Mapping effects in choice-response and go/no-go variants of the lexical decision task: A case for polarity correspondence

Previous research has shown that responses to words are faster and more accurate in the go/no-go version of the lexical decision task (LDT) than in the choice-response version. This finding has been attributed to reduced response-selection demands in the go/no-go task. Here, we test an alternative account assuming similar response-selection demands in the two tasks, but an additional impact of polarity correspondence between stimuli and responses in the go/no-go task. Positive and negative polarities have been described as a frequent characteristic of binary stimulus and response dimensions. Only for the go/no-go version of the LDT, an apparent polarity difference between go and no-go responses exists, with go responses having the same polarity as words and no-go responses the same polarity as nonwords. Thus, compared with the choice-response LDT, in the go/no-go LDT, go responses to words should be facilitated by polarity correspondence, and go responses to nonwords should be inhibited by polarity noncorrespondence. In this study, each participant performed a go/no-go LDT and a choice-response LDT. Responses to words were faster and more accurate than responses to nonwords, and-consistent with the alternative account-this difference was larger in the go/no-go LDT than in the choice-response LDT. An analysis of performance by means of sequential-sampling models that take into account a decaying influence of irrelevant stimulus features supported the effect of polarity correspondence in the go/no-go LDT. This analysis suggested an effect in the choice-response LDT as well, though of a smaller size and a faster decay.