Letter Similarity Data Set Archive

This web site archives data from many published experiments on letter similarity. It is presents stimuli and data from many different attempts to capture the similarity structure of the alphabet. Not all identified data sets have been transcribed and placed in this on-line archive. References for complete list are available here. An in-progress paper on these matters is available on request.

Summary of experiments reporting letter similarity matrices.
Reference Method Case Typeface
Cattell, 1886 Naming errors B Latin serif
Naming errors B Fraktur
Sanford, 1888 Naming errors of distant stimuli L Snellen
Naming errors of brief stimuli L Snellen
Naming errors of brief stimuli L Old-style Snellen
Dockeray, 1910 Naming errors of stimuli in periphery L 10-pt. Roman old-style
Roethlein, 1912 Confusable letter sets B 16 different fonts
Hartridge, 1922 Naming of distant stimuli U Green's Letter Set
Korte, 1923 Naming of distant stimuli B Antiqua
Naming of distant stimuli B Fraktur
Banister, 1927 Naming of distant brief stimuli U Green's Letter Set
Naming of distant brief stimuli U Green's Letter Set
Naming of distant stimuli U Green's Letter Set
Tinker, 1928 Naming of brief stimuli B Bold serif font
Ovink, 1938 Naming of distant stimuli B 11 different fonts
Hodge, 1962 Letter reading errors B Uniform-stroke alphabet
Gibson et al, 1963 Children's matching of target to set of random choices U Sign-typewriter
Children's matching of target to set of similar or dissimilar letters U Sign-typewriter
Popp, 1964 Forced choice confusions of children L Century-style
Bell, 1967 Naming errors of brief stimuli U Long Gothic
Naming errors of brief stimuli L Murray
Dunn-Rankin, 1968 Similarity preference of letter pairs L Century Schoolbook
Dunn-Rankin et al, 1968 *Shape congruency L Century Schoolbook
Kuennapas, 1969 Subjective similarity ratings L Sans serif Swedish alphabet
Uttal, 1969 Naming errors of brief masked stimuli U 5x7 dot matrix
Laughery, 1969 *Feature Analysis U Roman block letters
Gibson, 1969 *Feature Analysis U Roman block letters
Fisher, et al., 1969 Naming errors of 200-ms stimuli U Futura medium
Naming errors of 400-ms stimuli U Futura medium
Naming errors of brief stimuli U Leroy lettering set
Townsend, 1971a Naming errors of brief unmasked stimuli U Typewriter font
Naming errors of brief masked stimuli U Typewriter font
Townsend, 1971b Naming errors of brief unmasked stimuli U Typewriter font
Bouma, 1971 Naming errors of distant stimuli L Courier
Naming errors of stimuli in periphery L Courier
Geyer, 1973 *Feature analysis U Roman block letters
Engel, et al., 1973 Naming errors of brief stimuli L Century Schoolbook
Loomis, 1974 Tactile letter identification U 18x13 matrix
Briggs, 1975 *Feature analysis U Roman block letters
Mayzner, 1975 Naming errors of brief stimuli U 5x7 dot matrix
Thorson, 1976 *Overlap values based on feature analysis U Roman block letters
Geyer, 1977 Naming errors of brief dim stimuli L Tactype Futura demi 5452
Coffin, 1978 *Fourier spectra similarity U 128x128-pixel block letters
Podgorny and Garner, 1979 Same-different choice RT U 5x7 Dot matrix chars.
Subjective similarity ratings U 5x7 Dot matrix chars.
Gilmore, et al., 1979 Naming errors of brief stimuli U 5x7 Dot matrix chars.
Kikuchu, et al., 1979 Tactile letter identification U 17x17 Dot matrix chars.
Craig, 1979 Tactile letter identification U 6x18 Dot matrix chars.
Keren and Beggen, 1981 *Feature analysis U 5x7 Dot matrix chars.
Johnson & Phillips, 1981 Tactile letter identification U Sans serif embossed letters
Geyer & Gupta, 1981 Naming Confusions for nine letters U Three fonts (dots, filled, stroked)
Loomis, 1982 Visual identification U Blurred Helvetica
Tactual identification U Helvetica
Paap, et al., 1982 Naming errors U Terak
Gupta, et al., 1983 Naming errors of brief dim stimuli U 5x7 Dot matrix chars.
Naming errors of brief dim stimuli U Keepsake
Phillips, et al., 1983 Naming errors of small visual stimuli U Helvetica
Tactile identification U Sans serif
Gervais, et al., 1984 Naming errors of brief stimuli U Helvetica
*Similarity of spatial frequency spectra U Helvetica
VanNes, et al., 1983 Naming errors of brief peripheral stimuli L 12x10 pixel matrix-least confusable (IPO-Normal)
Naming errors of brief peripheral stimuli L 12x10 pixel matrix-most confusable
VanderHeijden, et al., 1984 Naming errors of brief stimuli U Sans serif roman
Blough, 1985 Pigeon's 2-alternative letter matching U 5x7 dot matrix
Blommaert, 1988 *Fourier spectra similarity L 16x32 pixel matrix courier
Heiser, 1988 *Choice model analysis of confusions U Sans serif roman
Jacobs, et al., 1989 Saccade times to matching target L 9x10 pixel matrix
Saccade errors to distractor L 9x10 pixel matrix
Boles and Clifford, 1989 Subjective similarity ratings B Apple-Psych letters
Courrieu, 1989 Children identifying targets that matched uppercase reference L Printed script
Watson, 1989 Naming errors of low-contrast stimuli U 5x9 pixel font (gacha.r.7)
McGraw, et al., 1994 Letter identification with keyboard L ``Gridfont'' chars.
Reich, 2000 Naming of tiny or peripheral letters U Sloan Letters
Liu & Arditi, 2001 Naming of tiny crowded or spaced letter strings U Sloan Letters
Mueller et al. (in prep) Forced choice identification of letter with distractor: @-mask U Courier
Forced choice response latencies: @-mask U Courier
Forced choice identification of letter with distractor: #-mask U Courier
Forced choice response latencies: #-mask U Courier
Note. In the Case column, ``L'' indicates lowercase, ``U'' indicates uppercase, and ``B'' indicates both cases were studied. Methods denoted with an * were measures developed by analyzing the visual form of letters, and not directly based on empirical data.
This archive was created and is maintained by Shane T. Mueller, Ph.D., who can be reached in regards to this archive at [stmuelle at obereed dot net].