The effects of visual similarity on proofreading for misspellings

Abstract

In three experiments, subjects read prose passages and circled misspellings in them. The misspelled words were created by replacing a single letter with another one. The visual similarity between the correct letter and the one that was substituted for it increased the percentage of proofreading errors. The results suggest that proofreaders search through a visual representation of the text and that a hierarchical feature test is applied to this representation, according to which subjects give first priority to resolving letter envelope and second priority to discriminating additional visual features. A sophisticated-guessing decision rule is also implicated for the misspellings that do not alter letter envelope: Subjects are tolerant of missing letter features but are intolerant of additional features.When individuals proofread for misspellings, do they search a phonetic representation of the text or a visual one? Several studies have provided evidence for the role of phonetic recoding in proofreading tasks. For example, Corcoran (1967) conducted a proofreading experiment in which letters were omitted from various words in a prose passage and subjects were asked to indicate the locations at which letters were missing. As in his earlier study of letter cancellation (Corcoran, 1966), Corcoran found that subjects made more errors on silent es than on pronounced es in the proofreading task. Similarly, Corcoran and Weening (1968) found that subjects made more errors in a proofreading task on silent ks than on pronounced ks. In another proofreading study by MacKay (1968), subjects were asked to read sentences containing misspelled words and to indicate the words in which they perceived a misspelling. Two types of misspellings were introduced: phonetically compatible ones, which would be pronounced in the same way as the original word (e.g., "work" misspelled as ''werk''), and phonetically incompatible ones (e.g., ''work'' misspelled as ''wark''). Phonetically incompatible misspellings were noticed more frequently than compatible ones. (See Cohen, 1980 On the other hand, several more recent studies have demonstrated the importance of visual factors in proofreading. Holbrook (1978) found effects of the visual similarity between the correctly spelled and misspelled words in a proofreading task. Likewise, Haber and Schindler (1981) have reported that misspellings are easier to detect when the correctly spelled and misspelled words differ in shape than when they are similar in shape.One goal of the present study was to determine the nature of the text representation that is searched by proofreaders. The second aim was to investigate the search process in more detail, in order to determine the nature of the information processing strategies, or rules, employed by proofreaders. EXPERIMENT 1In the first experiment, the type of the to-be-detected misspelling was systematically manipulated. Misspelled words were created by substituting a single letter in a word with another one. (A similar "substitutionproofreading" task was u...

The effects of perceptual condition on proofreading for misspellings

“…These information processing rules were supported by two related findings in the study by Healy (1981): (1) Subjects searching for misspelled words made many more proofreading errors when a correct letter was replaced by another letter that had the same visual features except for one that was missing (as when e was replaced by c) than when the letter substitution was made in the reverse order (as when c was replaced bye).…”

“…Healy (1981, p. 459) noted that her results provide strong evidence for a hierarchy of levels of discrimination but do not provide any information concerning the temporal sequence in which those levels are processed. In accord with the work of Lupker (1979), it seems likely that subjects first resolve letter envelope and then process additional features, but it is also possible, given Healy's (1981) findings, that the two levels of discrimination are processed in parallel, with more weight given to those features defining envelope than to other features. If the envelope of the letter conforms to their expectations, the proofreaders adopt a sophisticated-guessing decision rule in which they tolerate the absence of letter features but are intolerant of additional features.…”

“…One reason why it is important to test for the influence of perceptual condition on processing strategy is that a number of previous investigations have yielded evidence inconsistent with the sophisticated-guessing decision rule, and those investigations used perceptual conditions that differed from the ones used in the proofreading study by Healy (1981). For example, in a tachistoscopic letter-recognition experiment, Lindsay and Norman (1977, p. 267) found more occasions when C was reported as G than when G was reported as C, in accord with sophisticated guessing, but they did not find, for example, that 0 was reported as Q more than Q was reported as O.…”

“…In fact, in an analysis of a tachistoscopic letter-recognition experiment, Massaro and Schmuller corrected for possible response bias by computing a d' index and found that the asymmetries previously evident in the confusion matrix were thereby eliminated. Healy (1981) controlled for response bias in her study by comparing passages typed in lowercase and uppercase characters. Since the asymmetries she observed were typically much larger in one typecase than in the other (i.e., the asymmetry for c and e was greater in lowercase, whereas that for c and g was greater in uppercase), they cannot be attributed solely to the subjects' biases to report specific letters, and therefore they can be taken as support for sophisticated guessing.…”

Demonstrations of the Ease of Detecting Additions over Deletions

“…Although not an absolute fi nding in all cases, many studies have found a cognitive bias in favour of processing positive information in diverse areas of psychology. For example, visual search is easier when detecting a feature's presence rather than absence (Neisser, 1963;Treisman and Gormican, 1988;Treisman and Souther, 1985), proofreading is easier when detecting added rather than deleted letters (Healy, 1981), conditional reasoning is more likely to include searches for positive than for negative evidence (Klayman and Ha, 1989;Wason and Johnson-Laird, 1972), sentencepicture verifi cation is faster for positively-stated sentences than for negatively-stated sentences (Clark and Chase, 1972) and positive affective information is better retained in memory as people age than is negative information (Mather, 2003;Mather and Carstensen, 2003).…”