Abstract
There is a great deal of evidence supporting the idea that, when a stimulus is processed fluently, it is more likely to be judged as pleasant. However, this influence of fluency on preference judgement seems to depend on several experimental conditions. So we tried to better understand these conditions via a comparison with recognition and by manipulating some aspects of the procedure (test format) and material (similarity and figure–ground contrast of the stimuli). Two experiments showed that some conditions maximally induce the use of processing fluency in a preference judgement, as in a recognition task. We discuss the implications of these findings for the well-documented discrepancy-attribution hypothesis (Whittlesea & Williams, Citation1998, Citation2000).
Notes
1In addition, there is a phenomenon known as the “test-pair [0]similarity effect”, which shows that YN recognition accuracy decreases with increasing target/lure similarity, but 2AFC recognition accuracy is not affected or increases slightly with similarity (e.g., Hintzman, Citation2001; Norman & O'Reilly, Citation2003). This pattern can be explained by data showing that familiarity contributes more to correct recognition in 2AFC than in YN tasks (Bastin & Van der Linden, Citation2003; Parkin, Yeomans, & Bindschaedler, Citation1994) and by data showing that familiarity is less strongly affected by target/lure similarity than recollection is (Norman & O'Reilly, Citation2003).
2For YN tasks, the d' values were obtained from hits (liked target) and false alarms (liked distractor), and for the 2AFC tasks, they were computed from the proportion of correct responses by means of a corrected formula (Macmillan & Creelman, Citation1991).
3In addition, these findings provide evidence for the “test-pair similarity effect”, which shows that YN recognition accuracy decreases with increasing target/lure similarity, whereas 2AFC accuracy increases slightly or is not affected by similarity (e.g., Hintzman, Citation2001; Norman & O'Reilly, Citation2003).