Recognition-memory models and ranking tasks: The importance of auxiliary assumptions for tests of the two-high-threshold model

• It has been argued that recognition rankings falsify two-high-threshold theory. This claim assumes invariance of lure detection, which does not hold empirically. Without this assumption, theory explains ranking judgments. Lure detection improves when testing lures jointly with strong (vs. weak) targets. We conclue measurement models need careful adaptation to new paradigms. The question whether memory should be measured assuming continuous strength (signal theory) or discrete states (threshold become a prominent point discussion. In light limitations associated receiver operating characteristics, comparisons the rival based on simple qualitative predictions derived from their core properties were proposed. particular, K -alternative tasks ( ARTs) yield conditional probability targets being assigned Rank 2, given they 1, is higher for than weak finding incompatible (2HT) model (Kellen & Klauer, 2014). However, we show incompatibility only holds under auxiliary assumption detecting invariant target-strength manipulations. tested in two different ways: by developing versions 2HT tailored ARTs and employing novel forced-choice-then-ranking tasks. Our results can explain increases 2 target strength. effect due larger lure-detection probabilities test displays are ranked (as compared conclude vary recommend allow variation. As such compatible ART performance, our work highlights importance carefully adapting