The Effects of Providing Headings in Braille Text
J.Hartley; M.J.Tobin; M.Trueman
Abstract: Twenty-four blind subjects are studied to assess whether headings in braille text facilitates readers' comprehension and, especially, recall of the material. Results show that there was a favorable disposition to the headings when the participants were questioned, but upon being tested for recall no significant effect was found.
In an earlier article, Hartley (1986) discussed how current research on text design might contribute to the setting of braille text. Hartley distinguished between three areas of research. These were:
o research on the language of text (ways of making texts easier to read);
o research on the layout of text (ways of using typography and space to convey text structure); and
o research on 'access structures' (ways of using devices such as headings and numbering systems to aid search and retrieval).
In this paper we concentrate on the use of headings. Previous research with sighted readers has examined the effects of headings on search and retrieval tasks and the effects of headings on readers' immediate and delayed recall. Generally speaking, it has been shown that headings do aid search and retrieval, but it has not been shown so clearly that headings always aid recall. Nonetheless, in a metaanalysis of some twenty studies of the effects of headings on recall, Hartley and Jonassen (1985) concluded that a cautious estimate of the effect-size for the presence of headings was 0.27. This means that for each experimental group there was, on average, a modest effect due to the presence of headings about the size of one-quarter of a standard deviation above the mean of the control group.
In addition to examining search, retrieval and recall, other studies have examined the effects of the frequency of headings, the position of headings (embedded in the text or placed in the margin), the length of headings (e.g., one-or two-word headings versus phrases), and the style of headings (e.g., headings written in the form of questions versus statements). In reviewing this literature, Hartley and Trueman (1985) concluded that it is useful to have several headings, and that phrases are more helpful than single words. However, it does not appear to matter much whether headings are embedded in the text or positioned in the margin, or whether headings are written in the form of questions or in the form of statements. Finally, we may note that participants in experiments on headings usually express favorable remarks about the provision of headings in text (e.g., Klare, Shuford, & Nichols. 1958; Charrow, & Redish, 1980).
As noted, the findings described above are restricted to studies with sighted readers. In this paper, we outline the results that we obtained with blind readers when we investigated the effects of inserting headings into braille text. Such an enquiry, as far as we can tell from Lorimer's (1982) annotated bibliography, has not been carried out before.
Method
Materials
Two texts were used in this enquiry. Each was based on a current British university prospectus and they explained what students could expect if they studied psychology at the University of Keele or political science at the University of York. Each text was modified from its original version so that the two texts were equivalent in difficulty and length. Each was printed in braille in two versions-one with and one without headings. The passages were both two-and-a-half thermoformed pages long, and each contained a total of approximately forty lines. Apart from the headings (which were inserted with one line-space above and no extra line-space below), the texts of the two versions matched each other line for line and page for page.
Both texts were divided into five paragraphs, and in the headings conditions four headings were inserted appropriately above the main paragraphs. The headings were written in the form of questions. For example, the headings for the Keele passage read:
o What happens in the first year of the main course?
o How does this change in the second year?
o What happens in the final year?
o How does the subsidiary course compare with the main course?
Short-answer recall tests were devised for each passage. There were ten questions on the Keele passage and nine on the York one. Both tests were marked out of ten (by allowing an additional mark for a correct answer to one complex question on the York passage).
Participants
Twenty-four blind people (16 men and 8 women) took part in this experiment. The median age of the participants was 59 years and their ages ranged from 17 to 80 years. The backgrounds of the participants were varied-they included retired professional people, college students and manual workers.
Procedure
There were four men and two women in each condition. Each participant was asked to read two passages-one with and one without headings-in one of four counter-balanced orders as follows:
o Keele passage with headings, York passage without headings;
o Keele passage without headings, York passage with headings;
o York passage with headings, Keele passage without headings; and
o York passage without headings, Keele passage with headings.
This procedure allowed us to assess the effects of the headings in the texts while keeping counter-balanced any effects caused by passage order, practice fatigue.
Each participant was asked to read the first passage twice, at his or her own speed, and then to answer the test questions given orally. These answers were recorded (manually) by one of our research assistants. The procedure was then repeated with the second passage. After completing both of their passages and tests, the participants were asked if they had noticed any differences between the two passages. If they did not comment on the headings in their reply, this particular difference was pointed out to them and they were asked to comment on how useful-in general they found headings to be in braille text.
Results
The answers to the test questions were marked independently by two markers who were unaware of whether the participants had read a particular passage with or without headings. The few discrepancies that appeared between the two marks given were resolved by discussion between the markers.
Table 1 shows the mean scores obtained with their standard deviations for men, women and overall. A two-way related analysis of variance (conditions x sex) showed that (i) the presence of headings in the two texts had no significant effect (F = 2.28, df 1, p = 0.15; (ii) there was no significant difference between the performance of the men and the women (F = 3.48, df 1, p = 0.18); and (iii) there was no significant interaction between these variables (F=0.19).
The responses to the questions about whether or not the participants had perceived anything different about the two passages were also analyzed. The results obtained (which are shown in Table 2) indicated that almost none of the participants answered this question in terms of headings. (Most commented on the differences in procedures between the two universities, or on how text A was harder to read than text B or vice-versa.) Nonetheless, when asked what they thought about the usefulness of headings in general, most of the participants made highly positive remarks, and this finding was statistically significant (X2=9.25, p<.0l, two-tailed test).
Discussion
The results of this experiment are reasonably clear. The headings did not significantly affect the participants' recall performance but, on the whole, the participants were favorably disposed toward headings. The difference between recall scores of the headings and the noheadings conditions, while not being significant, was in the expected direction, and indeed this difference was just over one-quarter of a standard deviation-as found in the meta-analysis reported by Hartley and Jonassen (1985). Perhaps, if we had been able to test more participants or to use a more homogeneous group of participants, then our findings might have been more positive.
In this study, there was a remarkable range of scores obtained on the test questions (from 2 to 10 for each version), and several participants commented on how the texts had little interest for them. Nonetheless, as noted above, most participants spoke in favor of headings. Some typical comments were:
Having headings does help, of course, when you are scanning back.
Headings prepare you for what is to come.
Headings might have been more useful in a longer piece.
Concluding remarks
The findings of this experiment, while not conclusive, do suggest that many blind readers perceive advantages in headings. It might be useful, therefore, to consider inserting more of them into braille text. If such headings were numbered (with, say, the numbers 'outdented' into the margin) and if they were more clearly delineated by space (e.g., two lines above and one below) or by space and typographic coding (e.g., two lines of space above, heading in italic, no extra space below), then it is highly likely that such procedures would help blind people perceive the overall structure of the text, and move more easily to and fro within it. In addition, headings may help readers to remember better the content of the text-although, of course, this has not been shown in this enquiry.
References
Charrow, V.R., & Redish, J.C. (1980). A study of standardized headings for warranties. Technical Report No. 6 Document Design Project, 1055 Thomas Jefferson St. N.W., Washington, DC 20077.
Hartley, J. (1986). Current research on text design and its implications for the setting of braille. (Paper submitted to the Journal of Visual Impairment & Blindness).
Hartley, J. & Jonassen, D. (1985). The role of headings in printed and electronic text. In D. Jonassen (ed.) The Technology of Text, Vol. 2. Englewood Cliffs, NJ: Educational Technology Publication.
Hartley, J., Trueman, M. (1985). A research strategy for text designers: The role of headings. Instructional Science, 14 (2), 99-155.
Klare, G.R., Shuford, E.H., & Nichols, W.H. (1958). The relation of format organization to learning. Educational Research Bulletin, 37, 39-45.
Lorimer, P. (1982). The braille code and the teaching of braille reading and writing: An annotated bibliography. Part II of J. Lorimer, M J. Tobin, J. Gill, & J. L. Douce. A Study of Braille Contractions. London: Royal National Institute for the Blind.
James Hartley, Ph. D., head of the Department of Psychology at Keele University, Staffordshire, UK; Michael Tobin, Ph. D., director of the Research Centre for the Education of the Visually Handicapped, the University of Birmingham, UK; Mark Trueman, M.A., lecturer in psychology at the Polytechnic, Wolverhampton, UK.
The authors are grateful to Mrs. E.W Hill, Mrs. J.R. WHittaker and Mrs. Alice Slaney for their assistance
Table 1. The mean scores (and standard deviations) for men and for women reading text with and without headings
Headings No Headings
Men 6.9 6.3
(N =16) s.d. 2.3 2.1
Women 5.0 4.7
(N=8) s.d. 2.4 2.0
Total 6.3 5.8
(N = 14) s.d. 2.5 2.2
Table 2. The number of participants who (a) noticed headings and (b) liked headings in general
(a) noticed (b) liked headings
Yes No Yes Neutral No
2 22 15 4 5