Modern memory researchers rely heavily on the encoding-retrieval match, defined as the similarity between coded retrieval cues and previously encoded engrams, to explain variability in retention. The encoding-retrieval match is assumed to be causally and monotonically related to retention, although other factors (such as cue overload) presumably operate in some circumstances. I argue here that the link between the encoding-retrieval match and retention, although generally positive, is essentially correlational rather than causal—much like the link between deep/elaborative processing and retention. Empirically, increasing the functional match between a cue and a target trace can improve, have no effect, or even decrease retention performance depending on the circumstance. We cannot make unequivocal predictions about retention by appealing to the encoding-retrieval match; instead, we should be focusing our attention on the extent to which retrieval cues provide diagnostic information about target occurrence.
, vol. 10, no. 5-6, pp. 389-395, 2002