Establishing the psychometric properties of measures of implicit memory

The literature currently permits the coexistence of two competing theories of implicit memory: the levels of processing approach and the multiple memory systems approach. Attempts to conceptualise implicit memory have been impeded by using tests of unsound psychometry. This research selected tests most commonly used in research and applied settings, which were most likely to be psychometrically sound, and examined their internal consistency and construct validity. Seventy-two participants were given the Visual Reproduction subtest (Wechsler Memory Scale, 3rd ed.), Rey Auditory Verbal Learning Test, mirror tracing task, magnetised dart-throwing task, speeded adaptation of word-stem completion task, and tachistoscopic identification. Tachistoscopic identification and mirror tracing were internally consistent, while the speeded adaptation of word-stem completion and the magnetised dart-throwing task were unreliable. Analysis of encoding strategy failed to support the levels of processing account of implicit memory. Exploratory factor analysis supported a multiple memory systems model of implicit memory.