27.2 Text Layout, Structure, and Access

So far I have discussed matters of typography that I believe help readers to find their way around a text and to grasp its underlying structure. I now turn to discuss those devices that are specifically used by writers and designers to further help readers in this respect. I have called this section "structure and access" because these devices-perhaps unwittingly-both clarify the structure of the text and also help the readers gain access to it. Readers do not simply read instructional text from the beginning to the end: they skim, search, reread, etc. Devices that help them do this include tides, contents pages, summaries, outlines, headings, and subheadings, and numbering systems. In addition, authors use such linguistic devices as "signals" to help readers follow the organization of their arguments (Waller, 1979).

27.2.1 Titles

Titles aim to describe the content of a text in the fewest words possible; but these are often supplemented with a subtitle. Such succinct descriptions help to focus attention and expectations, and studies have shown that titles affect the readers' perception and interpretation of ambiguous text (e.g., Bransford, 1979). However, it is to be hoped that instructional text will not be ambiguous! So one would hardly expect titles to have much effect on the comprehension of instructional text, although they may aid later recall of what the text was about. Unfortunately, I know of no research on typographic variables connected with the setting of titles (e.g., type sizes, typefaces, weights, etc.) and none on the more interesting problems of using different title formats (e.g., statements, questions, quotations).

27.2.2 Summaries

Summaries in text can have different positions and roles. Beginning summaries tell the readers what the text is about; they help the readers decide whether or not they want to read it, and they help the readers who do read it to organize their subsequent reading. Interim summaries summarize the argument so far, and indicate what is to come. End summaries list or review the main points made, and thus aid the recall of important points in the text. End summaries can use the more technical vocabulary introduced in the text; beginning summaries might not. Research on the effectiveness of author-provided summaries has been reported by Hartley and Trueman (1982), Reder and Anderson, (1980), and Sherrard (1988). Research on the effectiveness of reader-generated summaries has been reported by, among others, Annis (1985) and Kirby and Pedwell (1991).

Summaries can be typeset in many different ways: in medium, bold, or italic, in large or small type, boxed in, etc. There is no research to my knowledge on the effect of such typographic variables in this context, although there is some indication that readers dislike journal abstracts set in a smaller . type size than the main body of the text (Hartley, 1994c).

27.2.3 Outlines

Outlines can have much the same function as a summary, although it is likely that outlines depict the structure of the text more clearly. Often outlines are provided in a graphic form, sometimes in the form of a tree diagram or flowchart (Guri-Rozenblit, 1989). Such displays facilitate understanding and recall in at least two ways. Firstly, readers can see the organizing structure of the text all at once. Secondly, readers can follow different routes within this structure - comparing and contrasting different parts-in the order of their choice: The argument is no longer linear, and it is not obscured by lengthy paragraphs of text. Research reviewing the effectiveness of outlines has been reported on and summarized by Glover and Krug (1988); Glynn, Britton, and Muth (1985); Hall, Dansereau, and Skaggs (1992); and Lambiotte and Dansereau (1992).

27.2.5 Headings

Headings may be written in the form of questions, statements, or (like here) with one- or two-word labels. Headings may be placed in the margin or in the body of the text.

In a series of experiments with 12- to 14-year-old school children, Mark Trueman and I investigated the role of different kinds of heading (questions versus statements) and their position (marginal versus embedded). We concluded that headings significantly aided search, recall, and retrieval, but that the position and the kinds of heading that we used had no significant effects with the texts that we employed (Hartley & Trueman, 1985). More studies still need to be carried out on factors such as the:

• Nature of the text (technical versus semiliterary)
• Frequency of headings
• Typographic denotation of headings of different levels (primary, secondary, tertiary)

Additional research indicating the effectiveness of headings has been provided by Spyridakis and Standal (1987), Townsend et al. (1990), and Wilhite (1989)

27.2.6 Questions

Questions may be interspersed in the text itself, or presented in a list at the end of a chapter to provide material for exercises. There is some indication that readers ignore questions given at the ends of chapters, so it might be more appropriate to consider how best they can be embedded in the text. It appears that factual questions, placed in a passage before paragraphs of relevant material, often lead to specific learning, whereas similar questions placed in the passage after the relevant content will sometimes lead to more general learning as well (see Allington & Weber, 1993; Hamaker, 1986; Hamilton, 1985). The level of difficulty of these questions, too, may be important (Allington & Weber, 1993; Armbruster & Ostertag, 1993; see also 30.6, 31.2.13).

Some of our earlier research suggested that headings in the form of questions were particularly suitable for less-able readers, but our more recent (better-designed) studies failed to confirm this (see Hartley & Trueman, 1985). None the less, it might be important to consider headings in this form for certain texts.

27.2.7 Sequencing

There has been little research on the sequencing of sentences or paragraphs within instructional text, apart from work with programmed instruction (see Chapter 20). Some of this work suggested that violations in natural sequences provided little difficulty for most readers. However, just what is a "natural sequence"? Posner and Strike (1978) contrast 17 different ways to show that sequencing is not a simple matter, and Van Patten et al. (1986) develop these issues further.

There are some situations, however, where we might all agree that the sequence used is unhelpful. Take, for example, this odd sequence of instructions I once found for using an electric razor:

1. To gain access to the heads for cleaning, press the button on the side of the appliance (see Fig. 4).
2. To remove the razor from its packaging....

Readers find it easier to follow a sequence in which the events match the temporal order in which they occur. Compare "Before the machine is switched on, the lid must be closed and the powder placed within its compartment" with "The powder must be placed in its compartment and the lid closed before the machine is switched on."

27.2.8 Sequencing Lists

It is fairly common in instructional writing to find sentences containing embedded lists of items such as this: under 10%

Five devices that aid the reader are (i) skeleton outlines for each chapter, (ii) headings in the text, (iii) an end summary, (iv) a glossary for new technical terms, and (v) a comprehensive subject and author index.

However, research suggests that readers prefer text that has such lists or numbered sequences spaced out and separated, rather than ran on in continuous prose. The above example would be better thus:

Five devices that aid the reader are:

• Skeleton outlines for each chapter
• Headings in the text
• An end summary
• A glossary for new technical terms
• A comprehensive subject and author index

27.2.9 Numbers in Text

Numbers are often used to signal the structure of a piece of text. Lorch and Chen (1986) showed that when making a series of points within paragraphs it was helpful to list and enumerate them. Commentators suggest that it is best to use Arabic numbers when there is an order or sequence in the points being made: "Bullets," as used above, are perhaps more appropriate when each point is of equal value.

The structure and organization of a piece of text can often be made clearer for the reader by the use of numbered paragraphs. Such numbering systems can be used to organize information in many different ways, e.g., the heading numbers in this book. There has been little research on the effectiveness of such systems. Many people undoubtedly feel that they are valuable, particularly for reference purposes. But such systems can be abused if they are overdone, and they can lead to extraordinary confusion (see Smith & Aucella, 1983; Waller, 1980).

27.2. 10 Verbal Quantifiers

When numerical data are presented in text, prose descriptions often seem more comfortable to readers than do actual numbers. Everyday words that act as rough quantifiers, e.g., "nearly half the group," seem adequate for most purposes and are handled with reasonable consistency by most people (Moxey & Sanford, 1993). Young children, of course, may have greater difficulties with some of these terms (Badzinski, Cantor & Hoffner, 1989).

Research by Hartley, Trueman, and Rodgers (1984) suggested that the following phrases can be used with confidence with adults:

None the less, it may be better (or at least clearer for the reader) if more exact verbal equivalents of numbers are given. For example:

Verbal descriptions of probabilities are also more comfortable for most people than are actual probability statements. People are less consistent, however, in their interpretations of verbal descriptions of probability than they are in their interpretations of verbal descriptions of quantity (Moxey & Sanford, 1993). If precision is required, actual quantities can be given with a verbal quantifier For example, one can say "nearly half the group - 43% - said ..." or "There was a distinct chance (p < 0.06) that..."

27.2. 11 Signaling

A rather different way of making text organization more explicit is to use verbal "signals." Signals have been defined by Meyer, Young, and Bartlett (1989) as "noncontent words that serve to emphasize the conceptual structure or organization of the passage." Words and phrases such as however, but, or on the other hand signal to the reader that some form of comparison is to be made. Similarly, words and phrases such as firstly, secondly, '.three reasons for this are ..., a better example, however, would be..." signal the structure of the argument (and comparisons with subsections). Likewise, words and phrases such as therefore, as a result, so that, in order to, because signal causal relationships. Studies have shown that such signals help readers grasp the underlying structure of the author's argument (e.g., see Rice, Meyer & Miller, 1989). However, there may be some confusion in the future over the use of the term signal. I now find it being used to cover devices such as headings, overviews, and summaries (Lorch, Lorch & Inman, 1993).

27.2.12 Conclusions

This section on structure and access has shown that there is a good deal of research available on the variety of methods used to help readers grasp the structure of a text and to gain access to it. However, most of this research is uncoordinated and atheoretical. Most researchers focus. on one device or another, and few consider the effects of several such devices in combination. Few, too, carry out lengthy programs of research that aim to investigate systematically the myriad factors affecting the effectiveness of particular devices. (Some exceptions to this criticism are Dwyer's work on illustrations, Dansereau's work on outlines, Meyer's work on signals, and possibly my own on headings.) Such theories as there are thus buried below a welter of specific instances rather than being subjected to any rigorous analysis that might, in the long term, lead to deeper understanding.