That likely depends heavily upon the timing, motives, and feelings about the event when they give their post-hoc estimate of it (which in this case was up to a year later, when their feelings about the other person are likely to be very different). Most under-estimates of BAC are from people with a biased interest in reducing culpability for actions that are made wrong by them being drunk, like DUI. The increasingly prevailing view is that higher BAC makes one less culpable for sex acts, and the other party more culpable. Thus, anyone trying to reduce culpability for actions they now regret can use inflating their claimed BAC at the time, so long as being drunk doesn't make them criminally liable. IOW, when it comes to doing or saying stupid and impulsive things, people very often inflate their recalled BAC to blame their foolishness on the alcohol or drugs. In fact, criminals do this all the time when their crime isn't really made worse by being drunk. It is no more a crime to commit theft while drunk than sober, so drugs are often used as a defense or excuse when trying to reduce sentences for such crimes.
Also, not just in my experience, but the scientific literature, people are likely to bias their memory of their objective and subjective experiences to match their current emotions. Again, that is mainstream science not my personal experience. Applying these established factors to this issue, if they no longer feel like they would have sex with that person, they will tend to mis-recall or reinterpret various aspects of the situation in ways that match their current unwillingness. Attributing the sex to intoxication is an easy scape-goat to downplay their willingness at the time.
Any person they willingly had sex with before was someone they had some positive feelings about at the time. Probability dictates that they would feel less positive (and in many cases quite negative) about many of those people a year later. Combined with these basic features of human memory and interpretation of the past, this strongly predicts that a significant % of women would mis-remember aspects of the sexual encounter in ways that paint it as more negative and more determined than it was by factors other than their feelings for the person.
This is not a sinister attempt to lie on their part, but a substantiated unconscious influence of current states on the reconstruction of past events. FYI: all memories are actually reconstructions filled with inferences rather than literally pulling up a preserved "filed" from storage. Remember that there is no evidence that any of these 15% in the study have ever claimed or viewed the incident as rape or in any way criminal. Which means there is no real consequence for them getting it wrong to preserve their sense of emotional coherence over time (other than the unfounded conclusions reached by those biased toward accepting the responses as veridical and blindly ignoring the likely errors).
And part of that background was their own beliefs that drugs made sex more likely and more fun, which positively predicted whether they gave a "yes" response to the incapacitated question. Also, self-reports of experiences need not be as uninterpretable as these are. Had they given the women a precise definition of "incapacitated" based upon specified criteria, like states of consciousness and awareness that they were having sex and who with, then we might have some idea of what those experiences refer to. IT would also help if the later parts of the question about "unwanted" were asked in separate questions and clearly stated whether the sex was wanted at the time and distinguished that from whether they gave indication of consent. In sum, the problem is not a lack of an objective BAC measure, but the lack of even a subjective measure that is at all interpret-able in terms of "rape" by any defensible definition whether legal, moral, or in the opinion of the women themselves.
