The effects of perceptual condition on proofreading for misspellings

In two experiments, subjects proofread text in which misspelled words were created by replacing a single letter with another one. The following perceptual conditions were included: clear-cut photocopies of typewritten text, clear-cut text presented on a CRT screen, ditto copies that varied in legibility, and text that included extraneous noise characters superimposed on some letters. In all conditions, subjects adopted a hierarchical feature test that gave first priority to resolving letter envelope and second priority to discriminating other letter features. When clear-cut text with no extraneous noise was proofread, subjects used a sophisticated-guessing decision rule that tolerated misspellings involving missing letter features (as when c replaced e) but not added ones (as when e replaced c). This asymmetrical rule was modified, however, when subjects were exposed to text that included extraneous noise that was confusable with the letter features. In those circumstances, subjects adopted a decision rule that tolerated misspellings involving missing features or any added features that resembled the noise.