1994) proposed that word frequency effects do not occur on a lexical decision task (LDT) when postmasked target exposure duration is sufficiently brief because such a task prevents verification--their hypothesized locus of the word frequency effect. In making this assertion, they proposed that the activation interpretation ofA. R. Dobbs, A. Friedman, and J. Lloyd (1985) and of P. A. Allen, M. McNeal, and D. Kvak (1992) was flawed. However, evidence that Paap and Johansen's conclusions were

more »

... and that their experimental design contained flaws is provided here. In Experiment 1 of the present study, word frequency effects were evident on an LDT at the 75% accuracy level proposed by Paap and Johansen as being sufficiently low to prevent verification. In Experiment 2 the mental lexica of participants from the same population as that used for Experiment 1 contained very-low-frequency words. Thus, the present results are consistent with an activation locus. Paap and Johansen (1994) , in an article published in this journal, questioned the validity of research that found effects of word frequency on recognition at brief exposure durations (Allen, McNeal, & Kvak, 1992; Dobbs, Friedman, & Lloyd, 1985) . In arguing that word frequency effects are not observed in word recognition tasks with brief exposure durations, Paap and Johansen unfairly characterized our earlier work (Allen et al., 1992) and used flawed methods that invalidate their conclusion that word frequency effects are eliminated under brief exposure durations. We discuss these issues and report a new experiment, the results of which are consistent with our earlier finding that word frequency effects occur under conditions in which, according to Paap and Johansen, they should not. Whether word frequency influences word recognition performance at brief exposure durations is critical to Paap and Johansen's (1994) theory because their activationverification model "eschews the possibility that encoding is frequency sensitive" (p. 1130). According to this model, "because verification relies on a comparison that involves