he Internet has grown dramatically, particularly as a source for online training and education. Web-based teaching and learning is advocated as a cost-effective supplement or replacement of traditional media. As more post-secondary institutions and corporate training departments adopt an online delivery model, the need to understand the effectiveness of this new medium takes on increasing immediacy. Formal research in optimizing the presentation of online information will allow developers to create online materials that maximize users' comprehension.
The presentation of information provides the foundation for online learning. The physical layout of online content is a key component in the overall presentation. One aspect of layout is the arrangement of text, and given that most current online content is textual, this factor is fundamental in determining the effectiveness of a presentation format. However, the current popular literature offers conflicting advice as to what constitutes optimal text layout.
Lynch and Horton (1999) have argued, "the lines of text on most Web pages are much too long for easy reading" (p.85). Lynch and Horton therefore advise Web designers to constrain the width of text passages in order to make it easier for users to read. Reducing line length results in non-text blank space, or whitespace, surrounding the text as interpreted by most browsers. Bernard, Chaparro, and Thomasson (2000) examined users' search times and preferences for Web sites differing in amount of whitespace. They found no difference in search time as a function of amount of whitespace but users preferred enough whitespace to separate different units of information. Spool, Scanlon, Schroeder, Snyder, and DeAngelo (1998) completed a usability study of major commercial Web sites, and found that an increase in whitespace corresponded to a decrease in performance on a search task. Thus, the findings of Bernard et al. and Spool et al. indicate that more text on the screen might be more appealing and easier to read than less text surrounded by more whitespace.
Authors of print design texts (e.g., Wilson, 1974) have advised that text presented in a print format be displayed in narrow columns. Print research supports this assertion by indicating that narrow columns of text are read faster and are comprehended better (Tinker, 1963). Early research on reading from computer monitors also supports the notion that short line lengths and increased whitespace aids reading speed and comprehension (van Nes, 1986; de Bruijn, de Mul, & van Oostendorp, 1992). Manipulations of text layout to optimize reading speed from computer monitors focused originally on whitespace variables like spacing and indenting (Muter & Maurutto, 1991).
Updated research is needed on reading from computer monitors. Advances in technology have significantly increased the quality of monitor displays and may have changed the effects of line length on reading and comprehension. As well, people are beginning to use the Web more and more and experience in reading from computer monitors while browsing or learning from the Web might make a difference in the effect of line length on reading and comprehension. Dyson and Kipping (1998) found that, although participants preferred shorter line lengths for reading from the Web, their reading comprehension was not affected by narrow versus wide text layout.
Given the inconsistencies in the research and in the design literature, we performed our own study of text layout in order to increase understanding of the optimal text presentation format. There are many components of text layout, including line length, whitespace surrounding the text, spacing between lines, size and shape of characters (font size and type), spacing between characters (kearning), and alignment of text (justification) (Gould, Alfaro, Barnes, Finn, Grischkowsky, & Minuto, 1987). We chose to focus solely on line length and whitespace. Web developers can more easily control these two factors than other factors, such as font size and type, which are affected by system controls.
Although our primary concern was disentangling the line length issue, we included an examination of total whitespace as a function of "text density" because, as line length varies, so does the amount of surrounding empty space. Some Web pages are sparse _ others are packed with text and graphics. Thus, we varied whitespace by changing the amount of text displayed in order to examine the effects of this potential confound to the manipulation of line length.
Participants read short prose passages in wide or narrow formats and with the surrounding whitespace filled or not filled with additional, unimportant information. We defined line length as the number of characters, including spaces, which are displayed as one line in the application window. We defined whitespace as the amount of information presented; in the one-column/paragraph condition, only the to-be-read text was presented and in the two-column/paragraph condition, additional, meaningless text was presented next to or below the to-be-comprehended text. Combining these line length and whitespace factors, we created four conditions: Narrow passages, with approximately 55 characters per line, were either presented alone, in a single column of text with white space to the right of the text, or in a two-column format, with uninformative text presented in a second column to the right of the text. Wide passages, with 115 characters per line, were either presented alone with white space below the text or in a two-paragraph format, with uninformative text presented in a paragraph below the to-be-read text.
We assessed the effects of line length and whitespace on comprehension by having participants read the texts and answer multiple choice questions about each text. To determine whether experience with reading off a computer screen might affect the results, participants also completed a computer experience questionnaire.
The presentation of information provides the foundation for online learning. The physical layout of online content is a key component in the overall presentation. One aspect of layout is the arrangement of text, and given that most current online content is textual, this factor is fundamental in determining the effectiveness of a presentation format. However, the current popular literature offers conflicting advice as to what constitutes optimal text layout.
Lynch and Horton (1999) have argued, "the lines of text on most Web pages are much too long for easy reading" (p.85). Lynch and Horton therefore advise Web designers to constrain the width of text passages in order to make it easier for users to read. Reducing line length results in non-text blank space, or whitespace, surrounding the text as interpreted by most browsers. Bernard, Chaparro, and Thomasson (2000) examined users' search times and preferences for Web sites differing in amount of whitespace. They found no difference in search time as a function of amount of whitespace but users preferred enough whitespace to separate different units of information. Spool, Scanlon, Schroeder, Snyder, and DeAngelo (1998) completed a usability study of major commercial Web sites, and found that an increase in whitespace corresponded to a decrease in performance on a search task. Thus, the findings of Bernard et al. and Spool et al. indicate that more text on the screen might be more appealing and easier to read than less text surrounded by more whitespace.
Authors of print design texts (e.g., Wilson, 1974) have advised that text presented in a print format be displayed in narrow columns. Print research supports this assertion by indicating that narrow columns of text are read faster and are comprehended better (Tinker, 1963). Early research on reading from computer monitors also supports the notion that short line lengths and increased whitespace aids reading speed and comprehension (van Nes, 1986; de Bruijn, de Mul, & van Oostendorp, 1992). Manipulations of text layout to optimize reading speed from computer monitors focused originally on whitespace variables like spacing and indenting (Muter & Maurutto, 1991).
Updated research is needed on reading from computer monitors. Advances in technology have significantly increased the quality of monitor displays and may have changed the effects of line length on reading and comprehension. As well, people are beginning to use the Web more and more and experience in reading from computer monitors while browsing or learning from the Web might make a difference in the effect of line length on reading and comprehension. Dyson and Kipping (1998) found that, although participants preferred shorter line lengths for reading from the Web, their reading comprehension was not affected by narrow versus wide text layout.
Given the inconsistencies in the research and in the design literature, we performed our own study of text layout in order to increase understanding of the optimal text presentation format. There are many components of text layout, including line length, whitespace surrounding the text, spacing between lines, size and shape of characters (font size and type), spacing between characters (kearning), and alignment of text (justification) (Gould, Alfaro, Barnes, Finn, Grischkowsky, & Minuto, 1987). We chose to focus solely on line length and whitespace. Web developers can more easily control these two factors than other factors, such as font size and type, which are affected by system controls.
Although our primary concern was disentangling the line length issue, we included an examination of total whitespace as a function of "text density" because, as line length varies, so does the amount of surrounding empty space. Some Web pages are sparse _ others are packed with text and graphics. Thus, we varied whitespace by changing the amount of text displayed in order to examine the effects of this potential confound to the manipulation of line length.
Participants read short prose passages in wide or narrow formats and with the surrounding whitespace filled or not filled with additional, unimportant information. We defined line length as the number of characters, including spaces, which are displayed as one line in the application window. We defined whitespace as the amount of information presented; in the one-column/paragraph condition, only the to-be-read text was presented and in the two-column/paragraph condition, additional, meaningless text was presented next to or below the to-be-comprehended text. Combining these line length and whitespace factors, we created four conditions: Narrow passages, with approximately 55 characters per line, were either presented alone, in a single column of text with white space to the right of the text, or in a two-column format, with uninformative text presented in a second column to the right of the text. Wide passages, with 115 characters per line, were either presented alone with white space below the text or in a two-paragraph format, with uninformative text presented in a paragraph below the to-be-read text.
We assessed the effects of line length and whitespace on comprehension by having participants read the texts and answer multiple choice questions about each text. To determine whether experience with reading off a computer screen might affect the results, participants also completed a computer experience questionnaire.
In my opinion there is no generalized answer to this question. Anyone who presents an answer to this question that is not based upon analytics is blowing smoke - and the answer that they give will not apply to your different situation. Too many factors are at play and these include... how the pages are linked, where those links are placed, how interesting the writing is, who your traffic is, how badly your traffic wants or needs your content, format of the writing.If you want the answer to this question you should take your best shot at the format then run analytics. When you have some information on a lot of visitors then change the format and run again. Repeat using a few different formats and you will learn something, plus you will have an answer that is based upon data and not on hot air.If you don't use analytics you are simply guessing at the answer. Most of my guesses are wrong.Also, other variables are at play... I have lots of long articles on my site and some of them are multipage and some are the single long page. In addition to user engagement, the single vs long page formats can have a dramatic difference in search engine ranking, traffic inflow - and Income! Again, analytics is the only way to measure these things. Combine user engagement with income and rankings and you have a multivariate puzzle that would challenge a statistician - yet it is easy to tell when something bombs or performs strongly.I have articles on my site that I have tested with long pages, short pages, intermediate pages, different linkage formats - over millions of pageviews and I am always amazed at the massive differences that are observed and that I am unable to have guessed them in advance. Huge differences in traffic, engagement and income between formats.I have done very little content generation on my site for the past several months. Instead I've spent almost all available time testing, tweaking, rebuilding. The goal is to find the best intersection of user engagement, SEO and income. The result is a 50% increase in income per visitor. But, although I think that I have learned a lot I can almost always improve upon my first guess at performance if I watch what happens and then modify accordingly. All of this is done with clicktracks, crazyegg, weblogexpert and income data.One final factor that is very important... after you have done all of the above you take a look at your site and ask if you have damaged or enhanced the linkability of the article. If you double your income but stink up your site you probably