UI Design Newsletter – January, 1999

Reading from paper versus screens: A critical review of the empirical literature, Dillon, A., Ergonomics, 35 (10), 1297-1326 (1992)

About seven years ago, Dillon (1992) reported a survey of the literature that compared reading from paper with reading from computer-based monitors. This paper is still referred to by many when discussing both performance and preference issues related to reading from paper versus monitors.

In general, he found that reading from paper tended to be preferred for tasks requiring much manipulation and navigation. Also, he reported no differences between paper and monitors in:

• accuracy,
• comprehension,
• fatigue, and
• the number of eye movements made by subjects.

However, the finding that is probably quoted most often is that silent reading from paper tended to be reliably faster (20-30%) than reading from monitors. This observation, which was reported by several studies in the 1980s, is reported in several current text books and is generally considered still to be true.

A recent study by Gujar, Harrison and Fishkin (1998) used a proofreading task to determine reading speed and comprehension. The authors introduced word substitution errors into the text, which forced their subjects to read and understand the text in order to find the errors. The substituted words varied grammatically from the original word. For example, the noun "cake" was replaced with the adjective "fake." The articles were all text, contained no graphics, and were written using 10-point Times Roman font.

Subjects read from a variety of different formats:

• a sheet of paper lying on a desk,
• paper displayed in a frame without glass,
• paper displayed in a frame under glass (to evaluate possible reflection problems),
• a monochrome monitor,
• a 21" color monitor, or
• a large image projected on a screen using an overhead transparency

They found no statistically significant differences in reading time or detected errors when reading from the different formats. Even though there were no performance differences, the subjects rated reading from paper significantly better than reading from the other 5 conditions.

The new, high resolution monitors allow users to read as fast from screens as they do from paper. It will be interesting to see how long it will take before this fact is generally known, and accepted as true by usability professionals.