Introduction

I was motivated to write this article having discovered the studies undertaken by Adam Parker as part of his PhD at Bournemouth University, UK, supervised by Tim Slattery and Julie Kirkby. The thesis and published studies inspired me to revisit the trail of research on line length and provided me with some of the missing stepping stones through the bibliography. My personal interest stems from the idea that the new research might explain some of my own results from over 20 years ago, which challenged received wisdom. Because of this, I was never entirely confident about the reliability of my results.

Background

The common view

The consensus among typographers is that we should avoid long line lengths when designing continuous texts. Also, line spacing should be greater with longer lines. The research which supports these recommendations has reported that long line lengths decrease reading speed and increase the number of regressions (right-to-left saccades). In particular, the return sweep, the saccade from the end of one line to the beginning of a new line, is identified as problematic with long line lengths.

The explanation

The problem with return sweeps is that they can be inaccurate: undershooting (failing to reach the beginning of the next line) or locating the wrong line. The recent research from Slattery’s lab (referred to in the introduction) describes undershooting as an under-sweep. When the return sweep does not reach the beginning of the new line (undershoots), it is followed by a small leftward saccade to move back towards the beginning of the line. This is described as a corrective saccade. Undersweep-fixations are the brief pauses after the inaccurate return sweeps, before the corrective saccades (Parker et al., 2017). (See Figure 1.)

The assumptions

We have assumed that the inaccuracy of return sweeps and the need for corrective saccades are undesirable errors and so long line lengths should not be used. This is expressed in a more nuanced way within the field of psycholinguistics. “A tacit assumption in the field of eye-movement reading research has been that undersweep-fixations are simply the result of oculomotor error, reflecting little to no influence of ongoing linguistic processing.” (Slattery & Parker, 2019, p1948). Undersweep-fixations are very short, about half the duration of other fixations, i.e. about 130 ms rather than 250 ms (Slattery & Parker, 2019, p1949). Slattery’s lab has explored return sweep eye movements in greater depth and updated our knowledge of the reading process. Their research challenges the view that inaccurate return sweeps are necessarily problematic.

The evidence

Paterson and Tinker (1940) found that longer lines result in more regressions than shorter lines and these are “chiefly at the beginning of each line” (p576), i.e. after a return sweep. This was probably one of the earliest studies to report this finding. Later studies confirmed these results and clarified the nature of the regressions, describing them as corrective saccades (Beymer et al., 2005; Heller, 1982; Hofmeister et al., 1999; Parker & Slattery, 2020). Because return sweeps are longer than saccades within a line (see Figure 1), they were thought to be more susceptible to saccadic range error (McConkie et al., 1988). Longer lines increase the likelihood of undershooting the beginning of the line and requiring correction, a small saccade to the left.

Challenging the assumptions

New evidence

Surprisingly, reducing the oculomotor error associated with return sweeps does not increase reading speed. A study by Slattery and Vasilev (2019) set the first word on a line in bold and found that return sweeps landed closer to the left margin than normal setting and the probability of making a corrective saccade was also reduced. Despite the more accurate return sweeps, reading was not faster. This finding questions the explanation given for longer reading times for texts with longer lines, namely that “the smooth functioning of oculomotor habits” is upset (Paterson & Tinker, 1940, p577). This explanation may be too simplistic.

Another explanation

Why aren’t undersweep-fixations and the addition of corrective saccades slowing down reading? When we fixate a little to the right of the left margin (undersweep-fixation), we appear to be gaining some benefit from this fixation that helps subsequent reading of the line. This is evidenced by shorter time fixating on initial words on a line after an undersweep-fixation (Parker & Slattery, 2019). Also, a word is more likely to be skipped or receive a shorter fixation if it has received an undersweep-fixation prior to left-to-right reading (Slattery & Parker, 2019). A similar pattern of results is found in adults and children (Parker et al., 2020). Children as young as six can also obtain something useful from these oculomotor errors. Mechanisms are established early in reading development which control reading across line endings.

We benefit by briefly processing words during the undersweep-fixation, e.g. a preview of line-initial words in parafoveal vision (see Figure 2). Also, we give some attention to the word at the undersweep-fixation and this facilitates later lexical processing (Parker et al., 2020). The precise nature of the information extracted during under-sweep fixations is not yet established. However, we may be identifying letters and retaining their representations for subsequent processing, i.e. when recognizing words in left-to-right reading (Slattery & Parker, 2019).

Reconciling new knowledge with previous research

The original research

Attributing slower reading of long lines to the error in the landing location of return sweeps appears to be questionable. Therefore, it is worth looking at the origins of reports of slower reading. Along with others, I have often cited the work of Tinker, co-authored with Paterson. However, we should note that their research focused on text legibility and providing practical guidance for typographers, printers, and advertisers. Tinker was far less interested in the nature of visual processing (Legge, 2007), although he did do some eye movement studies.

Unfortunately, the validity of the studies undertaken by Tinker and Paterson is uncertain. A flaw in their method was noted in Parker’s thesis (p25) and referred to in Parker et al. (2019, p2–3) which led me to take a more detailed look at the early work. This revealed that Tinker and Paterson (1929) failed to separate the potential effects of the line length from effects of the content of the texts that were read in their study. The order of reading the texts was also not varied. Consequently, differences in reading speed could be influenced by the difficulty of specific texts, practice from previous reading, or line length. They did make a concerted effort to demonstrate that their results were due to typographical variations, and not these unwanted factors. But they did not substantially revise their experimental design to remove the confounds. (See future article for more detail.)

As it is not possible to adequately disentangle the effect of line length from the text content or the order of reading, the internal validity of Tinker and Paterson’s findings is questionable. This is extremely unfortunate as their data on typographical variables, including typeface, type size, line spacing, and line length comes from numerous experiments involving over 30,000 people. This research is some of the most extensive work on legibility (Morrison & Inhoff, 1981). Perhaps because of this, there have been few subsequent studies on line length in print.

Further doubt

There is one study which questions the difficulty with long lines in text using material printed in a line length far beyond Tinker’s safety zones (Tinker, 1963, p106). The researchers found that “an extreme line-length does not place an undue strain on the reader” (Hartley et al., 1978, p183). This study involved school children of around 11–13 years old, rather than university students who made up the majority of Tinker and Paterson’s participants. The long line length of 42 ems, about 178 mm or 110–115 characters per line (cpl), was not recommended. However, the results of Hartley et al. do suggest that even if the landing location of the return sweep needs to be corrected, this may not be as undesirable as previously assumed.

But, if we look only at research into print legibility, we have inadequate data to either support or refute Tinker and Paterson’s evidence that reading speed is slowed down with long lines.

Line length on screen

When research moved into investigating line length on screen, rather than in print, the results from some studies (including my own) started to reveal a trend towards faster reading of longer line lengths (Dyson & Kipping, 1998; Duchnicky & Kolers, 1983; Kolers et al., 1981; Shaikh & Chaparro, 2005). However, the results were not entirely consistent. This is probably unsurprising as researchers used different reading tasks, checks on comprehension, and line lengths. The longest line length across studies ranged from 70 cpl to 132 cpl.

Some of the work came from usability studies which were published online for rapid dissemination or presented at conferences. The accounts are therefore not always sufficiently detailed, and they may not have been subjected to rigorous peer review. Faster reading of long lines was not found by all researchers: two studies by Bernard et al. (2002, 2003) found no differences across three line lengths. Also, I did not replicate my results when I changed some aspects of my original study; long lines were not read consistently faster than shorter lines. However, in no case were long lines read more slowly than shorter lines, which seemed to contradict Tinker and Paterson’s findings and common views as noted above. Therefore, it did appear that the optimal line length for screen, at least in terms of speed of reading, may be different from print.

Removing inconsistency

In seeking to reconcile why longer line lengths may not slow down reading on screen but do so when reading print, I outlined some differences, e.g. visual angle, time spent scrolling. But although physical differences between reading from screen and reading print still exist, we now have direct evidence to explain why longer lines were read at least as fast as shorter lines. Readers can process words during the brief fixation when undershooting the start of the new line. This saves time in subsequent processing. Now we might also acknowledge that there is greater consistency between the range of optimal line lengths for print and screen.

Theory and practice

The new research from Slattery’s lab has updated our knowledge of eye movements in reading. Specifically, the studies have shown that we continue to read efficiently despite errors in locating the beginning of a new line. Surprisingly, this facility develops with little reading instruction as children show a similar pattern of results to adult fluent readers.

Although the new findings on return sweeps have gone some way towards exonerating long line lengths, I am hesitant to unreservedly recommend their use on screen to support efficient reading for a few reasons:

• Some readers, such as those with dyslexia, may benefit from shorter lines (Schneps et al., 2013)
• Further research is required to clarify whether inaccurate return sweeps may also locate the wrong line, as proposed by various researchers. If so, there may be an impact on reading speed. If such errors are more problematic than undershoots, how much line spacing is required to alleviate the problem?
• One of the more consistent outcomes across studies is that long line lengths are least preferred or judged as least easy to read (Bernard et al., 2002; Dyson & Kipping, 1998; Youngman & Scharff, 1998). A moderate line length is preferred which would also be the most familiar format.

But it would be exciting if designers were to challenge received wisdom, explore the unfamiliar, and work on overcoming any possible reluctance from readers to engage with different ways of presenting text.

Notes

References

Bernard, M.L., Fernandez, M., & Hull, S. (2002). The effects of line length on children and adults’ online reading performance. Usability News, 4(2). https://www.researchgate.net/publication/252707646_The_Effects_of_Line_Length_on_Children_and_Adults’_Online_Reading_Performance accessed 31 March 2021.

Bernard, M. L., Fernandez, M., Hull, S., & Chaparro, B. S. (2003). The effects of line length on children and adults’ perceived and actual online reading performance. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 47(11), 1375–1379. https://doi.org/10.1177%2F154193120304701112

Beymer, D., Russell, D. M., & Orton, P. Z. (2005). Wide vs. narrow paragraphs: an eye tracking analysis. In M. F. Costabile & F. Paternò (Eds.), Human-Computer Interaction - INTERACT 2005. Lecture Notes in Computer Science (Vol. 3585, pp. 741–752). Springer. https://doi.org/10.1007/11555261_59

Burt, C., Cooper, W. F., & Martin, J. L. (1955). A psychological study of typography. The British Journal of Statistical Psychology, 8, 29–57. https://doi.org/10.1111/j.2044-8317.1955.tb00160.x

Duchnicky, R. L., & Kolers, P. A. (1983). Readability of text scrolled on visual display terminals as a function of window size. Human Factors, 25(6), 683–692. https://doi.org/10.1177%2F001872088302500605

Dyson, M. C. (2005). How do we read text on screen? In H. van Oostendorp, L. Breure, & A. Dillon (Eds.), Creation, use, and deployment of digital information (pp. 279–306). Lawrence Erlbaum Associates. https://doi.org/10.4324/9781410613035

Dyson, M.C. (2018). Legibility: How and why typography affects ease of reading. Centro de Estudios Avanzados de Diseño.

Dyson, M. C., & Beier, S. (2016). Investigating typographic differentiation: italics are more subtle than bold for emphasis. Information Design Journal, 22(1), 3–18. https://doi.org/10.1075/idj.22.1.02dys

Dyson, M. C., & Haselgrove, M. (2001). The influence of reading speed and line length on the effectiveness of reading from screen. International Journal of Human-Computer Studies, 54, 585–612. https://doi.org/10.1006/ijhc.2001.0458

Dyson, M. C., & Kipping, G. J. (1997). The legibility of screen formats: are three columns better than one? Computers & Graphics, 21(6), 703–712. https://doi.org/10.1016/S0097-8493(97)00048-4

Dyson, M.C., & Kipping, G.J. (1998). The effects of line length and method of movement on patterns of reading from screen. Visible Language, 32(2), 150–181. https://visiblelanguage.herokuapp.com/issue/115

Hartley, J. (2013). Cyril Burt and text design: A brief note. Educational & Child Psychology, 30(3), 33–36.

Hartley, J., Burnhill, P., & Davis, L. (1978). The effects of line length and paragraph denotation on the retrieval of information from prose text. Visible Language, 12(2), 183–194. https://visiblelanguage.herokuapp.com/issue/46

Heller, D. (1982). Eye movements in reading. In R. Groner & P. Fraisse (Eds.), Cognition and eye movements (pp. 487–498). Deutscher Verlag der Wissenschaften.

Hofmeister, J., Heller, D., & Radach, R. (1999). The return sweep in reading. In W. Becker, H. Deubel & T. Mergner (Eds.), Current oculomotor research: Physiological and psychological aspects (pp. 349–357). Springer. https://doi.org/10.1007/978-1-4757-3054-8_49

Kolers, P.A., Duchnicky, R.L., & Ferguson, D.C. (1981). Eye movement measurement of readability of CRT displays. Human Factors, 23(5), 517–527. https://doi.org/10.1177%2F001872088102300502

Kruk, R.S., & Muter, P. (1984). Reading of continuous text on video screens. Human Factors, 26(3), 339–345. https://doi.org/10.1177%2F001872088402600309

Legge, G.E. (2007). Psychophysics of reading in normal and low vision. Lawrence Erlbaum Associates. https://psycnet.apa.org/record/2006-20917-000

McConkie, G. W., Kerr, P. W., Reddix, M. D., & Zola, D. (1988). Eye-movement control during reading. 1. The location of initial eye fixations on words. Vision Research, 28(10), 1107–1118. https://doi.org/10.1016/0042-6989(88)90137-x

Morrison, R.E., & Inhoff, A.-W. (1981). Visual factors and eye movements in reading. Visible Language, 15(2), 129–146. https://visiblelanguage.herokuapp.com/issue/58

Parker, A. J. (2019). The return-sweep in reading [Doctoral thesis, Bournemouth University]. BURO. http://eprints.bournemouth.ac.uk/32170/

Parker, A.J., Kirkby, J.A., & Slattery, T.J. (2017). Predictability effects during reading in the absence of parafoveal preview. Journal of Cognitive Psychology, 29(8), 902–911. https://doi.org/10.1080/20445911.2017.1340303

Parker, A. J., Kirkby, J. A., & Slattery, T. J. (2020). Undersweep fixations during reading in adults and children. Journal of Experimental Child Psychology, 192, 104788. https://doi.org/10.1016/j.jecp.2019.104788

Parker, A. J., Nikolova, M., Slattery, T. J., Liversedge, S. P., & Kirkby, J. A. (2019). Binocular coordination and return-sweep saccades among skilled adult readers. Journal of Vision, 19(6), 10. https://doi.org/10.1167/19.6.10

Parker, A.J., & Slattery, T.J. (2019). Word frequency, predictability, and return-sweep saccades: Towards the modeling of eye movements during paragraph reading. Journal of Experimental Psychology-Human Perception and Performance, 45(12), 1614–1633. https://doi.org/10.1037/xhp0000694

Parker, A.J., & Slattery, T.J. (2020). Spelling ability influences early letter encoding during reading: Evidence from return-sweep eye movements. Quarterly Journal of Experimental Psychology, 74(1), 135–149. https://doi.org/10.1177%2F1747021820949150

Paterson, D.G., & Tinker, M.A. (1940). Influence of line width on eye movements. Journal of Experimental Psychology, 27(5), 572–577. https://doi.org/10.1037/h0054498

Rayner, K., & Pollatsek, A. (1989). The psychology of reading. Lawrence Erlbaum.

Schneps, M.H., Thomson, J.M., Sonnert, G., Pomplun, M., Chen, C., & Heffner-Wong, A. (2013). Shorter lines facilitate reading in those who struggle. PloS ONE, 8(8), e71161. https://doi.org/10.1371/journal.pone.0071161

Shaikh, A.D., & Chaparro, B.S. (2005). The effects of line length on reading performance of online news articles. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 49(5), 701–705. https://doi.org/10.1177%2F154193120504900514

Slattery, T.J., & Parker, A.J. (2019). Return sweeps in reading: Processing implications of undersweep-fixations. Psychonomic Bulletin & Review, 26(6), 1948–1957. https://doi.org/10.3758/s13423-019-01636-3

Slattery, T.J., & Vasilev, M.R. (2019). An eye-movement exploration into return-sweep targeting during reading. Attention, Perception, & Psychophysics, 81(5), 1197–1203. https://doi.org/10.3758/s13414-019-01742-3

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

Tinker, M.A., & Paterson, D.G. (1929). Studies of typographical factors influencing speed of reading: III Length of line. Journal of Applied Psychology, 13, 205–219. https://doi.org/10.1037/h0073597

Vasilev, M.R., Adedeji, V.I., Laursen, C., Budka, M., & Slattery, T.J. (2021). Do readers use character information when programming return-sweep saccades? Vision Research, 183, 30–40. https://doi.org/10.1016/j.visres.2021.01.003

Youngman, M., & Scharff, L. (1998). Text width and margin width influences. Paper presented at SWPA 1998. https://laurenscharff.com/research/textmargin.html accessed 31 March 2021.