Word frequency and receiver operating characteristic curves in recognition memory: evidence for a dual-process interpretation

Abstract

Dual-process models of the word-frequency mirror effect posit that low-frequency words are recollected more often than high-frequency words, producing the hit rate differences in the word-frequency effect, whereas high-frequency words are more familiar, producing the false-alarm-rate differences. In this pair of experiments, the authors demonstrate that the analysis of receiver operating characteristic (ROC) curves provides critical information in support of this interpretation. Specifically, when participants were required to discriminate between studied nouns and their plurality reversed complements, the ROC curve was accurately described by a threshold model that is consistent with recollection-based recognition. Further, the plurality discrimination ROC curves showed characteristics consistent with the interpretation that participants recollected low-frequency items more than high-frequency items.

Figure 1

Predicted receiver operating characteristic (ROC; left column) z-transformed ROC (z-ROC; right column) curves for single-process models (top row), high threshold models (second row), dual-process models (third row), and dual threshold models (bottom row).

Figure 2

Hits to old items and false alarms to similar and new lure items in Experiment 1 as a function of word frequency. Error bars depict 95% confidence intervals.

Figure 3

z-transformed receiver operating characteristics (z-ROCs) from Experiment 1 as a function of word frequency. Triangles represent performance for high-frequency items, open circles represent performance for low-frequency items. Functions for low-frequency items are dotted and functions for high-frequency items are dashed. The top panel depicts the z-ROCs for old–new discrimination, with the best-fitting linear trend. The bottom two panels depict the z-ROCs for old–similar discrimination. The left panel depicts the best-fitting linear trend and the right panel depicts the best-fitting regression model with quadratic components.

Figure 4

Receiver operating characteristics (ROCs) from Experiment 1 as a function of word frequency. Triangles represent performance for high-frequency items, open circles represent performance for low-frequency items. Functions for low-frequency items are dotted and functions for high-frequency items are dashed. The top panel depicts the ROCs for old–new discrimination, with the best-fitting ROC function generated by the Rockit maximum-likelihood estimation algorithm. The bottom two panels depict the ROCs for old–similar discrimination. The left panel depicts the best-fitting ROCs generated by the Rockit maximum-likelihood estimation algorithm and the right panel depicts linear regression fits.