So again I ask … what lesser level of proof should be required to warrant expulsion for “inappropriate behavior”? *
Would there be any open hearing of the facts at all or would even an anonymous accusation (woman’s identity protected) be enough?
Let us be clear: what is being proposed is that a possibility of guilt is enough to expel a student on grounds of “inappropriate behavior” (with major consequences to his/her life course) without any real burden of proof. Better to punish a hundred innocent people than let one guilty individual go free, eh?
I also ask why the expectation is that non-college environments are more acceptable places for women to be victimized? As established, young women in non-college environments are at higher risk of sexual violence than those in college environments. Why should not the same standard of guilty unless proven innocent apply there as well?
As far as the statistics go, RP, assuming a completely random representative sample is the big assumption that is being questioned. The study states that at one particular mid-sized city school they were able to get 1882 men at busy street corners to stop and take a survey of undescribed length in which sexual violence questions were contained as part of many for $3 or $4. Unmentioned is what day of week or time of day or if the busy street corners were near the bars or near the library or what. Of that selected group they found seventy-something individuals who admitted to multiple rapes or attempted rapes mostly by way of having sex with someone when they were too intoxicated to refuse. Pretty much the same repeat offender seventy-something also admitted to being guilty of battery and child abuse.
Is there any reason to believe that a sample of men willing to take some undescribed amount of time to participate in a study for 3 or 4 bucks at unspecified locations and times of day or week is representative of all males on the campus? Is there reason to believe that this campus is necessarily representative of all campuses? What would a matched age control done in a non-campus environment (not done) demonstrate?
No you do not generalize from one small sample without controls to the population as a whole on the assumption that a population is random and representative of the whole. If using a small sample and attempting to make such a generalization you do lots of work to have a true random representative sample and can reasonably prove it is. Without such proof the assumption is that it is not. I would not be surprised if the finding was real, and even more pronounced in a non-college environment, but this study is poor evidence of such.
*Note: I am not saying colleges should not be legally able to expel a student for any reason that is not on the basis of a protected class or otherwise be able to be just as at-will as employers. I am asking if doing so on suspicion alone, without any real burden of proof, is justified, fair, ethical, or wise.
It is perfectly possible that UMass is uniquely full of rapists. But in the absence of any reason to believe that, this is not a design flaw.
The survey was also conducted in English (I presume). That alone doesn’t make it flawed merely because we could speculate that rapists tend to speak French.
If you believe that the design led to non-random sampling as to the proportion of rapists, it’s really on you to spell that out. Merely observing that they generalized from a sample–which was ITR Champion’s criticism–is not a persuasive argument because that’s true of all sampling. You need to provide some reason for thinking the sample was too small, or that it oversampled the relevant variable.
Nobody is proposing that. IME, the level of proof required in student conduct hearings is generally a preponderance standard, with the burden falling on the complaining party.
I’m not sure what you’re talking about specifically, DSeid. You say “let us be clear about what’s being proposed” and you put “inappropriate behavior” in quotes, but it isn’t at all clear that you’re responding to any particular point.
BrainGlutton said he’d be OK with somebody being expelled for rape based on “considerably less” than the proof needed for conviction. I said that where the facts are clear that a student acted inappropriately in a sexual misconduct case, it makes sense that the university’s disciplinary policy, in lieu of a criminal prosecution, should be able to address the situation.
Where are you getting this idea that expulsion based on a bare accusation – “suspicion alone, without any burden of proof” – is what anyone is after?
The burden of proof lies on those who assert that the study was valid and demonstrated something. IOW no one has to prove that the sample was invalid - the researchers have to prove that it was valid.
(But I’ll give you an argument anyway: It doesn’t make any sense to put the burden on the research to disprove every possible speculation about how survey design might have biased the sample. In the real world, researchers have the burden to disprove obvious sample problems. The notion that UMass just happens to have more rapists it not an obvious sample problem.)
I agree with DSeid. One of the problems with the sample is that it is not representative of all campuses. UMass is a commuter college that has no dorm facilities and a high proportion of nontraditional students (part time, older). There are other problems as well. And yet it is being used to determine policy for traditional colleges.
I would like to note thatDr. Mary Koss, who conducted the study that was the original source of the 1 in 5 statistic, has this to say about Lisak’s study:
Dr. Mary Koss—a scientist, feminist, and acclaimed expert on the subject of campus sexual assault—says the psychologist who popularized the serial predator theory of student-on-student violence has misrepresented his research for years.
“It’s one of the most egregious examples of a policy with an inadequate scientific basis that lives on because it offers a simplistic solution,” said Koss, in an interview with Reason.
I think a lot of people are hesitating to speak up about this because of politics. The criticisms are coming from right wing and libertarian sources, and IMO the progressives are worried that acknowledging the criticism will affect government funding, and resources for victims. To be fair, that probably is the primary motivation behind the criticism. But I can’t in good conscience defend this study on the merits.
If they were self-selected, that could be a problem. If there is some reason to think UMass is different, also fine. These are different from the wrong criticism of the general idea of small samples.
Things were fine as far as Sterrrett knew, for a while. Then CB’s mom found CB’s diary, and contacted the school. That’s when things started going haywire.
He was ordered to stay away from his dorm room, while the administration conducted an investigation.
They talked to him once, by Skype, and then:
I also looked at the National Crime Victimization Survey. It says the prevalence rate for rape/sexual assault in 2014 was .06%. (Table 4, page 5.) It was the least prevalent type of crime they track - other than “serious intimate partner violence”, which was .05%. Those are 2014 numbers, not lifetime numbers.
If you look at multi-year rates, according to the Slate article “0.8 percent of noncollege females age 18-24 revealed that they were victims of threatened, attempted, or completed rape/sexual assault. Of the college females that age during that same time period, approximately 0.6 percent reported they experienced such attempted or completed crime.”
I originally thought the 1/5 number came from this study: The Campus Sexual Assault (CSA) Study, which is what the Marshall Project says. But the study actually says “Data indicate that 13.7% of undergraduate women had been victims of at least one completed sexual assault since entering college” - which is not in in five. The study sent out invitations to student to participate in a web survey, but the participation rate was low: almost 60% didn’t participate. Or to put it differently, the 42% who did participate were self-selected.
There is NO reason to believe that the men sampled are representative of the males of UMass/Amherst, let alone of college males in general, let alone college males as opposed to males anywhere else. There is no way this study should ever have been accepted by any reputable journal.
If one wanted to create a representative male sample of a particular college population going to where the light is good (some arbitrary busy street corners) and asking some unknown huge amount of passersby to take some unknown prolonged length of time to fill out a survey for three or four bucks and waiting until eventually some arbitrary sufficient number finally did not say “Uh, no. Busy now. Class to get to.” is not the way to do it. And if one did then minimally one would report how many said “No.” and would collect demographic information from the fraction that did stop to at least document that on obvious items like age, race, majors, drinking and drug habits, relationship status, etc., they are similar to the male university population as a whole.
Using small sample sizes to extrapolate to a whole depends upon rigorously showing that the small sample is representative of the whole. Not say assuming that sampling people outside a busy donut shop that happens to be across form your office is representative of all of a city’s BMI and nutrition habits and thus representative of every city’s populations BMIs and nutrition habits.
More reasonably the college enrollment list could have been used and a certain population randomly chosen and then balanced to reflect the composition of the male campus population would be chosen and offered enough money to get a high participation rate. The composition of those who declined would be monitored to make sure there was no obvious factor correlating with that and the make up of the final sample compared to the known general male campus population composition, along with the refusal rate reported.
There would still be concerns about confounders but they would be minimized.
If you pull a sample at random from a population, you can extrapolate safely to the population that the sample was taken from. You cannot extrapolate safely to a much larger population.
For example, if you took a random sample of people in Utah and asked them whether they were Mormon, you could use the results to form conclusions about the percentage of Utah residents who are Mormons. You could not extrapolate the results to the percentage of Americans who are Mormons, or the percentage of human beings who are Mormons.
I have asked for some clear explanation of what lesser burden of proof would be accepted. Not a lawyer, so “preponderance” don’t mean diddley to me without some examples. About that I am indeed ignorant. Give me a hypothetical accusation of rape that would have the evidence highly probably not be sufficient to win at trial and that would meet that standard.
Again, not a lawyer, but most rape cases that are of the sorts being included in these college surveys are of the “too inebriated/high to consent” or “I said ‘No.’” variety, having had sex is not in question, significant physical violence is not usually alleged, and the cases already depend mostly on how much the accuser is believed about lack of consent versus how much the accused is that there was mutual consent. The only difference I can imagine in such cases is that if significant doubt exists over who is being truthful then “not guilty” is often a court’s response, while in such cases a school tribunal is free to say that in the case of significant doubt over who is truthful the accused should be expelled. Unclear if the accuser would be protected from cross examination in that circumstance as well out of well intended concern to protect her from being re-victimized. Is there another scenario that you see as probable?
Yes, that equals “suspicion alone, without any burden of proof.”
I remain waiting for anyone to offer an explanation as to why non-college environments should be allowed to be so much more risky for young women and only women in college environments (statistically at, if anything, less risk than same aged women in non-college environments) are the population of concern.
Oh an interesting factoid according to the Lisak study: More than one out a hundred of college aged men have committed multiple acts of child sexual abuse as well and about 2.5% have beaten up an intimate partner. Now I actually do not doubt these numbers as the CDC’s surveys, done with much greater care, come up with general population numbers on intimate partner violence that are significantly worse and the vast majority occurring first before age 25, and many before age 18. That report contains scary and sad statistics to be sure, but they do also provide the actual context: colleges are relatively safer places than the rest of the country is. The rest of the country just does not apparently rank as much concern.
Oh. “Inappropriate behavior” was in quotes because of the position that a man expelled under this standard would, based on parallel to “at will employment” not being expelled for “allegation of being a rapist” lest they open themselves up to a defamation lawsuit, but just for being inappropriate in some unspecified way that warranted expulsion.
Only if the university had a policy against sexual conduct between students; and if a university claims a policy has been violated then they should have to have proof and an appeal process to deal with the assured abuses.
The real solution is to make sexual activity between students dealt with like sexual activity in the workplace. Make the rules unambiguous. Is this realistic? Prob not.
Just to be sure about something, am I remembering correctly that Dseid and Camille would typically not be on roughly the same “side” as ITR Champion in a dispute about a socio-political issue?