new domain Experimental Aging Research 16 3 151-154
Charness, N, Schumann, C E and Boritz, G M (1992) Training older adults in word processing : effects of age, training technique and computer anxiety International Journal of Technology and Aging 5 1 79-106
Gregory, M and Poulton, E C (1970) Even versus uneven right-hand margins and the role of comprehension reading Ergonomics 13 4 427-434
Hartley, J (1994a) Designing Instructional Text, 3rd edition Kogan Page London
Hartley, J (1994b) Is this chapter any use? Methods for evaluating text. In J R Wilson and E N Corlett (eds) Evaluation of Human Work, 2nd edition Taylor & Francis
London
James, S, Lewis, A and Allison, F (1987) The Comprehensibility of Taxation: A Study of Taxation and Communications Avebury Aldershot
Jonassen, D H (1982) Advance organisers in text. In D H Jonassen (ed) The Technology of Text Educational Technology Publications Englewood Cliffs N J
Lipman, P D and Caplan, L J (1992) Adult age differences in memory for routes: effects of instruction and spatial diagram Psychology and Aging 7 3 435-442
Meyer, B J F, Marsiske, M and Willis, S (1992) Text processing variables predict the readability of everyday documents read by older adults. Paper available from the authors, Department of Psychology, Pennsylvania State University, Pennsylvania 16802
Meyer, B J F and Rice, G E (1989) Prose processing in adulthood: the text, the reader and the task. In L W Poon, D C Rubin, B A Wilson (eds) Everyday Cognition in Adulthood and Late Life Cambridge University Press Cambridge
Meyer, B J F, Young, C J and Bartlett, B J (1989) Memory Improved: Reading and Memory Enhancement Across the Life-Span Through Strategic Text Structures Erlbaum Hillsdale New jersey
Michael, D and Hartley, J (1991) Extracting information from flowcharts and contingency statements: the effects of age and practice British Journal of Educational Technology 22 2 84-98
Morrell, R W and Park, D C (1993) The effects of age, illustrations and task variables on the performance of procedural assembly tasks Psychology and Aging 8 3 389-399
Morrell, R W, Park, D C and Poon, L W (1990) Effects of labeling techniques on memory and comprehension of prescription information in young and old adults Journal of Gerontology 45 4 166-172
Poulton, E C (1967) Skimming (scanning) news items printed in 8 point and 9 point letters Ergonomics 10 6 713-716
Poulton, E C (1967) Skimming lists of food ingredients printed in different sizes Journal of Applied Psychology 53 1 55-58
Prince, J H (1967) Printing for the visually handicapped Journal of Typographic Research l l 31-47
Rice, G E, Meyer, B J F and Miller, D C (1989) Using text structure to improve older adults' recall of important medical information Educational Gerontology 15 527-542
Shaw, A (1969) Print for Partial Sight Library Association London
Taub, H A (1984) Underlining prose material for elderly adults Educational Gerontology 10 401-405
Taub, H A, Baker, M T and Sturr, J F (1986) Informed consent for research: effects of readability, patient age and education Journal of the American Geriatrics Society 34 601-606
Taub, H A, Sturr, J F and Monty, R A (1985) The effect of underlining cues upon memory of older adults Experimental Aging Research 11 4 225-226
Thompson, D N and Diefenderfer K (1986) The use of advance organisers with older adults of limited verbal ability. Paper available from the authors, Department Educational Foundations, Georgia State University, Atlanta, Georgia, USA 30303
Vanderplas, J M and Vanderplas, J H (1980) Some factors affecting legibility of printed materials for older adults Perceptual and Motor Skills 50 923-932
Walmsley, S A, Scott, K M and Lehrer, R (1981) Effects of document simplification on the reading comprehension of the elderly Journal of Reading Behavior XIII 3 237-248
Woods, J H and Bernard, R M (1987) Improving older adults' retention of text: a test of an instructional strategy Educational Gerontology 13 107-120
Zachrisson, B (1965) Legibility of Printed Text Almqvist &
Wiksell Stockholm