Abstract
A long-lasting dispute within typography is whether serif or sans serif fonts are the most legible. However, several parameters affect font legibility; not just whether serifs are present or not. An experiment was conducted to test whether reading performance was more affected by serifs or by the contrast of the font's letter-stroke. We created four fonts, such that their stylistic features were isolated; serifs and letter-stroke contrast. Type of serif was manipulated to be serif or sans serif, letter-stoke contrast was manipulated to be low or high (i.e., 2 by 2, factorial design), and a lexical decision task was implemented. That is, participants were exposed to a pair of words, one after the other (separate intervals), and one word was always misspelled (swapped two of its middle letters). Participants indicated in which of the two intervals the correctly spelled word was placed (i.e., two-interval, forced-choice task). As a baseline condition, three lexical identification thresholds were obtained via the QUEST algorithm and the font Helvetica. These three thresholds were averaged and used to tailor each participant's stimuli intensities, which were 9 levels of font size. The method of constant stimuli was implemented where the font size varied each trial. There was a main effect of letter-stroke contrast such that low-contrast letter-stroke elicited lower lexical identification thresholds (M = 61.12 points), relative to those elicited by high-contrast letter-stroke (M = 63.49 points), F (1, 14) = 5.36, p = .036. Neither the main effect of type of serif nor its interaction with letter-stroke contrast were significant, which indicated that the serif parameter did not significantly influence lexical identification. These results lend support for an account through which letter-stroke with low contrast, and not serifs, facilitates observers' lexical identification thresholds.