More on Linespacing

Typography Tip #3 recommends setting linespacing in Word to a multiple of 1.2. This has the effect of making each line of 10 point text 12 points tall. Is this a good recommendation?


Miles Tinker ran a huge research program investigating typographic variables from the 1920s to the 1950s. Linespacing is one of many variables he thoroughly investigated. He collected data with multiple sizes of text and multiple line lengths. His standard methodology measured people’s reading speed while simultaneously carrying out a comprehension task that ensured that the text was fully understood. This ensured that he was only measuring reading speed differences and not comprehension differences.

When examining reading speed with 10 point text, Tinker found little difference between reading speed with text set solid and adding 1 point of linespacing (multiple of 1.1). A statistically reliable speed advantage was found by adding 2 points of linespacing (multiple of 1.2). And adding more linespacing past 2 points did not further improve reading speed. 2 points of linespacing appears to be the critical amount of extra space needed to separate lines, as 2 points was also optimal for both 8 and 12 point text sizes.

Linespacing with 10 point text:

Amount of Additional Linespacing

Reading Speed Difference in Percent

Set Solid (control)


1 point


2 point


4 point



Bonus: In the comments for Typography Tip #3, Adam Twardoch asserts that the line length effects the amount of needed linespacing. Tinker’s data does not back up this assertion. This table shows that 2 points of linespacing performed the best at each line width tested.

Linespacing with 10 point text. Reading speed difference in percent compared to 19 pica line (3.2 inch) with 2 points of additional linespacing:

Line Width

Set Solid

1 point linespacing

2 point linespacing

4 point linespacing

9 pica (1.5 inch)





14 pica (2.3 inch)





19 pica (3.2 inch)



0.0 (control)


31 pica (5.2 inch)





43 pica (7.2 inch)





Cheers, Kevin Larson

Miles A. Tinker, Legibility of Print, Iowa State University Press, 1963.

Comments (17)

  1. John Hudson says:

    It doesn’t surprise me that the desirability of increased linespacing for longer line lengths is not reflected in reading speed tests. But there are both aesthetic and functional grounds for this practice. Aesthetically it makes the text look less cramped and better balanced. Functionally, I believe this openness makes it easier to scan the text, locate specific words, etc. — what might be termed ‘text entry operations’ — which are important ways in which we engage with documents other than reading them through from start to finish.

  2. Hi,

    Great post, love those numbers. I hadn’t seen the actual results of such studies before.

    To get a better feel for what the numbers were actually saying, I put together a PDF document showing each of the layouts used in the study. It’s gratifying to find that the layouts with poor-reading speeds do look uglier!

    I’ve made the PDF available here, with some further comments:



  3. Will Robertson says:

    The commenting system encoded that closing > as part of the URL above. Oops. Just delete the "%3E" to get to the right place.

  4. Kevin Larson says:

    John, I agree that more leading improves the aesthetics of longer lines. I expected a closer match between aesthetics and reading performance.

    Thanks for the visualization Will!

  5. SirPavlova says:

    Any chance that we could see some comparative data between the 8, 10, & 12 pt sizes mentioned? I’d be interested to see which is optimal, & by how much.

  6. Urm, I hate to point this out, but on average, Word’s "single spacing" already has the equivalent of 20% leading. It is not the same as being set solid.

    Worse, Word’s single and multiple spacing is simply not based the point size alone, but is based in part on the font bounding box. It has very little to do with the kinds of measurements Tinker was doing.

    I’ll post a more detailed explanation on my blog later today, with pictures and commentary.

  7. I was a typesetter some 20 years ago (electronic and Linotype) and 1.2 was almost always the specified spacing. Normal text sizes were 10/12 (10pt type, 12pt leading) and 12/14. this study only confirms what a lot of us already knew.

    However, I’m with Thomas (fix your URL!) concerning Word’s dependence on the bounding box, though this is really more the fault of the type designer than the OS.

  8. james kreines says:

    Sorry for the off topic comment, but I was still hoping that fontblog might go into more detail abotu whether there is a way to get cleartype to work with displays oriented in portrait mode. There is conflicting info out there on the web. I asked about it in response to this post:

    Kevin Larson then said in his response that Greg was going to go into some fascinating detail in response, but I never saw anything more from Kevin or Greg on this. Is there any worthwhile source for more information?


  9. Why have the posts stopped?  Whats going on?

  10. Ben Darlow says:

    I’m surprised that the recommendation only calls for a 120% type size linespacing. I’ve been using larger linespacing settings than this for some time in web content I publish using CSS (sadly IE6 doesn’t support the line-height attribute, but IE7 will remedy this). For a type size of 11px I’ll typically use 16px or even 18px as the line-height value. I’d say 120% is still way too cramped. I’ll have to check later, but I’m pretty sure Robert Bringhurst recommends a far larger linespacing value in ‘Elements of Typographic Style’.

  11. IEBlog says:

    For Beta 1, we discussed some of the technical improvements (like domain highlighting, multi-line paste,

  12. Ранее мы уже обсуждали некоторые из технических усовершенствований (выделение имени домена, многострочная

  13. fontblog : More on Linespacing

  14. &#160; Internet Explorer 8 Beta 1에서 주소 표시줄에 대해서 실시한 몇가지&#160; 기술적인 안내 (도메인 하이라이트 기능, 여러 행 붙이기 기능, 클릭