What's wrong here? Where to begin ...

A: There is neither a premise nor an indicated purpose.

B: As a research project/experiment, there are absolutely no controls.

C: There are built-in biases:
1) Any given physical format such as the 8½ x 11 that you arrogantly call 'letter size' biases against those that don't have access to that size. For most of the rest of the world, A4 is 'letter size'.
2) Each story in its own font? The content will have more of an impact than the font selection, and there may be a question of appropriateness of a given font to a given narrative.
3) Sampling bias: You're asking people to PRINT OUT this material? Only those willing and able to do so would therefore be counted; in other words, interested parties with time on their hands and a willingness to spend money on ink and paper.
4) Sample size. For any results to be meaningful, there would have to be an adequate sample size.

And more.

This would only be very interesting, I'm afraid, in reference to a specific design of a specific text (on specific stock, in a specific format ...) — in other words, the kind of choices that typographers make every time they start a project: what font is best for this book? If you wanted to design a text setting with a few samples and ask people which made for more comfortable reading, I think you would get some responses — but that wouldn't be an experiment anymore, just an everyday design decision.

This would only be very interesting, I'm afraid, in reference to a specific design of a specific text (on specific stock, in a specific format ...) — in other words, the kind of choices that typographers make every time they start a project: what font is best for this book?

Which is why I specified 8.5 x 11, common white paper.
The point size of the text is debatable. 13?

Yes, but the readability of it depends on the relationship of the content of the text to the typeface, as well as the layout of the page. Does the text have a lot of capitals or numbers or italics or paragraph breaks? What is the subject matter? How wide is the column? What is the leading? And, since all the typefaces you've chosen are indeed serious, readable faces that could be used for substantial texts, the more important question becomes whether it's appropriate to the tone and content of whatever you're typesetting. What I'm trying to say is that I don't think tests of "readability" have much relevance or interest outside of a specific project. All these faces are readable enough that what is more likely to have an effect on the reader is their appropriateness — which depends on artistry and history, not science.

How many works of fiction have you ever read in your life that were set 13pt on letter size paper? If the point of this test is to look at sans serif types in the context of long texts of the kind normally set in serif types, surely the first thing to control are all the other aspects of typography normally associated with such texts. That means either a smaller format akin to a book or a two-column layout akin to a magazine, and it definitely means a smaller type size than you suggest.

Why voting on a shortlist? That isn't even remotely scientific. If you want to limit the number of types considered, you need to come up with reasons for the selection that are meaningful to the exercise, e.g. types with particular different characteristics.

With regard to the problem of the text itself influencing readability, this may be a real issue given your proposed text selection method. If using different texts, one wants to be able to ensure similar levels of syntactical complexity, vocabulary, etc., which is difficult unless one has a source for rated texts. Picking random 'entertaining public domain stories' won't provide good controls. Here's a better idea, I think: pick a single story of reasonable length and use a different font on each page. Produce a set of numbered PDFs with different fonts used on different pages, and ask participants to pick one at random (or assign them one at random) and record which they read. The idea here is that you avoid e.g. the same font always being used for the most exciting and riveting part of the story.

Science is hard.

hhp

And, ideally, give this test to people who aren't typographers.

I don't think the average person will cut a letter sized piece of paper down after printing, or even look for book sized paper at the stationary store to print on. But they likely will have a printer and some 8.5 x 11" paper lying around.

In regard to subject matter and content influencing comfortability, I totally agree. However, it would be easy to use php to serve a given font in a random choice from the four texts.

Or perhaps a more elegant way to do it, would be to simply use only Sherlock Holmes short stories?

There's hundreds of them available in the public domain, right?

All you need do is go to Project Gutenberg to pick up a few texts. But whether I can be bothered to special-order 8½×11″ paper is a different Sheila.

"Special-order?"

It may come as news to many North Americans that in most of the world 8½×11″ is not generally available. A4 is the standard international paper size that is closest in dimensions to US Letter size, but it is narrower and taller (and much more nicely proportioned to my mind).

But it's not narrower enough.

hhp

Given typical type size, leading and margins for correspondence, and considering this in terms of number of words on a line, A4 seems to me a very good size and quite narrow enough.

I like the two column idea for sure.

@Ryan Maelhorn:
Or perhaps a more elegant way to do it, would be to simply use only Sherlock Holmes short stories?

There's hundreds of them available in the public domain, right?

At least until recently, they were only public domain in the U.S., but still protected by copyright in most of the rest of the world.

@Joshua – The standard letter format in Europe (and, yes, I am a Native European) is A4. One scant-noticed thing about the A formats is that is that you can cut one A(n) sheet into exactly two A(n+1) sheets since the ratio between height and width is sqrt(2).

Useful pages:
http://en.wikipedia.org/wiki/Paper_size
http://en.wikipedia.org/wiki/ISO_216

Ah, right, I knew this; I just wasn't thinking. The A sizes work according to the golden ratio, the reason that half of one sheet is the next size down, which does make them more attractive and more sensible than the American sizes.

B5 > A4.

hhp

The A sizes work according to the golden ratio

No. As Reynir already mentioned, they’re based on the "square root of 2" ratio: 1/1.41

B5 > A4

Actually, no:

I meant it's better. :-)

hhp

A4 is really pretty close to Letter. We could probably just set one version for both. You know make sure the "text body" (not up on my book setting terminology) is the same size on both A4 and Letter. So there would have to be slightly different margins. This would save us from having to re set the type for both versions.

Also, no one has anything to say about my original contenders? Anyone agree, disagree? Any other suggestions to try? I was thinking about Univers, but decided most of the other typefaces listed have a sort of 'flare' quality to them. (think albertus). I'd love to try one of Hubert Jocham's fonts, like Bent, Spring Sans, or TeleVoice; but I don't own them.

Meanwhile I'm trying to find good texts.

.. good texts ..

Well then, why not select typefaces that have 'text' in their name.
Naturally that would include a great many blackletter and calligraphic fonts.
What's in a name, anyway.

Ryan: A4 is really pretty close to Letter. We could probably just set one version for both.

When I need to prepare documents that will be printed by clients in countries with different standard paper sizes, I use a custom page size in InDesign that has the width of A4 and the length of US Letter, i.e. the smaller of the dimensions of the two papers. This ensure that everyone can print the document without the print area being trimmed or any margins being uncomfortably small.

Ryan,

Set up your pages half-size of whatever "standard" you want, because the finished page will be closer to "ideal" book size. Then, generate a PDF file.

Next, use a very useful program like Quite Imposing to set up a print run which will allow you simply to chop the finished output in half, then place one half on top of the other. Bind same in a manner you see fit. Nothing could be simpler…really.

I have a commercial proposal for a franchise operation called Bibliophile Junction, which allows for on-demand printing from a rights-managed library, or as an adjunct service for eBook authors. Some of us old-timers simply do not like eReaders. OTOH, we will probably all be dead soon enough, and thus will millions of trees be saved…

Trial 1:

(I make no claims as to being a book setter)

http://nonbookfonts.com/xtra/a_gentleman_vagabond_chap1_font1.pdf

http://nonbookfonts.com/xtra/a_gentleman_vagabond_chap1_font2.pdf

http://nonbookfonts.com/xtra/a_gentleman_vagabond_chap1_font3.pdf

http://nonbookfonts.com/xtra/a_gentleman_vagabond_chap1_font4.pdf

I also want to make clear that I dont intend this to be an absolutly scientific test that proves beyond a shadow of a doubt which sans font is best for long text setting. But rather to raise discussion and practical thinking about the matter, and to give a few readily available examples to draw from.

As this moves on, hopefully with the help of others here, I will try to set these texts in a better, more comfortable to read manner.

Thank you to all who participate.

This is full of typos and typographical oddities that will probably have a greater impact on fluency of reading than the typeface.

To me Kievit and Caspari do a better job but both need more leading. I would also suggest increasing the top margin. I'm not sure about the typeface body size but 13 is quite big for comfortable reading. If I remember correctly Kievit wors quite well set 9.5/11 as a body copy face.

Regarding my margin suggestion. It's better to increase the lower margin and to have one of the side margins larger than the other.

Yes, I will have to set a new text, something without these 'typographical oddities.' I will increase the leading, and the top and bottom margin. Not sure if there would be any reason to have unequal side margins though, as this isn't going to be bound like a book.

Assimetry is better for visual interest than perfect simmetry.

Symmetry is a lost opportunity.™

hhp

So, is this the only text, or is there one for each font? for each user? for each size?

Just the 1 text in four fonts for now. Basically trying to come up with a format that makes the most people happy. Hopefully that will translate into the most amount of people reading it as possible.

I'm still open to suggestions for different fonts to try. The four that I used for the text so far are the four I think work best for long text reading, but I have been wrong before...

Caspari and Kievit are interesting to compare to eachother. At about 10 point (which is what all the texts are set at), to me they look fairly similar and seem to almost even have the same "mood." But when you blow them up and look at them in 200 point or so, you notice they are very different. Caspari is close to being mono-line, and all it's curve look pretty polished, whereas Kievit is not mono-line at all, and all it's lines and curves looks, well I can't come up with a better term than "hand drawn." Yet again, to me, at long text size, they feel remarkably similar to one another.

The Scala Sans sample I find the most interesting, however. It has such an "official," almost "clinical" appearance. Might be a bit too "formal" looking for casual reading (long text reading for pleasure, as opposed to research, education, etc...), but there is something very pleasing and "right looking" about it for body text, for say, a new prescription drug brochure.

Legato seems to take on very stoic, serious tones when set in long copy. I would love to see a copy of the bible set in Legato. As with Scala Sans, though, might be a bit too formal for long pleasure reading.

It's catchy. Opportunity for what?

.

@rs_donsata – An opportunity for putting a bit of tension and interest into it, is my guess.

The bad justification and hyphenation is killing these settings. I wouldn't read anything set this way, regardless of the typeface. Since you are using a two column layout, I suggest you go look at a well typeset magazine such as Lapham's Quarterly and try to understand what makes it good.

John is right, the typesetting is very buggy. You can improve it much by simply using hyphenation and by using some paragraph bleeding instead of full line separations. Avoid widows and orphans.