Retention Interval

Retention interval refers to the amount of time that elapses between the end of a witness’s encounter with a perpetrator and any subsequent testing of the witness’s memory for that encounter. Testing of a witness’s memory for a perpetrator’s identity is obviously important whenever the prosecution seeks to prove that the perpetrator and the defendant are indeed the same person. When eyewitness testimony is provided, the trier of fact must decide whether the testimony is accurate. Unless the trier of fact believes that human memory operates with the fidelity of a video camera, he or she will need to estimate the strength of the witness’s memory at the time of his or her memory being tested. To increase the precision of the estimate, the trier of fact needs three pieces of information: An estimate of the original strength of the witness’s memory representation of the perpetrator’s face, the length of the retention interval, and the nature of the forgetting junction. The forgetting function is the curve that describes the strength of the memory trace over the course of the retention interval.

Inasmuch as the trier of fact ordinarily has access to a relatively precise measure of the length of the retention interval, with both the time of the incident in question and the time of the memory test being well established, the problematic pieces of information are an estimate of the original strength of the witness’s representation of the perpetrator and knowledge of the course of the forgetting function during the retention interval. Let us first consider what is known about the nature of the forgetting function. Researchers interested in how memory for the human face is affected by the retention interval have conducted several dozen published studies wherein they have assessed memory accuracy after two or more different retention intervals. Assessments of the average effect size for the retention interval (measured in standard score units) taken across all these published studies have revealed that, statistically speaking, one can safely conclude that memory traces of human faces encountered but once previously will be weaker at longer retention intervals than at briefer ones. However, simply knowing that memory for unfamiliar faces is less accurate at longer retention intervals does not specify the time course of the forgetting function. The trier of fact would like to know just how rapidly memory strength declines for an unfamiliar face.

Published surveys of the opinions of psychologists who qualify as experts in the science underlying the psychology of testimony have shown that more than 80% of them believe that the nature of the forgetting function for the human face follows the same form as that of the forgetting function first described by the early experimental psychologist, Hermann Ebbinghaus, and reproduced in introductory psychology texts. That is, the experts believe that the forgetting curve declines rapidly right after viewing of a perpetrator’s face and then levels off over time. It turns out that when theoretical forgetting functions are fit to retention interval data from studies wherein three or more retention intervals were tested, theoretical functions that fit the data very well describe a forgetting function that is mathematically quite similar to that of Ebbinghaus.

Given that there are theoretical forgetting functions that make relatively accurate predictions regarding the memory accuracy of the typical witness in a laboratory or field experiment, memory accuracy at any particular retention interval, it should not be surprising that one can “work backward” from the earliest tested retention interval to make a prediction as to what the original strength of the witness’s memory representation was. When this estimate is translated into a proportion correct measure of accuracy, one then has a reasonable estimate as to the maximum level of accuracy expected for the typical witness under the conditions prevailing in the experiment—or in more realistic situations, to the extent that conditions are the same as in the experiment in question. The expectation is that memory accuracy will only decline from this level at forensically typical retention intervals. Interestingly, the retention interval most frequently encountered by the British police has been reported as 1 month. One theoretical forgetting function that fits empirical data well makes the prediction that the strength of the memory trace for an unfamiliar face at a 1-month retention interval, depending on a number of factors, would likely be in the range of 40% to 60% of its original memory strength.

Researchers have identified a number of factors that affect initial memory strength and, therefore, the amount of strength remaining after any retention interval. Longer exposures to an unfamiliar face, better lighting, and greater facial distinctiveness (as compared with the typical face) have all been shown to increase initial memory strength. Estimates of witness memory accuracy when tested with a lineup or photo spread have also been shown to be a function of how distinctive (or similar) the unfamiliar target face is relative to the other faces presented. A high degree of similarity will produce a lower estimate. Events occurring during the retention interval can also seriously affect witness memory accuracy. For instance, exposure to mug shots before the ultimate memory test, followed by a memory test that includes one of the faces from the mug shots, increases the probability of erroneously selecting the face seen in the mug shots rather than at the crime scene.


  1. Deffenbacher, K. A. (1986). On the memorability of the human face. In H. D. Ellis, M. A. Jeeves, F. Newcombe, & A. Young (Eds.), Aspects of face processing (pp. 61-70). Dordrecht, The Netherlands: Martinus Nijhoff.
  2. Deffenbacher, K. A. (1996). Updating the scientific validity of three key estimator variables in eyewitness testimony. In D. Herrmann, C. McEvoy, C. Herzog, P. Hertel, & M. K. Johnson (Eds.), Basic and applied memory research: Vol. 1. Theory in context (pp. 421-138). Mahwah, NJ: Lawrence Erlbaum.
  3. Wixted, J. T., & Ebbesen, E. B. (1991). On the form of forgetting. Psychological Science, 2, 409-115.

Return to the overview of Eyewitness Memory in Forensic Psychology.