Type readability and empirical testing

polymeris's picture

Hello everyone,
I am an amateur typographist, enthusiastical but not very experienced. I'm pleased to find this forum, in hope I may learn a little.

I have an hypothesis about word readability I have been developing, and putting into practice creating a variant of a standard typeface. I am almost done with that. But before I go into details and publish either the modified type or my assumptions, I'd like to empirically test them!

Do you have any ideas how one would go about that? Any established methods?

I figured I could put up a webpage that would flash words taken randomly from a dictionary file, rendered in either the original type or the variation, and then queries people to see if they can remember the words. See is if there is any significant difference... What do you think?

Just to clarify, with "word readability", I mean the ability to quickly recognize words, abbreviations, etc. (but not long sentences) written in the type, in a setting e.g. on the road, where time matters and conditions may not always be ideal. Not books or small type.

Thanks in advance for your comments!

Nick Shinn's picture

You should investigate the research that has been done along these lines before reinventing the wheel :-)

Two Typophile contributors who have made significant contributions: Kevin Larson, cognitive psychologist with Microsoft, and James Montalbano of Terminal Design, creator of the Clearview typeface. Get googling...

polymeris's picture

Thanks for the pointers, Nick. Very interesting reads!

William Berkson's picture

There is lots of discussion of this over the years on Typophile. Here is one thread from a year ago. Also if you look at the posts of typophilers Kevin Larson and Peter Enneson, most of their threads relate to readability. Also, here is my account of my own effort to maximize reading comfort in a text face.

Nick Shinn's picture

...my own effort to maximize reading comfort..

Bill, we all do that.
In fact, your approach to Caslon was very similar to mine re. Scotch Modern, a genre which you consider to have poor readability.
I don't believe our approach was informed by scientific empiricism or scientific readability theory in any way (as opposed to long established theories of readability held by the typographic trade), it's just post-facto rationalization of what is quite simply a type revival done according to one's personal taste, facilitated and informed by the state of the art at a particular point in time -- that is to say, what is technically and economically feasible.

Did you do empirical testing during font development?
AFAIK, the only legit type designers who have are James Montalbano (Clearview) and Jason Smith (Mencap).
However, I'm not sure how truly scientific Smith's work was, or whether it was more along the lines of focus group research.


I recently came across some interesting paragraphs in The Shocking History of Advertising (1952) by E.S. Turner, set in Times Roman BTW.

"The leading writer on psychology in advertising (which was taken rather more seriously in America) was Professor Walter Dill Scott, Director of the Psychological Laboratory at Northwestern University, Chicago. His book The Psychology of Advertising was issued in Britain in 1909.

"In the previous century psychology had been an abstract study, couched in intimidating jargon. In order that practical use might be made of its conclusions, 'psychological laboratories' like the one at Northwestern University had been set up, and those psychologists who could pass on their knowledge in simple terms had begun to address advertising clubs. Many of the truths they announced had been discovered, empirically, generations ago. It is open to argument whether the psychologist gave the advertiser new ideas, or whether he merely told the advertiser which of his ideas were good ones (which the advertiser probably knew already)."

Note that the author equates empiricism with general practice, rather than laboratory experiment!

In assessing the relevance of readability science, and attempts to use it to prescribe styles of typeface or typography, it should be borne in mind that similar attempts to influence behavior -- which is the true nature of the agenda -- "better advertising through science", have been going on for over 100 years, with extremely limited success. A little awareness of the cultural nature of the endeavor might perhaps give some pause to any belief that "readability science" is of any practical use at all in typography, or ever will be.

Consider the sitability of chairs.
The sleepability of beds.
The driveability of cars.
The walkability of shoes.
The wearability of shirts.
The memorability of names.
The lookability of makeup.
The cookability of vegetables.
The drinkability of beverages.
"Readability" is not a scientific term.

William Berkson's picture

I am on the trail of finding who first distinguished between readability and legibility, and I think it was a scientist, but I have yet to confirm that.

If we define legibility as the ability to distinguish one letter from another quickly in isolation, and readability as the ability to read whole words quickly, comfortably, and with good comprehension, I think the scientific validity of the distinction is shown by the first visual in the article I linked to, among other things. There a word with individually highly legible letters is very difficult to read.

As I explained in the article, the only experimental testing in developing the font was with my own eyes. I agree with you about the poor results of research so far, but I am not dismissive of the promise of future research efforts, as you are. Based on the success of scientific methods in other fields, I am optimistic that such research will eventually bear fruit.

I don't see how our processes could have been identical, since we bring standards that are in a few points in conflict, and those differences show in the final results.

Nick Shinn's picture

As I said, it's a question of personal taste, which we share.
These faces are about the love of old letterpress, and trying to port that quality into the digital realm.

Here I compare Scotch Modern with a typical digital rendering of the didone style.
Like you, I have not simplified and cleaned up the glyph, but attempted to incorporate the trace of the letterpress process, with a rounded finish, not square cut. Also, I have kept the horizontal flow I saw in the original. And varied the width of hairlines/serifs.

The micro-detailing we have employed is a contemporary, practical method now, with pre-press and press gain being so small in the digital, high-res (offset lithography) workflow.

As "independent" designers, these face of ours are not typical, in the way that, say, Adobe fonts of these genres are geared towards a broad corporate market that company represents, or Font Bureau types of these genres target the periodical market. (Even though FB publishes your Caslon.)

I would say these are the deciding factors in our common approach, not anything we've read about reading theory.

William Berkson's picture

Nick, we both were concerned with ink spread, but that was by no means my only concern.

You write, "It's a question of personal taste."

My decisions were not only a question of taste, but also of my being guided by my theories of readability, as partly explained in the article. There were developed by reading, reflection, discussion with others, looking at type, and drawing type and testing it.

My theories are not the same as yours, and so my process somewhat different as well. It's not only a matter of taste.

Kevin Larson's picture

Sofie Beier used empirical measures to test the letter recognition rates of different letter designs in three different typefaces. The typefaces were all new designs, and the empirical data was used to create the final designs. Two methods were used to reach her findings, a distance threshold (as with Clearview Hwy) and a time threshold.

Beier, S. & Larson, K. (2010). Design Improvements for Frequently Misrecognized Letters. Information Design Journal, 18(2), 118-137.

Chris Dean's picture

@polymeris: “I have an hypothesis about word readability I have been developing…”

What is your hypothesis?

Taking a new study like the one Kevin mentioned and looking at it's reference section is a great way to start a literature review. An alternative strategy, if you or any of your colleagues have access to the Web of Science Citation Database is to find an old study of relatively high profile and create a citation map. This shows you every article that referenced it in the future. By using both methods you end up attacking the problem from two sides at once. One of the older studies I have come across:

Roethlein, B. (1912). The relative legibility of different faces of printing types. The American Journal of Psychology, 23(1)l, 1–34.

And while many of his studies are confounded, Miles Tinker is also a good place to start as most of his writing is an easy read.

See also:

Garvey, M., Pietrucha, T. & Meeker, D. (1997). Effects of font and capitalization on legibility of guide signs. Research on Traffic Control Services: Transportation Research Record-series, 1605, 73–79.

Zineddin, A., Garvey, R. & Pietrucha, M. (2003). Effects of practice on font legibility. Proceedings of the Human Factors and Ergonomics Society 47th annual meeting.

Zineddin, A. (2002). The impact of familiarity and word-superiority effect on large format text applications. Unpublished dissertation. Pennsylvania State University.

I believe Zineddin still works with the Federal Highway Administration in Washington DC, but it has been some time since we have spoken. He is very personable and was more than willing to share his work with me.

Good luck.

Kevin Larson's picture

I’m a big fan of that Roethlein paper. The amount of raw data in the paper is fantastic, and just doesn’t happen anymore. I think it would a lot of fun to replicate that test 100 years later (now) to see if letter legibility has changed in that time.

Syndicate content Syndicate content