Letter Similarity Data Set Archive

Geyer & Gupta, 1981

Geyer, L. H., & Gupta, S. M. (1981). Recognition/confusion of dot matrix vs. conventional font capital letters. Perception & Psychophysics, 29, 280-282.



The authors used a subset of nine uppercase letters (E, F, I, L, T, H, M, N, & W) presented tachistoscopicly. Letters subtended .55% vertical angles, and duration was controlled for correct recognition across fonts of 50% (no other details given). There was apparently data from 24 participants, each of whom took part in 2 font conditions.


Three fonts were used: dots, "stroked" font letters with serifs, and "filled", which had letter forms identical to the dots, but the spaces between the dots were filled in. In the image below, the first line is the dots condition, the second line is the filled, and the third line is the stroked.


Transcribed confusion matrices for the three conditions are found here: Dots, Filled, Stroked. For each, rows sum to roughly 1.0, implying that each row indicates a presented stimulus, and each column represents a response. No guarantee about accurate transcription is implied.

Back to main Letter Similarity Data Set Archive Page