Automatic vigilance for negative words in lexical decision and naming: comment on Larsen, Mercer, and Balota (2006)

An automatic vigilance hypothesis states that humans preferentially attend to negative stimuli, and this attention to negative valence disrupts the processing of other stimulus properties. Thus, negative words typically elicit slower color naming, word naming, and lexical decisions than neutral or positive words. Larsen, Mercer, and Balota analyzed the stimuli from 32 published studies, and they found that word valence was confounded with several lexical factors known to affect word recognition. Indeed, with these lexical factors covaried out, Larsen et al. found no evidence of automatic vigilance. The authors report a more sensitive analysis of 1011 words. Results revealed a small but reliable valence effect, such that negative words (e.g., "shark") elicit slower lexical decisions and naming than positive words (e.g., "beach"). Moreover, the relation between valence and recognition was categorical rather than linear; the extremity of a word's valence did not affect its recognition. This valence effect was not attributable to word length, frequency, orthographic neighborhood size, contextual diversity, first phoneme, or arousal. Thus, the present analysis provides the most powerful demonstration of automatic vigilance to date.