Eye movement in reading comics

I’ve posted a few studies that have looked at how people’s eyes move across comic pages (here and here), and I recently found another. This short study looked at when people’s eye movements skip panels or go back and re-read them. 
They found that people spend more time reading panels with text than with just images, and that panels without text are more likely to be skipped and to be read with peripheral vision. Unusual panel arrangements (i.e. non-horizontal then vertical arrangements) also possibly led to jumping over panels (as was found in another study as well). After skipping these panels, participants then backtrack and re-read them.
These findings are consistent with previous studies that compared the eye-movements of expert and non-expert comic readers. Non-experts tend to focus more on text and read more erratically throughout a page. Experts tend to read more smoothly and focus more on the images. 
General studies like this are interesting, though I’d really like to see more studies that specifically target specific issues. Are there particular features of page layouts that motivate skipping panels? Are there features of layouts that impede on the actual comprehension of panels? Once we get beyond these very basic sorts of “what do eyes do generally” studies, we can really start exploring how looking at eye-movements can tell us about the comprehension of comic pages.

Chiba, Shinichi , Takamasa Tanaka, Kenji Shoji, and Fubito Toyama. 2007. “Eye Movement in Reading Comics.” In Proceedings of the 14th Annual International Display Workshops, 1255-58 


  • Thanks for this interesting post! I quite agree with you. Meanwhile, Nakazawa's study was only on two 12 years-old girls, and one of them (the "expert") had knowledge of the manga to be read – so a generalization of his findings there is excessive.

  • Sorry in advance if I make it too long.

    I've been reading some of your stuff and some studies you quote in your papers and blog. I've noticed that most of them take for granted (or assume) the fact that the panel layout conditions reading more than the dynamic lines inside a panel (for instance, an element forming a diagonal towards the panel to the right, while blockage would direct the reader to the panel below).

    Save from some specific exceptions, I personally agree with that assumption (that is what experience tells me, at least). However, I had an argument with an artist a couple of days ago. He tells me (he was told by his teachers, all professional comic artists)that coincidental strokes drawing an invisible line that goes from one panel to another are considered a mistake by artists and editors, since they may direct the readers to the wrong panel ('superindex' would be the literal translation from Spanish; I don't know if there is an equivalent to the concept in English).

    Our discussion was whether it is the gutters or these implied lines of continuity which excercise more power when orienting the reader. I have no doubt it is the panel layout that provides orientation when reading comics, although I admit internal elements can play a part in certain cases.

    Have you got any thoughts about this?

  • Thanks for the comment! I've heard similar comments about the internal elements of panels directing readers towards the next panel.

    From my experience (i.e. no data…) I'd say that panel layouts alone mostly direct a reader through a layout. The content then often interacts with that layout to tell you if you're right or wrong.

    However, I'd say that subtle directional cues in the panels, like the ones you describe, probably just reinforce those more overt cues, but don't constrain it enough to actually push someone the other way.

    So, if we were to do an eyetracking study where images in a blockage scenario were identical except for background lines that were directed at different subsequent panels, I'd guess people would still end up at the one directed by the layout. However, there might be some rapid eye movements going to the "wrong" panel suggested by the background lines.

    That's all to say that my guess is that such lines would contribute overall, but that their influence is fairly small.

  • Write a Reply or Comment