In this article, we address the high prevalence of false discoveries in recognition memory research. Using Monte Carlo simulations, our goal was to find a valid measure of performance that reliably separates the contribution of sensitivity (accuracy) from that of bias . Relevant to myriad tasks, most notably oldânew recognition memory, the simulations revealed that common measures confound sensitivity with bias, a finding termed the âmeasurement crisis.â As a solution, we propose a version of d-sub-a ( d a ). We ran comprehensive simulations to evaluate the validity of sensitivity measures, including P r = HR â FAR, A âČ, d âČ, and AUC g , in addition to d a . Memory âsignalsâ were randomly sampled from lure and target distributions. Sensitivity measures generated from iso-sensitive conditions that differed in bias were compared using t -tests, across thousands of simulations. For bias-independent measures, the rate of significant results should be 5%. We manipulated several parameters, including the form of the distributions (i.e., from three prominent models of recognitionâmemory: unequal variance signal detection [UVSD], double-high threshold [2HT], dual-process signal detection [DPSD]), the distance between their means, their relative variance, the placement of response criteria, the sample size, and the number of simulated trials. Results demonstrated that under most experimental scenarios, only d a was unaffected by changes in bias. In contrast, all common measures typically exhibited alarmingly high false discovery rates, exceeding 5%. The rates rose to 100% with larger sample sizes and a large number of trials. These findings indicate that d a warrants serious consideration as the default measure of sensitivity.