DIAGNOSING DIRECT ACCESS RETRIEVAL VIA
SPEED-ACCURACY TRADEOFF
Cue-based retrieval operates using an associative retrieval mechanism that accesses information directly (i.e., direct access) using cues available at the retrieval site. This means that retrieval speeds are constant, regardless of the amount of information to be searched through, or the recency of the desired information. This is in contrast to a search-based system, in which retrieval speed will be affected by both. In order to fully understand how the language system interacts with the contents of memory, a means of diagnosing direct access retrieval is necessary. The Speed-Accuracy Tradeoff method makes this possible.
The advantage of this method is illustrated in Fig. 1. Reaction or reading times represent one point (open circles) along a potential response function, but whether this point is associated with the rate of information accrual (slope) or with terminal (asymptotic) performance is unknown. SAT data (filled circles) disambiguates these cases by recording the entire response function. Here, the RT data reflects two conditions that differ in asymptotic accuracy alone, with the same rate of information accrual, shown by the point at which the response functions reach their 63% point (the two horizontal lines. ~ 450 ms). Equivalent speeds across conditions with different amounts of intervening material is the hallmark of direct access retrieval. Without the additional response curve data, it is not possible to distinguish this situation from one where the same RTs (open circles) reflect identical asymptotic performance, but different rates—a case in which asymptotes would overlap and the higher RT point in Fig 1 is well past the 63% point.
We have used this method with skilled readers to verify that syntactic and semantic interference affects the probability of retrieval but not retrieval speed, and that referential prominence affects retrieval speed. We have also shown that less skilled readers use direct access retrieval in the auditory modality, verifying that their reading difficulties are not due to faulty memory retrieval.
Click to see the full paper.
Current Collaborators
Representative Publications
Kush, D., Johns, C. L., & Van Dyke, J. A. (2018). Prominence-sensitive pronoun resolution: New evidence from the speed-accuracy tradeoff procedure. Journal of Experimental Psychology: Learning, Memory, and Cognition. [publisher | pubmed]
​
Johns, C.L., Matsuki, K. & Van Dyke, J.A.(2015). Poor readers' retrieval mechanism: Efficient access is not dependent on reading skill. Frontiers in Psychology—Language Sciences, 6:1552. [publisher | pubmed]
​
Van Dyke, J.A.& McElree, B. (2011). Cue-dependent interference in comprehension. Journal of Memory and Language. 65. 247-263. [publisher | pubmed]
​