There is now a GitHub site,
https://github.com/leifpowers/judgment-standards-code
which contains the spreadsheet and program to which the below refers. Please submit your fixes and additions.
The question about how we should handle rape reports boils down to the following considerations:
– What standard should we objectively use to decide whether someone has been raped or not. It will actually have to be a non-refusal, but I will discuss why that is in more detail.
– What specific judgment standard, operating on the actual evidence that we as law enforcement can observe, is the best? This is not as obvious; the tradeoff is between acceptable false positive rate and acceptable false negative rate, and based on the software model, it does not appear that any such thing exists. Furthermore, as you will see, this is actually mostly irrelevant, since many indicators of intent or resistance can be forged or manipulated.
– How can we prevent our standards being exploited, either by rapists to get away with their crimes, or by black widows looking to send people to prison that they don’t like? Based on the spreadsheet breakdown of exploitability, combined with the difficulties with judgment standards, it seems that the only realistic approach is to have a formalized approach for consent. This will add bureaucracy and certainly will not be followed anywhere near 100%, but it will provide a realistic path for many victims to receive justice and for non-rapists to be able to avoid getting trapped by black widows.
The first question is about whether we, seeing with the eye of God a given sex situation, would consider it rape. There are a few commonly seen valuations that people raise:
– Affirmative consent (must say yes), with a possible invalidation if a partner is drunk
– Nonrefusal, which is that if a partner is not actively resisting, then it’s not rape
Nonrefusal is the only worthwhile standard due to considerations such as:
– Marital rape vs. non-rape. The vast majority of marital sex acts do not involve explicit, unambiguous yeses at each step. It may even be the case that the majority of non-marital sex acts do not involve a legally suitable consent. Furthermore, in many cases, one or more of the partners is drunk. If we enforced the affirmative consent standard, we would send millions of people to prison who never did anything with which their partner would have a problem.
– Proof of consent, and being able to demonstrate it. Under affirmative standards, the individuals involved would have to provide the appropriate proof if someone raised the point of whether they had had sex legally. This would create a tremendous amount of daily overhead, and people just won’t do it.
There are two counterpoints raised to this:
– Just assume that the marital vow or informal living arrangement constitutes consent. This leaves people at the mercy of their partners if things get difficult prior to formal separation, but this is not the biggest flaw with this argument. The real issue is that this argument actually elevates a formal document or de facto arrangement as true consent, and not a per-encounter or per-sex-act consent. Effectively, once you make this argument, you are no longer arguing for affirmative consent as it is commonly understood, but rather some sort of nonrefusal combined with indisputable arrangement.
– Just ignore the proof of consent requirement until someone makes an actual accusation of rape. This is a highly exploitable recommendation, especially if intoxication invalidates consent, since rapists can argue that they were drunk too. It also imposes a very heavy investigative burden based on credibility and non-forgery from the involved parties. If applied with a threshold of proof beyond a reasonable doubt, it also will perform extremely poorly, since the consent has to be applied at the time of the sex act in question, and hence the evidence of that consent must be preserved, which is more or less equivalent to having credible eyewitness testimony or video evidence of the sex act. When applied to every sex act, that’s a lot of video evidence to preserve.
Hence, only some sort of nonrefusal standard would be a viable means by which to evaluate whether a particular sex act was actually rape.
The question of the best judgment standard is about the tradeoff of false negative and false positive. Regardless of even the oracle standard of whether someone was raped, one or the other of these is poor, as evidenced by the output of the program. False positive rate can be driven to levels approaching one percent, but only at the cost of significant (in many cases 50% or more) levels of false negative rate. Even the approach of simply taking the victim at her word outperforms many of these combinations of approaches. It is possible that under a more refined and potentially realistic probability distribution, that an approach could be found that does a better job of balancing the tradeoffs. However, the following scenario demonstrates the difficulty of truly improving upon this Pareto curve:
– Assume that the reporter is right 90% of the time, so that gives you a 10% chance of a false positive rate. Hence, it would make sense that you would only want to weed out reports that are clearly false. How would you do that? You would look for situations with contradictory evidence or eyewitnesses – that might get you some reports refuted, assuming your eyewitnesses are credible. You would look for situations where evidence was lacking – but lots of actual rapes have similarly lacking evidence, so that’s probably not a good place to look for improvements. The 10% false positive rate would be composed of some women looking to trap, some women looking to get back at their lovers for whatever reason, and some women who, through pressure from others or their own regret, are looking to enhance their own reputation by making the allegation of rape. All of these women would be willing and capable of deceiving their partners, and all of them would be willing to fake evidence and harm themselves in order to convince the authorities of their assertions. Consequently, there would have to be some sort of motive or credibility test as part of this, which now makes the evaluation of the actual act a function of the general, unspecific credibility of the parties involved in the sex act. That is to say, by taking this approach, the controversy no longer is a question about whether the woman was raped, but about whether the woman or the man has a better reputation and status.
So, if we assume that the reporter was raped 90% of the time, why would we not just take the person who reported rape at her word? The answer is that it creates an exploitable situation, where women can falsely accuse anyone without any check. In general, creating exploitable situations greatly increases the probability that reports and actions are taking place as exploits and not as earnest decisions made as the result of normal factors. We will not be seeing 90% accurate rape allegations, but some uncertain amount that is significantly less than that.
Consequently, our determinations must be based on measurements that are difficult for both parties to forge after the fact. The spreadsheet included in the bundle lists the means by which the various sorts of evidence can be made up after the fact. Some of them are fairly straightforward to manipulate. The ones that concern intent and are hard to spoof are the direct video or eyewitness evidence, if it exists, those related to legal matters (such as pre-consent forms and prior accusations) and the general opinions and assessments of individuals relating to the sexual history of the victim and accused. The intoxication state is also helpful, but whether it is asymmetrical between attacker and victim often is difficult to assess at the time of the crime report.
If we concern ourselves with the specific question of whether the specific sex act in question was rape, then, as far as I can tell, only direct consent forms or eyewitness/video evidence are direct indications that cannot be easily spoofed. Evidence about general sexual history helps to determine whether any such declarations of chastity are actually sincere; that is, if someone has a declaration of chastity on file, but is constantly having one night stands with people from singles bars, the declaration means nothing, and should be annulled.
Consequently, to avoid exploitation of the system by both rapists and black widows, people must file consent forms listing the partners and sex acts which they would permit. Those same reports must then go out to partners previously covered, and to partners newly covered, to prevent someone from pulling a bait and switch. In order to protect the virgins, the default for an individual is no consent.
When, inevitably, people have sex in contravention of their declarations, but no rape is alleged, their declarations will be struck, they will do weekend prison time, and they will have a demerit on their consent record. Were we to let them off the hook, people easily could continue the current practice of accusing anyone they want.
As for the process: when the victims get away from the rapists, they go to the hospital and submit to the rape kit. The DNA tests get done, the consent forms are checked, and the rapists, when identified, go to prison.
The hookers, playazzz, and sluts won’t have much protection under this system; only gruesome assaults on them are going to be punished. However, that’s the trade that has to get made in order to protect the people who are asking for accurate justice.