What sexual assault victims remember about rape has been the subject of much research and courtroom debate, especially when there is alcohol involved.
Some have argued based on outdated and very little empirical research that victims remember rape less well than other life events. For example, according to guidance for police interviewers posted on the website End Violence Against Women International (EVAWI):
One of the fundamental challenges to the credibility of sexual assault victims is that many – if not most – make statements to the law enforcement investigator or others that are incomplete, inconsistent, or just plain untrue. (See it in context here and here).
They then go on to explain that rape victims cannot remember what happened. These are strong claims. Not only will police interviewers read this guidance, but so will the media, potential jurors and other legal professionals, such as prosecutors, who make decisions about whether to bring charges in rape cases, and defense experts, who testify before juries to undermine the rape complainant’s memory evidence.
The claims may also impact what rape survivors make of their own experiences, their memory capabilities, their recovery and their willingness to report the crime to the police.
Thus, there should be good theory backed by a strong evidence base to support these claims given how consequential they are for people’s lives.
Rape Memory Research
The only research they cite is work by Koss and colleagues (1996), who set out to study in women whether rape memories are different from other types of memories. The guidance summarizes Koss et al.’s findings as follows:
The research reveals that memories of sexual assault – as compared with other types of memories are ‘less clear and vivid, less visually detailed, less likely to occur in a meaningful order, less well-remembered, less talked about, and less frequently recalled either voluntarily or involuntarily; with less sensory components including sound, smell, touch, and taste; and containing slightly less re-experiencing of the physical sensations, emotions, and thoughts than were present in the original incident’ (Koss, Figueredo, Bell, Tharan, & Tromp, 1996).''
EVAWI concluded from this research that rape investigators should be cautious and not trust the word of a survivor. How might legal practitioners who read this be affected in their decisions about interviewing complainants, or in bringing charges against defendants, or in how they cross examine complainants?
EVAWI's conclusions, however, are not warranted given the poor design of the study they cite. They also do not mention Koss and colleagues' findings contradict theory, other empirical findings, and clinical experience, not to mention the lived experiences of many rape survivors.
Rape Memory Research Confounds and Contradictions
The study design was as follows: About 3,100 adult women (their age ranged from 17 to 75 years, and on average was between 36 and 40 years) completed a self-report survey, rating their memory and emotions for either a rape that occurred when they were 14 years of age or older (30% of the women reported such a memory). Women who did not report rape were asked to rate their memory for another significant event of their choosing. The researchers then coded the significant events as pleasant versus unpleasant, dividing the nonrape group into two subgroups. The ratings of the groups of women (those who rated rape versus those rating another type of memory) were compared to draw the above conclusions.
There are several likely confounds arising from the design. It is not possible to conclude from it whether the type of memory (i.e., a rape versus nonrape memory) caused these differences in memory characteristics.
Memory age is one potentially a serious confound. Younger people are at an increased risk of rape (WHO, 2013). Koss and colleagues' participants tended to be middle aged. This suggests participants were by and large reporting memories that were 10 or more years old. The authors did not measure how long ago the event occurred nor did they control for this in the design (e.g., by asking people to report about events that occurred within a specific timeframe). However, among those who reported rape, the authors did state that the events tended to be old memories, with 93% rating a rape that occurred more than 2 years ago.
If the rape memories were older on average than other types of memories, this alone could account for the findings. If this were the case, we would conclude from their data:
When an event happened long ago, memory strength weakens over time, and the event will be less well remembered, less clear and less vivid, less detailed, less likely to occur in meaningful order, especially if it is not rehearsed.
Near the end of the paper, in the Discussion, the Koss and colleagues' briefly mention that memory age may be a confound (p. 430):
Other potential influences on our results include possible differences between the age at which the rape occurred versus the other unpleasant incidents, or differences in the time that had elapsed since the experience. Unfortunately, we cannot characterize any further the content of the nonrape memories, nor can we substantiate that the types of memories were equivalent in developmental stage or recency.
(Note that if a researcher did an experiment and could not tell us whether the delay between the to-be-remembered event and the memory test was the same across the groups that were being compared, the results would not be publishable.)
As another example, the Koss and colleagues did not control for whether the memory was part of a series of repeated events (e.g., recurring rape by a family member). The rape group may have been more likely than the nonrape group to be rating a memory from a series of events rather than a single event. Remembering a particular episode of sexual trauma may be affected if it is part of series (see this paper for a discussion of the issue).
In discussing the implications of the findings, Koss and colleagues and EVAWI seem to have lost sight of the fact that groups of women, not memories per se, were compared. Despite the myriad of ways that the groups potentially differ (e.g., the rapes may have been more likely to have happened at night and thus under poorer illumination compared to other types of events) the results are discussed in terms of how memory for rape is qualitatively different compared to other types of memories.
Koss et al. Findings Inconsistent with Other Research
The findings are also inconsistent with other empirical findings. Considering that past research tells us that rape survivors are at increased risk of PTSD and people with PTSD can have more vivid memories, the Koss et al. went to great lengths to explain why, in their study, that people who reported rape described memories that were less vivid and less clear than should be expected based on past research and clinical experience with PTSD.
To account for this discrepancy, the researchers proposed in Tromp et al. (which is a paper by the same researchers and uses the same dataset) that women who are raped cope with it through avoidance: They don’t think or talk about the rape because they are not able to put what happened to them into words. Thus, women in their study remembered rape less well than expected because they avoided talking about the rape.
Note that their reasoning is circular and based on conjecture. Koss et al. did not measure any of this or give a theoretical reason for why rape is harder to put into words than other events.
Koss and colleagues further opined in their papers that analyze this data set that memory for rape is less intense than other memories because women who suffered rape in their study were probably more likely to have been intoxicated during the attack, which may have ‘reduced the quality of memory stored.’ Again, there is no evidence for this conjecture. The researchers did not ask participants about whether they were intoxicated.
As an important aside, what kinds of messages does this research convey to victims? That it’s normal to not talk about rape? That people cannot rely on their memory of the rape to know what happened to them? If they report rape, people will not think they are able to remember, so don't bother?
Peace and colleagues (2007) followed up on Koss et al.'s research using an improved design. They asked 44 participants who had been raped and referred for counseling to rate their memories of a sexually traumatic event, a nonsexually traumatic event and a positive event. Each participant rated their memories for all three types of events. The events all had to occur within the same timeframe (2 years).
In a better controlled study, Peace et al. (2007) found that participants rated their memories for a sexually traumatic event as more vivid and better remembered compared their memories of other types of events.
Interestingly, Peace et al.’s participants who rated their memory as more coherent also gave higher likelihood ratings when asked if they were going to proceed legally with the case. This result should caution us against broadly generalizing research results to rape complainants who report rape to the authorities. Those who proceed might better remember what happened to them compared to those who do not proceed.
A study of domestic violence survivors replicated Peace et al.'s main results, similarly finding that women reported stronger memories for trauma experiences compared to their positive experiences.
This recent study also came to the same conclusion, though note that, unfortunately, it is not clear what event women in the study were remembering.
Still, we need to keep in mind that these studies use small numbers of participants and self report data. We do not know the extent to which findings generalize to other people. We also have no way of verifying how accurately participants were remembering the event or rating the characteristics of their memories. (Interestingly, Koss et al. are saying in effect that women in their study can't remember the rape but they can accurately assess how they remember it. This in and of itself is a research question.)
Consultants and Expert Testimony on Rape
Consultants who provide guidance about police interviews with rape victims or experts who proffer testimony in the courtroom must acknowledge that this is an active research area and that there are a range of theoretical positions, with some researchers arguing that traumatic events are not remembered any differently than other types of memories, others arguing traumatic memories are stronger than other types of memories, and still others arguing traumatic events are not well remembered.
Finally, I believe Koss et al. were well meaning and wanted their findings to help rape survivors. However, as researchers, we should think about the various ways in which our research will be put to use (e.g., case issuing decisions, expert testimony against rape complainants) and write carefully about our work.