Orthographic neighborhood size effects and associative recognition

Three experiments on the role of orthographic distinctiveness (as measured by neighborhood size [N]) in associative recognition are reported. A mirror effect was obtained, with high-N words receiving more hits and fewer false alarms than low-N words. This pattern was replicated in Experiment 2, where participants carried out a relational orienting task. However, the high-N advantage in hit rates was eliminated in Experiment 3 when subjects carried out an item-processing orienting task. The high-N advantage in associative recognition contrasts with the low-N advantage found in item recognition. This reversal of mirror effects between item and associative recognition is empirically similar to patterns found in studies of normative word frequency.