Remember that meme that got passed around about some study that showed that people parsed lower-case letters much more easily than they did all caps? Every word in the paragraph was misspelled, but with letters that were approximately the same shape as what they replaced-- a c where an o should be, an f where a t should be.
I can’t find it, no doubt because search engines can’t match up the misspellings with the words they are supposed to represent. If somebody has a link, I’d really appreciate it.
The study was done at Cambridge University. I don’t recall that it had anything to do with caps, except that mixing cases made it harder. But MiXinG CasEs always makes reading harder, regardless of whether the words are scrambled. Same for ALL CAPS.
Comics readers like myself who’ve enjoyed the habit for decades have little problems reading all caps lettering since that’s the standard for comic books and comic strips. It’s probably more a case of media conditioning: I expect all caps reading comics, but it’s usually off-putting running into it elsewhere. Also, in terms of social mores, it’s considered rude to do so on internet message boards.
As noted earlier, you conflated the issue of mixing letters with the legibility of capital vs. lowercase letters. Other posters have already discussed the mixed-letters study, but in case you were wondering about capitals vs. lowercase, type in all-caps has been shown to reduce reading speed by about 13-20 percent. Check out the excellent book Dynamics in Document Design by Karen Schriver for more info about typography and document design.
I can’t find any of the debunkings I read on this topic earlier, but as usual for email forwards and the like, there’s certainly not an actual university study cited in the mixed-up text, and a lot of what it says is quite untrue. More is known about the process of letter recognition than used to be, and it’s a much more complicated subject than most people imagine. Further, the text used in this paragraph is not very thoroughly mixed; take the same text, mix each word properly (leaving, of course, the first and last letters in place), and it becomes MUCH harder to read. I’ll look around later for the hard-to-read version I found.
Yeah, but how much prose is there in a comic book? You’re not generally reading entire paragraphs. I can parse “KAPOW” okay, too, but the “I have a dream” speech presents a greater challenge when in all caps.
Check out the link I posted above (a page by a cognition scientist at Cambridge). It suggests some properties that a word should have if its scrambling is to be easily readable. It shows some examples of difficult-to-read scramblings, such as:
There was no study done at Cambridge (didn’t read my own link closely enough), but there was a Ph.D. thesis by Graham Rawlinson in 1976 that showed this exact phenomenon. Rawlinson wrote a letter to New Scientist about his work.
Define “paying more attention to”. It’s harder to tell capital letters apart from each other, so you need to devote more attentional resources to figuring out what the letters and words are, which slows down your reading speed.
If you mean, do the readers take more time because they think all caps make the text more important, then no, that’s not a factor in the results.
It does depend on the writer, which comic strip/book and to some extent the skill of the letterer. Chris Claremont, Berklely Breathed, George Herriman, Bill Foster, Kevin Smith, early Stan Lee, Bill Watterson, Brian Michael Bendis, Dave Sim and Christopher Priest are writers whose works characteristically dump chunks of all-caps dialogue and expostion at you. I’m as comfortable reading Alan Moore’s PROMETHEA as I am FROM HELL. Even though the latter has traditional capitalization but IMO the lettering is harder to read because of the chosen font size.
I don’t think so. I’ve seen readability studies comparing reading speed in text samples that were mixed case, all lower case, and all upper case. These studies used tasks designed so the participants would pay the same attention to all samples and there were specific instructions to counter any “gee, this is all caps it must be important” reaction. The studies found that reading all-caps text was significantly slower than mixed case or all lower case because it eliminates the whole-word processing which we use to quickly read most words. All caps forces you to pay attention to individual letters, which is not the same thing as paying more attention to the content of the all-caps text.
The studies were motivated by US military use of all-capped text in almost all electronic communication. In this real-world case, people do have to pay more attention to the all-capped text because it’s so hard to read, and productivity would go up considerably if they would switch to mixed case. Unfortunately, I can’t find an online link to these studies. I’ll keep looking.
So why is it the case that all caps text is hard to read? Is it just a matter of what we’re used to? Or is it the smaller differences between the letters themselves? It’s been shown that the idea that we read by examining mostly the overall shape of the word is not true - each letter is examined (though the process is quite complicated) and it seems that with training, comprehension of all caps text shouldn’t pose much of a problem.
Can you provide a cite? The studies I’ve seen have indicated that there are several active strategies and whole-word recognition is one of them. If whole-word recognition succeeds, individual letter discrimination is not necessary. If whole-word recognition fails, then letter chunks and eventually individual letters are attended to. These finding are in line with our general intuition since the brain is very good at pattern matching and there’s no reason it wouldn’t use word-shape patterns as a reading strategy. The studies I’ve seen have been focused on short-text readability and target recognition, and they’re technically visual search tests, not simple readability, so we may be comparing apples to oranges. In any case, I’d be interested in the studies which showed whole-word recognition does not occur.