A Problem with Subjectively Rated Word Familiarity

: Part 1. Familiarity of Visually Presented Words

Abstract

We recently conducted a lexical decision experiment controlling various attributive values of stimulus words based on a popular Japanese database (Amano & Kondo, 2003), and found extremely high error rates for several stimulus words with relatively high familiarity. We thought that qualitative differences between raters in Amano and Kondo (2003) and the participants of our experiment might have caused this inconsistency. Accordingly, we conducted a lexical decision experiment, and then had participants rate their familiarity with the visually presented words. The mean familiarity ratings of our participants were significantly lower than those in the database, and the negative correlations with error rates and with lexical decision time were stronger for the familiarity ratings by our participants than for those of the database, suggesting that word familiarity should be rated by a homogeneous population.