How well does a spreading activation mechanism account for the auditory family size effect?

Words with more morphologically related words are recognized more quickly. We investigated the cognitive mechanisms underlying these family size effects in auditory word recognition, by studying how well they are accounted for by different versions of DIANA, a model of auditory word recognition. First, we found that part of the family size effects can be accounted for without incorporating morphological processing. This raises the question of what drives this part of family size effects. Second, we found that morphological processing in the form of the spreading of activation between morphologically related words does not account for auditory family size effects, at least not how we implemented this spreading. Third, we found that discriminative learning accounts somewhat better for family size effects, except for those that rely on phonological similarity. Together these results suggest that family size effects are more driven by phonological properties of the words than is commonly assumed. This Data Sharing Collection (DSC) includes: a) DIANA-based predictions (with and without spreading activation) of reaction times (RTs) in BALDEY, b) LDL-AURIS-based predictions of reaction times (RTs) in BALDEY. c) A BALDEY dataset enriched with three different family size (FS) measures. d) Control variables used in the associated experiment. e) All necessary scripts to derive the above materials Please note that the collection does not include the CELEX or CGN databases used for computing FS measures and training the LDL-AURIS model, as we do not have the license to share them. Researchers with access to CELEX and CGN can use the provided scripts to recreate the LDL-AURIS model and FS measures. Those without access will need to use the enriched materials provided in this collection.