HUMAN FACTORS, 1984, 26, (3), 339-345.

Copyright (c) 1984, The Human Factors Society, Inc. All rights reserved.

Reading of Continuous Text on Video Screens

RICHARD S. KRUK and PAUL MUTER, Department of Psychology, University of Toronto, Toronto, Ontario, Canada

Address correspondence to Paul Muter, Department of Psychology, University of Toronto, Toronto, Ont., M5S 3G3, Canada, muter@psych.utoronto.ca.

ABSTRACT. In three experiments, the finding of slower reading of text from a video screen than from a book was replicated, and several possible reasons for this effect were explored. Extra time (9 seconds) taken to fill the screen had no significant effect on reading time in the video condition. Similarly, varying the contrast ratio of the video image and the distance between the screen and the reader had no effect on reading speed. The format used in the video condition (39 characters per line and 20 lines per page) produced slower reading than did a format typical for books (60 characters per line and 40 lines per page), but this effect alone (9.5%) could not account for the difference in reading speed between the book and video conditions (24.1%). The reduced reading speed was partly overcome by avoiding single spacing, which produced 10.9% slower reading than did double spacing in the video condition.

INTRODUCTION

Research on optimal methods of presenting text on CRTs is mushrooming (e.g., Duchnicky and Kolers, 1983; Juola, Ward, and McNamara, 1982; Pastoor, Schwarz, and Beldie, 1983; Treurniet, 1980), but there remain many unanswered questions in this area. The present studies investigate the effects on reading of the time required to fill the screen, the number of characters per line and the number of lines per page, the distance between the screen and the reader, the interline spacing, and the contrast ratio of the video image.

Muter, Latremouille, Treurniet, and Beam (1982) found that extended reading of continuous text on television screens is feasible, but that reading was 28.5% slower from a television screen than from a book. There are several possible causes of this difference. First, slower reading in the video (television) condition could have been caused by the reduced number of characters per line or lines per page or both. Subjects in the Muter et al. study read text in the video condition with 39 characters per line and 18 lines per page; text in the book condition contained 60 characters per line and 40 lines per page. Past research (e.g., Duchnicky and Kolers, 1983; Kolers, Duchnicky, and Ferguson, 1981; Tinker, 1963) indicates that the number of characters per line can affect reading speed, though the effect depends on a number of factors.

A second possible cause of slower reading in the video condition was the time required to fill the screen. In the Muter et al. study, it took 9 seconds to fill the screen for each new page. This may have distracted subjects.

Third, the distance between the reader and the reading material was different in the two conditions. Research investigating the effects of distance on reading continuous text is scarce (Morrison, 1983), but Kak (1981) varied this distance from 35.6 to 111.8 cm and found that CRT reading rate decreased as distance increased. On the other hand, Morrison and Rayner (1981) have shown that perceptual span (the number of characters read within a fixation) remains constant as distance is increased from 36 cm to 71 cm.

Fourth, a difference in the contrast ratio may have affected reading. In a study of word recognition, Timmers, Van Nes, and Blommaert (1980) found that a reduction of the background-to-text contrast ratio from 10:1 to 1.14:1 resulted in a substantial increase in time to perform the task.

Finally, the interline spacing was different in the two conditions of the Muter et al. (1982) study, and this variable may affect reading speed (Kolers, Duchnicky, and Ferguson, 1981). Morrison and Inhoff (1981) have suggested that an increase in the blank area between lines decreases lateral masking (the interference of surrounding letters on word perception) and results in more accurate return sweeps.

EXPERIMENT 1

Experiment 1 was designed to replicate the reading speed result of Muter et al. with a different computer display; to assess the effect of format (number of characters per line and number of lines per page); and to determine the effect of delays in presentation.

There were two book conditions and two video conditions. In one book condition, text was printed with 60 characters per line and 40 lines per page; in the other book condition, text was printed with 39 characters per line and 20 lines per page. In the delayed video condition, 9 s was required to fill the screen for each page; in the instant video condition, the screen was filled within 0.5 s.

Method

Subjects.
Subjects were 17 female and 7 male undergraduates at the University of Toronto. Subjects in all three experiments had at least 20/20 vision, normal or corrected, as measured by an eye chart. No subject participated in more than one experiment.

Materials. Reading materials were four sets of short stories written by H. H. Munro (1976). Stories in the book conditions were presented in typewritten booklets made up of 21.5 cm x 28 cm paper. Text material in the book condition analogous to the typical book was printed with 60 characters per line and 40 lines per page. The number of words per page was approximately 400. Text material in the video conditions and in the second book condition was printed with 39 characters per line and 20 lines per page, yielding approximately 130 words per page. The character set was the same in both book conditions: approximately four characters per cm in width, produced by a Pica 10-pitch daisy wheel. Text in all conditions was left-justified.

In the video conditions, text was displayed on a 30.5 cm (diagonal) green monochrome video monitor (Amdek 100G). Display of text was controlled by an Apple II Plus microcomputer with an upper- and lowercase character generator. The character set was the same in both video conditions: approximately two characters per cm in width. The character set was designed on a 5 x 7 dot matrix. Luminance of the characters was 46.6 cd/m2 and luminance of the background was 6.6 cd/m2. An example of text presented in the video conditions is shown in Figure 1.

Text in the video conditions was delivered in full-page form; there was no scrolling. A new page was presented whenever any key on the keyboard of the computer terminal was pressed. In the delayed video condition, a total of 9 s was required to fill the screen. in the instant video condition, the screen was filled within 0.5 s after the subject pressed a key. Text appeared in the standard order: left-to-right and top-to-bottom.

Subjects sat approximately 150 cm from the video screen in a soundproof booth. A cardboard visor was affixed to the top of the monitor to reduce glare. The booth was illuminated from a light source located above and behind the subject. In the book conditions, subjects were permitted to hold the booklet in whatever manner they preferred.

Design. Each subject experienced all four conditions: instant video, delayed video, 60-character-per-line book, and 39-character-per-line book. Each subject read the four sets of short stories in the same order, but each subject received a different one of the 24 possible orders of the four conditions.

Procedure. Subjects were informed that they would be reading two sets of short stories from a video screen and two sets in booklet form, and that their reading speeds and comprehension would be measured. Subjects were told to read as if for pleasure, but that they were not permitted to turn to previous pages. Subjects read each set of stories for 5 min. At the end of each 5-min period, the experimenter noted how far the subject had read. Finally, after this 20 min of reading, subjects were given four comprehension tests, one for each set, each lasting 5 min. Comprehension tests, which were created by the authors, were based on the first three full book pages of the first story of each set, and consisted of 12 questions: two multiple-choice questions, four one-word-answer questions, four short-answer questions, and two long-answer questions. Tests were presented in the same order as the stories so that the time period between a story and its corresponding test was always the same.

Results and Discussion

Mean reading speed scores are listed in Table 1. As demonstrated in Muter et al. (1982), reading of continuous text was slower from a CRT screen than from a book. Subjects read somewhat more slowly when given a booklet with 39 characters per line and 20 lines per page than they did when presented with a booklet with 60 characters per line and 40 lines per page, but the difference was small (9.5%) compared with the overall difference between the book and video conditions (24.1%). This suggests that the difference in format cannot completely account for the difference between the book and video conditions. Furthermore, Table 1 suggests that delay in presentation is not important.

Statistical analyses support these conclusions. A one-way repeated measures ANOVA yielded a significant result, F(3,69) = 41.7, p < 0.001. Tukey tests revealed that the instant video and delayed video conditions were not significantly different from each other, but that all other pairs of conditions were significantly different: Reading in the 39-character-per-line book condition was faster than in the instant video condition, p < 0.01, and reading in the 60-character-per-line book condition was faster than in the 39-character-per-line condition, p < 0.05.

Mean comprehension scores are also given in Table 1. Only questions on material that had been read by all subjects were included in the analyses of comprehension scores. A one-way repeated measures ANOVA revealed no significant effect, F(3,69) = 1.18.

EXPERIMENT 2

In Experiment 2, the distance between the subject and the video monitor was varied. There were no book conditions.

Method

Subjects.
Eight male and 10 female undergraduates at the University of Toronto took part in the study.

Materials. Reading materials consisted of the first three sets of short stories used in the first experiment. The apparatus used in the experiment was the same as that used in Experiment 1. The only exception was a chin rest, which held subjects' heads at specific positions throughout the experiment. Three distances were used: 40 cm, 80 cm, and 120 cm. As in the instant video condition in Experiment 1, the screen was filled within 0.5 s after the subject pressed a key.

Design. The distance between the subject and the screen was a within-subject variable. Each of the six different possible orders of presentation of the three distances was used three times.

Procedure. The procedure was similar to that of Experiment 1. Subjects were informed that they would be reading three sets of short stories from three different distances. As there were now three sets instead of four, each reading period and testing period was increased to 6 min to increase sensitivity.

Results and Discussion

Mean reading speed scores are listed in Table 2. The ANOVA revealed no significant effect, F(2,34) = 1.36. This result is consistent with the finding of constant perceptual span over different distances (Morrison and Rayner, 1981), but contradicts the result of Kak (1981), who found that reading speed was faster for closer distances. Kak used distances of 35.6 cm, 73.7 cm, and 111.8 cm in a between-subjects design with four subjects per distance. She used a 22.9-cm monitor, also with an Apple microcomputer, so the visual angles of her characters were slightly smaller than those in the present study. The difference in results is perplexing. The difference cannot be attributed to a lack of sensitivity in the present experiment: Given our error mean square, her mean reading speeds (estimated from Kak, Figure 1) would have produced a highly significant result.

Mean scores for comprehension are also presented in Table 2. Again, the ANOVA revealed no significant effect, F(2,34) < 1.

EXPERIMENT 3

Both contrast ratio and space between lines (as a proportion of height of characters) were greater in the book condition than in the video condition in Experiment 1. The same was true in Muter et al. (1982). Experiment 3 examined the effects of varying contrast ratio and interline spacing on speed of reading from a video monitor.

Method

Subjects.
Nine female and three male students at the University of Toronto served as subjects.

Materials. The reading materials were the same as in Experiment 1. The apparatus was the same as in Experiment 2. Only one distance (80 cm) was used, but the chin rest was retained. The contrast ratio was varied from 4.6:1 to 83:1 by turning the contrast knob on the monitor. The luminance of the green phosphor was 44.8 cd m2 in the low-contrast condition and 82.1 cd/m2 in the high-contrast condition; the luminance of the background was 9.7 cd/m2 in the low-contrast condition and 9.9 cd/m2 in the high-contrast condition. Interline spacing was varied by inserting blank lines in the double-space condition. In order to equate number of lines of text per condition, the top and bottom portions of the screen were unused in the single-space condition. Thus, screen lines 5 to 14 inclusive were used in the single-space condition, and odd-numbered lines 1 to 19 inclusive were used in the double-space condition. The instant mode of presentation was again used.

Design. Both spacing and contrast ratio were within-subject variables. Spacing and contrast ratio were varied factorially in a 2 x 2 design. To ensure that each of the four conditions occurred in each ordinal position equally often, subjects were tested in three orthogonal 4 x 4 Latin-square designs. As in Experiments 1 and 2, the order of the sets of stories was the same for every subject.

Procedure. The procedure was similar to that of Experiment 2. Subjects read four sets of material for 5 min each.

Results and Discussion

Mean reading speeds for Experiment 3 are given in Table 3. Contrast ratio had no statistically significant effect, F(1,33) < 1, on mean reading speed. However, mean reading speed was 10.9% slower in the single-space condition than in the double-space condition. This difference was statistically significant, F(1,33) = 9.12, p < 0.01. The effect of spacing did not depend on the contrast ratio, F(1,33) < 1.

Kolers, Duchnicky, and Ferguson (1981) also found a significant difference between double and single spacing, but their difference was only 2.2% compared with the present 10.9%. This discrepancy may be somehow attributable to the fact that spacing was confounded with lines per page and words per page in the Kolers et al. study, whereas lines per page and words per page were constant in the present experiment. On the other hand, the data of Duchnicky and Kolers (1983) suggest that doubling the number of printed lines per page would have little effect on reading speed. A more likely explanation of the discrepancy is that in the single-space condition, the space between lines, as a proportion of the height of the characters, was apparently greater in the Kolers et al. experiment than in the present experiment.

As in Experiments 1 and 2, the stimulus conditions did not affect comprehension scores in Experiment 3 (see Table 3). The comprehension tests in the present experiments may have been relatively insensitive. Another possibility is that subjects tended to trade off speed and accuracy in such a way as to maintain a constant level of comprehension. Duchnicky et al. (1983) and Kolers et al. (1981) also found that comprehension levels did not vary as a function of a wide range of variables.

GENERAL DISCUSSION

In the present experiments, the finding of slower reading from a video screen than from a book (Muter et al., 1982) was replicated with a different display, several possible reasons for this difference in reading speed were ruled out, and two partial explanations for the difference were suggested.

Muter et al. (1982) found that reading white text on a blue background from a Telidon videotex terminal (Bown, O'Brien, Sawchuck, Storey, and Treurniet, 1980) was 28.5% slower than reading from a book. In the present Experiment 1, reading text on a smaller green-phosphor monitor driven by a microcomputer was 24.1% slower than reading from a booklet.

Experiment 1 ruled out delay of presentation of material as an explanation for the slower reading in the video condition: Reading speed was the same whether 9 seconds or less than 0.5 seconds was required to fill the screen with text.

Experiment 2 demonstrated that varying the distance between the reader and the video monitor from 40 to 120 cm had no effect on reading speed. This result is somewhat surprising but is consistent with the reasoning of Morrison (1983), who argued that the increase in retinal image size, as distance is decreased, is offset by the decrease in acuity, as the image falls into more eccentric portions of the retina.

Experiment 3 showed that increasing the contrast ratio in the video condition from 4.6:1 to 8.3:1 had no effect on reading speed. Timmers et al. (1982) found that contrast ratio did influence performance. There are three possible reasons for this difference in results: (1) the Timmers et al. experiment was conducted with paper materials; (2) it involved word recognition rather than reading; and (3) contrast ratio was varied over a wider range (1.14:1 to 10:1) than in the present experiment.

The present results suggest that there is no single cause of slower reading in the video condition. Apparently there are at least two causes: format (number of characters per line and number of lines per page) and vertical spacing. In Experiment 1, substituting the video format (39 characters per line and 20 lines per page) into the book condition produced a decrement in reading speed of 9.5%. Tinker (196) suggested that a reduction in book reading speed should be observed with excessively short and excessively long lines. The 60-character line is longer than the upper limit suggested by Tinker for optimal legibility, yet it was read more quickly than the 39-character line, though 39 is above Tinker's suggested lower limit for optimal legibility. On the other hand, the present result is consistent with data of Duchnicky and Kolers (1983) and Kolers, Duchnicky, and Ferguson (1981), who varied format within video conditions and found that fewer characters per line produced slower reading.

In Experiment 3, single spacing produced reading that was 10.9% slower than that produced by double spacing. This result suggest that tight vertical spacing in the video condition may contribute to slow reading, though the results of Kolers et al. (1981) suggest that single spacing may be a negligible problem with certain displays. Perhaps single spacing should be particularly avoided with computer displays in which the space between lines is small relative to the height of the characters.

Further research is needed to explore several other factors that may contribute to the difference in reading speed between book and video conditions. These include posture, image polarity (dark on light versus light on dark), character set, resolution, justification (left or full) and familiarity with the medium.

CONCLUSION

Reading from a video screen with 39 characters per line, 20 lines per page, and single spacing was substantially slower than reading from a book. There were no significant effects on comprehension. The present results suggest that, within the ranges studied, contrast ratio, time required to fill the screen, and distance from the screen do not affect reading speed, but that format (characters per line and lines per page) and interline spacing do.

ACKNOWLEDGMENTS

This work was supported by Research Grant U0149 from the Natural Sciences and Engineering Research Council of Canada to the second author. We thank Lochlan E. Magee, Nancy L. Williamson, and especially Robert L. Duchnicky for helpful comments.

REFERENCES

Bown, H., O'Brien, C. D., Sawchuk, W., Storey, J. R., and Treurniet, W. C. (1980). Telidon Videotex and user-related issues. In P. A. Kolers, M. E. Wrolstad, and H. E. Bouma, (Eds.), Processing of visible language 2 (pp. 473-479). New York: Plenum.

Duchnicky, R. L., and Kolers, P. A. (1983). Readability of text scrolled on video display terminals as a function of window size. Human Factors, 25, 683-692.

Juola, J. F., Ward, N. J., and McNamara, T. (1982). Visual search and reading of rapid serial presentations of letter strings, words, and text. Journal of Experimental Psychology: General, 111, 208-227.

Kak, A. V. (1981). Relationships between readability of printed and CRT-displayed text. In Proceedings of the Human Factors Society, 25th Annual Meeting (pp. 137-140). Santa Monica, CA: Human Factors Society.

Kolers, P. A. Duchnicky, R. L., and Ferguson, D. C. (1981). Eye movement measurement of readability of CRT displays. Human Factors, 23, 517-527.

Morrison, R. E. (1983). Retinal image size and the perceptual span in reading. In K. Rayner, Eye movements in reading: Perceptual and language processes (pp. 31-40). New York: Academic Press.

Morrison, R. E., and Inhoff, A. W. (1981). Visual factors and eye movements in reading. Visible Language, 15, 129-146.

Morrison, R. E., and Rayner, K. (1981). Saccade size in reading depends upon character spaces and not visual angle. Perception and Psychophysics, 30, 395-396.

Munro, H. H. (1976). The complete works of Saki. Garden City, NY: Doubleday.

Muter, P., Latremouille, S. A., Treurniet, W. C., and Beam, P. (1982). Extended reading of continuous text on television screens. Human Factors, 24, 501-508.

Pastoor, S., Schwarz, E., and Beldie, I. P. (1983). The relative suitability of four dot-matrix sizes for text presentation on color television screens. Human Factors, 25, 265-272.

Timmers, H., Van Nes, F. L., and Blommaert, F. J. J. (1980). Visual word recognition as a function of contrast. In E. Grandjean and E. Vigliani (Eds.), Ergonomic aspects of visual display terminals (pp. 115-120). London: Taylor and Francis.

Tinker, M. A. (1963). Legibility of print. Ames, IA: Iowa State University Press.

Treurniet, W. C. (1980). Spacing of characters on a television display. In P. A. Kolers, M. E. Wrolstad, and H. E. Bouma (Eds.), Processing of visible language 2 (pp. 365-374). New York: Plenum.