Jessica Ladd Calisto Project.mp3

Jessica Ladd Calisto Project.mp3: Audio automatically transcribed by Sonix

Jessica Ladd Calisto Project.mp3: this mp3 audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.

Oh, I'm Jess and I run Sexual Health Innovations. We are a nonprofit dedicated to creating technology to advance sexual health and well-being in the United States.

Callisto is an online sexual assault reporting system for college campuses. It is our main initiative and I say try, but it's not our only one. Which is why you might be like I. I try Callisto. Those are two different things. They are but one, those within the other. And I'm really excited to talk today because normally when I'm asked to speak, it tends to be a very short pitch and it has to not involve a lot of numbers and not involve a lot of qualifiers around those numbers, which I hate because I'm a huge data nerd. So I'm excited today to go through a little bit of giving you all a background of what sexual assault looks like on college campuses, what sort of the setting is why we see such high rates of underreporting. How what the concepts in game theory, which of those might apply to this problem and how information escrows might address it and then sort of walk you through how we're we're applying it in the case of Callisto and open up discussion to what other settings might a similar system work in? So about 20 percent of women. And somewhere between four to eight percent of men will be sexually assaulted at some point during their college career. And rates are even higher among trans students. And depending on the way you define sexual assault and the way you ask the question about assault, you'll get slightly different numbers. But study after study is finding numbers in around this range. Ultimately, though, actually who the survivors are is a lot less interesting than who the perpetrators are. And when you look at asking people basically, have you sexually assaulted someone? Not in so many words, but by asking them about their behavior.

Well, only men are serving. You almost never see women surveyed. So it's an unknown percent of when women are perpetrators of men and of college men. Somewhere between six and 11 percent of college men say on surveys that they have committed behavior that would amount to sexual assault and by six to 11 percent. What I actually mean is six or 11 percent. Because there's only been really two studies on this. And I want to walk you through a little bit of those two studies because they have sort of different findings. Now, the first of these studies is quoted a lot. Because for a long time, it was the only data we had. And it was done by David Lee SACC and published in 2002. And what Loebsack found when surveying college attending men and he surveyed over eighteen hundred men, he found that six percent of them so that they had committed assault. And at some point and he found a really high rate of repeat perpetration. So what he found is that of people who had committed sexual assault. About 40 percent of them had committed at once. And about 60 percent of them had had committed more. That it more than once in sort of this long tale of assaults, which means that when you look at from the survivors perspective. If you're wondering, was my assault the only one this person committed or was it their first one? If you use those same data and you break that out, you'd see that you have some proportion. It's the first assault, some proportion.

The second assault. But that almost 60 percent of it. It's it's the assailants. Third assault or more. So there's a pretty good chance if you have experienced sexual assault based on leasebacks research, that what you experienced was committed by somebody who has done it before or who will do it again. And that is kind of scary. It also has pretty big implications when we're talking about policy, because if it's these repeat offenders, we think, well, maybe they're predators. They're probably just really bad people. The only solution would be to lock them up. Um, and that leads towards a certain type of thinking. More recently, there is another paper that came out and this came out this year, and it sort of challenged this idea that these are serial perpetrators who are committing all this assault. And it was done by sward out to at all. And it was in JAMA Pediatrics. And what they really looked at was over time in a longitudinal study, do people who have assaulted before. Do they continue to assault in subsequent years? So they didn't ask about rates of assaults, like how many assaults they committed, which least acted, but they were trying to see. OK, do rapists stay raping or do they not? And they actually found that third didn't seem to be anything predictive about past behavior and future behavior of assault. So having assaulted people in high schools didn't predict assaulting people in college, having assaulted people freshman year, didn't seem to predict assaulting people sophomore year, which sort of tells a different narrative. And they don't directly contradict each other because they're measuring different things in different ways. But.

This can be interpreted as well, maybe these aren't sort of serial bad people who are always doing this forever. Maybe this is people who sort of start and then stop again for whatever reason. Maybe they learn not to or they get scared that they'll get caught. Who knows? So the truth about perpetrators, which is really who you need to understand if you're trying to present prevent assault in the first place, is that we don't really know much. We don't really know who they are. We don't really know why they do what they do. We don't really know how many times they do it. And we have to make a lot of assumptions or at best on that, based on the data that we do have. What we do know is that among people who are assaulted, less than 10 percent ever report their assault to their school or to the police. And the rate's a little higher among the general population than it is among college survivors. And the rate is lower among men than among women. So you see certain communities are far less likely to report, often due to increased shame or stigma around what happened to them or distrust of the police or other authority figures. And we do know that when college survivors do report, they on average delay. Eleven months. And this can have really big implications. So when we're thinking about the impact of underreporting and delay on the individual, we're really looking at a lost chance for closure and empowerment.

Right. Ideally, you would report ideally it would be taken seriously. Something would happen. And that way you are able to make meaning out of what happened to you. That way you're able to say, well, like, that was shitty, but at least something happened, at least a change that is positive occurred. Um, it means that among survivors who don't report often there is a feeling of guilt. Often we hear survivors in our focus groups or our surveys talking about how well I always really wondered what I should have done or I always wonder if I should have said something or if something else happened that I could have stopped. And that is not awesome, right? When you're already chances are beating yourself up over your own sexual assault because there's often a lot of self guilt. Now, you're also blaming yourself for the rapes of other people. Um, another impact of particularly the delay in reporting is a very low prosecution rate. So among the, uh, assaults that are reported to the school and more assaults tend to be reported to the school than the police and campus settings. Um, and among the assailants who are found to be responsible, which is the school's equivalent of guilty, uh, only less than a third of them are ever expelled. So you have a pretty low rate.

And I don't I couldn't find data on it, which is why I don't present it here. Uh, of of reported cases to the school, what percent result in a responsible finding. But of those that are found responsible, the consequences don't tend to be particularly severe. And when you look at reporting to the police, it's actually, uh, not a particularly rosy picture. So this is from an association called Rain. And here they're they're assuming that every one hundred rapes, about 32 percent get reported. Uh, that is true of the general population, as far as we can tell. But again, is lower among survivors but of campus assault. But when we look at among those reports, the police, only about six percent will result in the reported rapists spending a single day in prison. And, uh, that's a little disturbing. Uh, when you think, okay, you go to the police, you're trying to do the right thing, and then you only have a six percent chance at justice when you look at the falloff here. There's a lot of different things happening. Uh, sometimes it's survivors will get discouraged once they learn how hard this process is going to be and that they're going to have to kind of tell their story again and again and again and that their credibility is going to be attacked throughout. They tend to not want to, uh, go ahead with it. Uh, some of it's that often prosecutors don't really want to take a case that they don't think they can win.

And it's really hard to prove the majority of sexual assault cases. Because generally what you have is you have the word of the survivor, um, if they remember what happened and then you have the word of the perpetrator who says that they didn't do it and that's all you got. And when that's all you got, usually a prosecutor won't actually take the case. They'll only take the case. Usually if there's physical evidence or if they can demonstrate a pattern of behavior by the perpetrator. Um. Now, this conviction rate is actually similar to a lot of crimes. It's similar to what it is in broad burglary cases, but in burglary. The question is not did a crime occur or not? The question is who committed the crime? In the case of sexual assault, over 85 percent of the time survivors who actually know their assailant in college cases. So it's not the who. It's the do you actually believe this person that this thing occurred? Uh, do they remember it? And if they don't remember it, then how can they prove that it wasn't consensual? Which is, uh, hard because in the case of either injury, it leading to concussion or memory loss or drugs, which are used as part of the assault leading to memory loss, then it still remains really hard to prove a case.

So what does this mean, then, for society? What is this underreporting and blaming for society? It means that there's very little deterrent to assault and there aren't many studies asking college men about whether or not they'd commit rape if they thought they could get away with it. But the two that have been done both found a rate of around one in three men saying that they would. And that's scary. And it sounds too high. It sounds too high for me. It probably sounds too high for you. So I invite you to read the study itself. It was in Violence and Gender in 2014 by Edwards at all. It was a pretty small sample size. But when you asked, again, not asking would you rape someone, but would you do the behaviors that constitute rape? You found one in three college men in this setting saying that they would. So deterrent. Is important. Um, the other impact of underreporting is that repeat per person, per perpetrators, which we think might probably make up the vast majority of sexual assaults, but we don't really know. Uh. Continue to assault. The other aspect is that we don't have great data. So we have these two studies. But when you look at the data from actual reports, there's so few of them. And it tends to be only the worst of the worst cases that even get reported that it really inhibits our understanding of what is this look like? What do we do about it? How do we stop it from happening again? Now, there are a lot of different reasons why survivors don't report.

One is that they're worried that they won't be believed. And that is not a totally unfair worry given what happens when they do report to their school or the police. They lack certain information about the reporting process. Often it's a little hard to know. Like, oh, you're assaulted. So you go to campus security or you call nine one one. But then what happens? Who finds out? Do your parents find out? Do your friends find out? Does the whole world find out? Will this hurt your job? Prospects in the world will always be linked to your name. And that's really scary. They're often not sure if it's serious enough to report. And this is actually probably the number one reason for college survivors not reporting their assault. Often when we think assault and rape, we associate it with stranger rape and we associate it with violence. And that is not with the majority of campus assault. Looks like. They don't really want others to know often. And this has to do with not wanting others to see one as a victim. It also has to do with maybe not wanting to get the perpetrator in trouble, particularly if it's a friend or a family member or you're not sure if it's serious enough.

Right. You don't really want to get them in trouble for something that might not be that big of a deal and that you lack certain information about the assailant and by lack information so that you don't know who they are. Right. You know who they are. For the most part, it's that you don't know why they did what they did. And this motivator, this understanding, what what is it that caused this person to do? This thing is really important when you're trying to decide whether or not to report them, particularly when it might be somebody who you care about in other contexts. Now we get to the game theory portion of this. So, uh, uh, within game theory, there is a term called first mover disadvantage, which means that if you are the first one to take action on something, sometimes that is not advantageous for you. So one of the ways this comes up in the case of sexual assault is fear of retaliation. So often the first person who comes forward against a repeat assailant, particularly publicly, will be shamed. They'll be branded either a nut or a slut. So they'll be told the people will try to frame them as being crazy or being really promiscuous and being somebody who is, you know, making it up. Uh. And the other disadvantage is wrongdoing, uncertainty.

So, again, this sort of not being sure whether or not a wrong was actually done, not being sure how to categorize that, particularly if you're the first person to express concern about this wrong. And then there's the stag hunt problem, which is really this or sometimes called the hungry penguin problem where. And with the hungry penguin problem, you have all these little penguins, they're hanging on the iceberg and there's some seals down there and they're they don't really know what they want to go down. So there's some seals. And so they'll just kind of nudge one penguin off to see if it gets eaten. And that really sucks for that first penguin. Right. And something similar kind of can happen in the case of sexual assault. Not intentionally, but that the first person who comes forward. Right. Often pays the harshest consequences, particularly if they're not joined by others. And they have no way of knowing at the outset whether they will be joined by others. And what this means is that we're in a case where what is best for the individual is not best for the community, what is best for the community in the case of a sexual assault would be probably everybody reports. You should just report it will probably suck. Probably not be believed. Um, it probably won't result in anything. And it will be re traumatizing and really hard. But you should just do it because that's the best thing for the community.

But for a survivor who's been through a really traumatizing experience, being asked again to read, traumatized herself to give up control to a system that is broken in many ways is not fair. Um. And, uh. But there's this thing that's true that economists don't always put in their models, which is that individuals care about the community. Right. We're not always selfishly motivated. And that one of the main reasons why Survivor is you do a report despite all of these disincentives, is that they want to stop it from happening again. If they think that somebody might be a repeat offender, um, because they've been through that, they know how horrible it is. So even errors. And Kate Unkovic is a law professor at Yale Law and Kate is a PTSD student at Berkeley, but used to study with Ian. They wrote this really wonderful paper in the Michigan Law Review in 2012 called The Information Escrows. And I really encourage you to read it if you like, to nerd out about this type of stuff, because they do a really wonderful job of explaining the theoretical underpinnings of this and what are different applications of information escrows and what our information, escrows, information escrows are a third party service that sort of stores data for you and that if something happens, it will release that data to another party.

So examples of this are commitment escrows, which is, um, like if I say to you that I will do this thing and you say to me that you will do this thing if I don't do what I committed to do, you don't do what you committed to do. So, um, a case of that could be sort of blackmail. Right. So I promise to hold onto this data for you. If you pay me money. If you stop paying your money. Oh, all right. Release it. Uh, no. Example's Posthumus escrows. So a will. Uh, when I die, information is released. Another example is shared interest escrows, which is you only let two people know that they are both interested if they both are. So Tinder would be an example of this. Right. If both people. Right. Swiped, then it lets you know. But otherwise it doesn't. And then the last piece, which is what this really focuses on. This paper focuses on is allegation escrows. And which is where if two people file a complaint against the same person, uh, it goes forward and indicate to a really wonderful job of modeling out the different scenarios where allegation ESCOs might be used and how much more likely they would be to lead to somebody being reported and to lead to multiple survivors, multiple victims coming out against the same person. And basically what they found is that information escrows work best where you have a high probability of multiple harassers or offenders, where you have a lower probability of direct complaints, meaning not that many people directly go to the police or to whoever to say that they experience something that you don't have many people who, once the escrow is available, would stop directly reporting.

So you could worry that, you know, if you provide this alternate system to only report if there's a match that people wouldn't directly report anymore. And where you have a high proportion of people who aren't reporting right now who now enter into the escrow. And there are a few other sort of costs and benefits to think about, um, when we're thinking about the application of these. So one is claim selection that right now, when something is directly reported, it tends to be the worst of the worst. Right. It tends to be the really bad, uh, assaults or the really bad human rights violation that's reported. And what isn't reported tends to be the lower level stuff. And what's hard about the argument of just everybody should report is that police, for example, they just don't have a Knux enough sex crimes detectives to investigate every case. And so they'll often try to discourage survivors who have the sort of lower, lower weight cases from going forward because they just don't have the time to look into it. And what an information escrow means is that then the claims that come forward are those that are likely to lead to action.

Right. It's either the direct complaints, which tend to be the worst of the worst or the match complaints, which even though there are lower level in terms of the degree of pain that was caused or the degree of seriousness that's judged by others. Uh. You're able to have that sort of corroborating information. The other Bedfords, this single venue. So having all of these things go through a single way of reporting can be really positive to aggregate this information. Right now, what can often happen is depending on who you report to, they probably don't know about the other place that the same person was reported. So, for example, one thing that will happen with campus sexual assault is a given person will transfer schools. And the second school doesn't know that this person had already had previous complaints filed against them at the previous school. So if you can have a single venue that matches cases, that can be really helpful. And the other is quality of evidence. So chances are if you were to put information into an allegation as grow like this, when evidence is submitted, it would be submitted sooner. And that means you would be of higher quality because you'd be closer to the time of the event. It be easier to recall all of the details about the wrongdoing and the costs of information as grows are one is an imbalance of stales staleness.

So the benefit that you're providing to the victim, you're not providing to the accused. The victim is able to (use allegation escrows) to store certain information and save it for (later). The accused doesn't get the same benefit unless they are writing down every consensual sex act they've ever had. There is a concern about equity. That (allegation escrows) might lead to less weight on uncorroborated claims. that come in directly. And nobody else reports people might take them less seriously and that we need to worry about false and bad faith complaints. Right. What is happening if somebody is, um, falsely reporting for whatever reason? And are you creating incentives when you create a allegation? Uh, escrow to encourage these bad faith complaints? So how did we implement this? Uh, about seven weeks ago, we launched Callisto. Callisto is an online sexual assault reporting system for college campuses, and it really has four main components. The first is just information. It's just information about the process. Um, what will happen if you report to a given entity how it works? What happens after you report who is involved? Uh, what confidentiality does and doesn't exist. It includes a ability to directly report electronically. So filling out a form is sending it to. In this case, the school administration, but hopefully eventually the police.

Uh, right. There's no reason to force people to make their claim in person. If you can make it easier online. A way to create and store a timestamped record of an assault so that if people do wait 11 months or longer to report, they can still record what happened. So we have sort of a form that survivors fill out. It asks a lot of the same questions that would be needed for an investigation. And Stav said as a timestamp document that they can delete later, that they can add to later, that they can download as a PDAF or that they can electronically submit. And then the last piece is the matching piece and how the matching would work. Although I think you probably well, let's walk through a scenario is let's say Student A says. I only want to report student X is if they assaulted somebody else. So they enter their own contact information and student X's identity into the system. What that looks like is that they first verify their identity with their university email address. They then enter their preferred contact information. So in the event of a match, would they like to be called? They'd like to be emailed. Can you leave a voicemail message? Can you, uh, what time of day would they prefer to be contacted? And they enter the perpetrator information for matching.

And we ask them for their name, but their name is not actually for the match. What's used for the match right now is either Facebook. You are out because it's unique. It's easy for college students to find or campus email address. And depending on the campus, we're using different identifiers. Other unique identifiers that could be used would be user names on different types of social media, cell phone numbers. Right. There's certain pieces of information that stick around with us and that are becoming increasingly easy for other people to find. And then they confirm the report submission. And at this point, they enter their secret key. And briefly, what a secret key is, is it? It is a key that the survivor has that we don't have that is used to encrypt the details of their record other than those that are needed for matching. So a lot of the details of the narrative there of their assault are encrypted with this key so that in the event of a subpoena, not even we are able to access. Section eight, does this do and B does the same thing, system finds a match, and at that time the school staff is notified of the match and they're sent the contact information and preferences of both survivors and both survivors are notified. They're not notified who each other are. They're just notified that there's been a match that they should expect to be contacted.

And what's probably going to happen next? And so what we're really trying to do through Callisto is on the individual level, really try to give people the best chance they can at justice and give them a way of catharsis. Right. Even if it's just writing it down and saving it. That can be a step. It can be this sense of feeling like I took a step in the right direction. Uh, and, uh. And if they do want to report, we hope that the Times damping the evidence as well as the multiple survivors of the same assailant. Well, we'll really improve their experience for the community. This can be helpful because we're hoping that having an increased reporting rate will act as a deterrent or just even knowing about this system. Whether or not it even increases the reporting rate would act as a deterrent, that we hope that we'd be able to stop repeat offenders earlier. Right. As we saw earlier, if we take leasebacks, uh, research at face value, if we are able to stop offenders even after their second assault. Right. They already got away with, too. We would still prevent basically 60 percent of sexual assaults. And the last the last sort of theory of change application is data. So among these non reported assaults, we can get certain information from the form that survivors fill out. Uh, mostly just the answers to the multiple choice questions.

And we can aggregate that and give that back to the school. And when we do that, we have to be really careful not to out any given survivor through that aggregate data. Um, I'm happy to talk a little bit later about how we do that. But it means that that schools can start to have information about what is really happening, what time of year our assaults are occurring. Is it seniors assaulting freshmen? Is it freshmen assaulting freshmen so that they can create better prevention policies? We just launched on two campuses. We already have a survivors using the system of survivors that are creating records. Half are either putting it into matching or officially reporting it. And we have already officially reported through us a complaint that we were told would not have happened had it not been for Callisto. And so we're working to sort of get Callisto into all the schools in the country, ideally someday. But we're also starting to look at other applications. So there's some settings where you could see a very similar system working like the U.S. military for sexual assault or harassment. But you could also see a similar system working for sexual harassment in the workplace or for sexual assault in commune's has been something that's been brought up US or group housing situations where people don't feel comfortable reporting other people who live with them. But might to somebody who is like an unofficial house heter child sexual abuse is something where you see a lot of repeat perpetration.

Right. You see a lot of cases coming forward and then them all floating together. And it's been going on for years and years and years. Um, but it's a little bit harder to figure out how to implement because it's not within a closed community or network where you can verify identity in the same way you could see it being used for sexual harassment in an industry. Um, you all might have seen there's a big scandal in the astronomy world these days of a pretty well respected professor having sexually harassed lots and lots of students over the years. And it just kind of broke. And in the case of an industry, rather than sort of in a workplace, you could see those types of reports going to some association like the American Astronomical Society rather than the workplace. Uh, you could also see that working in with reports, going to unions, uh, instead of your workplace, if you're worried that H.R. might not take it seriously, uh, you could see it being used in supply chain management. So for companies that care a lot about the ethics being used in their factories, having the reports driving, going to the factory directors who might be sort of complicit in the harassment or wrongdoing, uh, it could go up to their boss's boss, but you could also see it beyond sort of the sexual health cases, which is where my nonprofit focus is to political corruption, corporate whistleblowing, war crimes, improve police brutality.

Right. There's a lot of things that we don't want to report. But we would potentially, if we knew that there would, other people would come with us. And that's where sort of I get excited about the broader implications of Callisto. We will be open source and Callisto in about a month. And we're hoping that lots of people can take it and apply it to other other scenarios, either. Either these or these are ones that people can make up. Uh. Because I think it is a really interesting idea. And we have yet to really see how much it'll pan out. But it could be pretty huge in terms of bringing voice and breaking the silence around a lot of these issues that we don't talk about and that go on, um, in our society and harm people. So thank you all. Uh, and I wanted to open it up a bit for questions and a little bit for a discussion in terms of what it could be, other applications of a system like this and what would be sort of worries around whether or not it would work in other applications versus, um, causes of excitement, because there's a lot of logistical stuff that has to work for this to work.

Can you talk a little bit about how you store the data? I think you mentioned that you could go in to that.

Mm hmm. So we act as the third party storage system so the school doesn't store the data. We store the data. We, uh, encrypt it with the secret key. Everything is stored with a username and a password associated with it. So you have your username, your password, your secret key user experience. Why is this isn't great? Because you have to remember those three pieces of information, the password. You can reset the rest of it. You can't, um. And if you lose your secret key, it's lost. The whole thing's lost. So we think of it like a security deposit box inside of a bank where the bank, uh. All of your details of your shoulder and security deposit box. And you you're the. With the key. The information that is the aggregate data were storing as counts. So any time that somebody fills out a multiple choice question and they're saying on campus, off campus, we sort of that counter goes up. That's our way of trying to keep track of it without it bringing back to us being able to individually identify a record that might out a survivor. And then the information that for the matching, like the identifier of the assailant and the contact information of the survivor, that is automatically sent in event of a match that we encrypt, but not with a private key. We encrypt it with Archey. And, um, we have zero strict controls around when we allow ourselves to decrypt decrypt that data.

Just click over here. Hi. Feedback from nothing.

I'm really quick, I'm just interested in the sort of analytics that you perform on the data. And particularly that if there's some sort of, you know, text analysis, natural language processing going on there, just because, you know, with the sort of. You know, the text that's inputted there. That could lead to potential matches which might be separate from the actual name identifier. And so, you know, um. And so I wonder if that is part of what you're doing as well. And also kind of related to that. Like, what do you think this could be a sort of an evidence collecting tool as well. And I don't know if there's a potential for, um, you know, adding pictures or any anything's like anything like that. But, um, I wonder if that's also in the mix.

Yeah. So for the evidence collecting part, uh, the original plan was to include that in the MVP, the viable products that we just launched, we didn't have time. We want to be really careful around the security of it and figuring out can we protect the security of that information. But you definitely could upload evidence. Theoretically, that could be video. That could be photos. Um, that could just be information about if you had a rape kit and where it was done. Um, we do invite people to write down whether or not anybody else sort of witnessed anything or if they told anyone else afterwards. Uh, you could also see a scenario where people are inviting other witnesses to write down what they saw as well, um, or where you're creating a sort of another version of it that allows witnesses to record what happened even if it didn't happen to them. All of those are potentially on the table at some point. And then the first question, text analysis. Yes. So we don't have enough data right now. Um, I think that once we hope that we're in a lot of schools and a lot of places, that is something that we could think about. Uh, right now, when we're encrypting the data, we'd we'd have to sort of do it before encrypting it and storing it to look for those words. Um. But if we could figure out how to do that, well, I think it'd be really interesting. And and you could start to look for patterns. So if you say only had physical descriptors about somebody, um, like a type of tattoo or eye color, hair color, height, and you felt like there was enough information you could potentially flag something. Um, the difficulty becomes a little bit. And what is enough of a flag to trigger an official report? Um, and or do you just sort of flag it for the individual and you let them know how many things that's been matched on and then they can decide. Um, so that that'll be interesting to see how that plays out.

Thank you. This is amazing, Weatherbee, um. Are you finding it difficult to create awareness in terms of how not in terms of what the product is, but in terms of, you know, hey, is my deal gonna be secure? Hey, what is the process like? Are they going to report it automatically, like, you know, that kind of awareness? Are you finding it difficult to create that around how the system works? Yeah, yeah, exactly. I mean, you know, are people more saying, well, um, you know, if I reported, is it really going to stay private or is it going to you know, at some point you're going to have to be like you said, you know, you might get subpoenaed or something like that. Yeah. And you'll have to report that information.

Yeah. We don't really know yet. Um, what we did do on our campuses when we launched seven weeks ago is we went to freshman orientation. We trained the arrays. We put sort of stickers and flyers and brochures everywhere. Uh, and we're linked to from the university email address. What we're trying to do is go in and train and have conversations with a lot of the people who survivers would likely come into contact with either people who work in the health centers are a student groups that do confidential counseling to make them aware of it. Uh, we also have text on the Web site that explains sort of how the matching works, but I don't know how persuasive it is. And when we see people who come to the home page at but then don't start a record, we don't know why. Uh, so what we are planning on doing is focus groups on campus and a year end survey of students to see if they know about the product, um, if they experience sexual assault and if they visited at all. And did they use it and why why not? And that might help us understand that a little better.

I've got a question for you. So in thinking about some of the other applications of escrows like this, several of us here actually are reading a book about the advent of statistical knowledge and like mass media surveys, like mass public surveys at the turn of the century. And it got me thinking about I was thinking about it a little bit while you were talking because like one of the chapters talks a lot about the Kinsey Report. Right. And the idea that, like generating this knowledge about what's normal practice was really like received by the public and this really interesting way. Right, where people thought maybe didn't feel like they were normal and then they found out what anyone else was doing. Then they felt not right. Or maybe they felt more abnormal. Right. But like that knowledge, like really that was knowledge that wasn't shared. That wasn't common knowledge before until that report came out. You could imagine maybe using ESCOs like this for things like that or thing. Probably more. Not so much. Kinsey aske stuff, but things that have to do with personal identity. So in context in which it's, you know, very stigmatized to come out, you could imagine it seems weird like register, you know, your sexual preference. Right. But something like that. Right. Where if you know that there are other people who share the preferences that you have or, you know, share characteristics with you, that that a system like this might make those processes easier for you.

Yeah, I think it's well. So now that mass surveys are happening more and more. It's interesting to see what's coming out about what's true of our world and what questions we do and don't ask. And it was shocking to me when even asking about sexual assault perpetration. How many people would on a survey say, yes, that they did do that? And, um, so I think one thing about surveys is always hard as well. How you interpret low response rates or missing data and stuff like that. But, uh, also, will people tell the truth? And people seem to offend. So it actually, I think, often tends to be almost a better source of information than something like Callisto, which is a bit self selected. But you do see through the Internet, particularly all of these subgroups finding each other. And you do see, um, that acting as a form of support through the Internet. Right. But can also lead to harm because people sort of end up in these echo chambers where they only hear each other. So when it's sort of a negative, um, sentiment that can be harmful, uh, versus one where you feel like reducing your isolation, which is obviously a great thing.

In terms of the actual report, do you ask any specific questions or is it kind of a blank form for them to fill out? And how did you decide which way to go? So we do ask specific questions.

Um, the the questions that we ask all follow, um, certain of flow that's used in trauma informed investigation. So there's a whole field that looks like when you're trying to interview somebody about a case of trauma, what questions do you ask? In what order do you ask them? How do you phrase things? And so we tried to mirror that and what we did. And you skip logic to not be asking questions that don't apply. Um, the, uh. One of the interesting things going through this was that I you really want to make sure that you're starting it with easy questions. It's similar to if you're doing survey design, you often start with the demographic questions you like kind of trying to build rapport through with survey design. The same thing ends up here where you start with sort of the more vague. And as people are are getting more into answering the questions, you're building their sense of confidence, you're building their sense of empowerment. And so that by the time that you're asking them, they're really sensitive stuff. Hopefully they're feeling more comfortable with this process versus asking right away, like who was it who was involved? Um.

Um, yeah. First off, I want to say it's really exciting to see this live. And this gets a little bit beyond what it's implemented as now. But I was talking to someone yesterday and she's from Afghanistan and she was telling a story about recently there was something that happened to a young woman and someone I like a terrible situation. Someone posted a video of it on Facebook and that video went viral and ends up I incriminating like 11 men that were involved in this case. And they go to jail. But then she's I asked her. And so, like, you know, what happens is men in jail and she's like, well, you know, oftentimes someone will bribe and they'll get out of jail. And so I guess my question is, what happens when we want to report things? But the police system isn't to be trusted. And I'm wondering if this can be used to also be creating like some sort of community justice body in different contexts, or if you guys have thought about that.

Yeah, I do think it's an interesting, interesting to think through what our other forms of justice when so many of our systems are broken and when there's so much sort of corruption or misaligned incentives. And a lot of the people that you report to this would include in cases of, say, reporting sexual assault to H.R. right or to somebodies manager is they might not want to take action for a lot of different reasons. And then what do you do? So in the case of that, you can see, OK, professional associations being a place to go or a union being a place to go. Right. There's other bodies that have might have incentives that are a little more aligned with that of the common good. In the case of sexual violence, if the police won't do anything and if the school won't do anything.

It's really hard to know because unless you wanted to create like a vigilante group, which is like, yeah, maybe not the best of it would be bad ass, but prolly not the best idea. Um then then it sort of. Who did those reports to go to.

And sometimes some people will say, well like what if you published it so you could see a case of if you have enough corroborating reports and no action is, is taken. Um, people going public like what happened with Cosby is if you tell people enough and nobody believes you going public and having the mass number of people to come with, you can result in action even if it doesn't result in justice in the criminal courts. So that that's one solution that could maybe work. That's definitely not something that we're planning on doing through Callisto. Um, but it is interesting to think through what are other ways to cause action that is hopefully positive to happen. Another way you could implement it is rather than a report going to anywhere anonymously connecting the survivors together, um, and then they can decide together what to do. The issue is if they then decide to report often that will be used against them because then they're seen as colluding and they're less likely to be taken seriously if you connect to them. The other aspect of it is a little bit, um, how do you ensure safety in that? So if you're anonymously connecting somebody, you have no way to verify reports when they come in, then you could be connecting like the perpetrator who's trying to figure out whether or not the person that he or she assaulted reported them together and that could result in harm for the victim. And so it's a bit of an edge case. It's hard to know how worried about that to be. But you could see misuse or you could see people fishing the system, which is deterred in part by having a report go to an official body.

And would you just say that because I'm thinking about so using this system not to report to anyone, but in case of corruption, for instance, in in many instances where the idea is less to, it's difficult to criminalize it. But maybe if the eve this you or the higher instances are maybe told that something is going to happen. But people need to to know that they haven't maybe they haven't seen really well. So they need to know that there are several who have witnessed the same thing. And so they can report it. And so do you think that this problem that you said, could you re explain it, the fact that if people are connected together, it may be simply called inducing? Do you think it may happen to in other instances like for corruption, not for sexual assault, potentially?

I think it depends a little bit how you connect to them. So let's say instead of allowing them to anonymously chat and send messages back and forth, all I do is I say, hey, there are five other people who've witnessed the same thing. If they're in, are you in? And like, what would be your threshold? Like, would you only be in if there are three others that are in, would you be in if there are one over there at in and then sort of gauging that and then automatically, if enough are in taking some action. Right. Which could be publicly shaming somebody and having people be able to anonymously back up their stories with the public, it could be going to journalists. One thing that is done in some cases is sort of crowdsourcing and mapping information about corruption like bribes. And so being able to see that on a map that's probably more easily done when it's what you're reporting is an area or you're reporting a company of wrongdoing than an individual. And I think the difficulty becomes a little bit as is liable. And in what case can somebody be accused of libel and what case it is somebody.

Potentially harmed by making some information public and in those cases where it's not so sure as there was nothing to send to any police or college officer, did you think that you might have a free rider problem where people just see that they're five or six or seven people and then they're like, oh, that's great, someone else is going to do it and I won't have to suffer the even if you were many, even if it's successful, you might suffer from it. Through many ways is and the B. Oh so it's great. I don't have to bother about it. I know that it's a very only rational way to think about it. You may have been might happen still.

So yeah, I think that's true. I think ideally what you would have is some enforcement mechanism. So if you say you're gonna come forward, if a certain number of people do and then that number people do, it's automated. So that's one thing that we do with Callisto's. If there's a match, it's automatically sent. Because if you have that case of like, hey, there's a match, other people are going forward, then maybe I ought to.

So as bad as the police are sometimes with these problems or with with handling sexual assault on college administrators can be even worse. They're just not you know, they have them and trained to handle evidence. They.

It's just.

Some random economics professor in some cases has banned the person who runs the bookstore and these committees aren't trained up. Do you require anything of your partners or are you just reporting service? Like, is there you know, if if someone's handling of a case is so egregious in the future that you've reported, would you not offer your service within that university anymore?

We don't we don't have this time. I could see down the road using Callisto once it sort of becomes a standard of practice as a way where we say we will only work with you if X, Y, Z happens or if your policy ensures that X, Y, Z will happen once a report is be submitted. We don't at this time.

Could you come back to this screen where the person is first reporting the identity of their perpetrator?

It seemed like it was reliant on Facebook.

Oh, yes. Yeah. So so what? So that sort of assumes that your perpetrator, it makes an implicit assumption that your perpetrator is someone who might be a student. And if it is like the the head of the bookstore or the security guard or a professor or something like that, they might not have Facebook. So I'm sure you've thought of that. Like, how did you how do you think around that? Yeah, with that.

So the way we are rived at Facebook for this iteration was. Callisto, in its current version, is designed for the use case of student being assaulted by other student and it most likely being a single act of assault. Between the two of them, that is a case. It is a common case, but certainly not the only case. When we asked students about what information they had about other students, Facebook, your elves came by far, far and away compared to the other things, the thing that most likely they were going to have. We sort of asked students, do you have Facebook? And they were like, well, yeah, I'm alive. I don't use it all the time. Talk to my friends. But they all had it. Other identifiers that could be used include campus email address. So at Pomona College, which is one of our pilots we use, we have this field that you thought you can also fill up, fill out campus email address. And if there's a match on either, it'll trigger you could see it being like a ton of different options, like fill out in all you have. You might worry a little bit, but the more options you have, the less likely that somebody would be to fill each individual one out. So there's there's some worry there, but, um, you could have a phone number, etc. You could also have a directory. The reason why we don't have a directory is in part because of this law called Phurba, which regulates what information schools can share with us. But if you had some sort of directory of all of the people in New York that I could search by certain characteristics and find the unique person based on, I don't know, five different piece of information about them, um, and maybe somebody says, like, I'm 100 percent confidence versus 90 percent confident. This is the person I don't know. But there could be other ways to identify uniquely individuals.

But it's just interesting that, like, this doesn't include things like identifying characteristics or something like that. Like you are kind of relying on somebody knowing a fair bit like the last name someone is a pretty high bar or having to really make themselves vulnerable and ask a lot of people around. Right.

So in the case of campus sexual assault, over 85 percent of survivors know their assailant like they know them personally. They can find their Facebook URL. So the cases outside of that, we can't at this time match on because we don't if you don't know them, we don't because we're not using other methods for matching or combination of data to match. We can't do the match. You can actually directly report. You can save it. You can use everything else. But you can't trigger the match unless you know them. But that is most cases of college assault. It's not all of them.

So I just have a quick follow up question about this. I don't understand the need to rely on Facebook at all. I don't. I mean, even even if you had several fields, you know, you just had one field that you could dump stuff into it. Searching and matching algorithms are good enough that you should be able to put together, at least with some reliability factor or a probability factor, that you've got a really good belief that you have a match and then a person, a person could intervene and decide if that's a real match or if it's just something that looks like a match. But I I'm guessing that you have a reason why you can't do that. Or is it because you're storing an encrypted? Is it like a design issue or is it just a choice?

I think some of it is. It's an edge case that for the MVP, we just didn't have the time to think through all of the different ways to do it. I think some of it's a little hard to know what those things would be and how confident you have to be in order to trigger the report. Right now, the messaging is pretty clear. It's like if there's a match on this unique identifier, then it'll go forward. If it was a case of fill out all of these fields and if we're like 90 percent confident, then we'll match like that could work. It's just a little bit less clear of message. But be curious as to what what I dunno what other pieces of information you think would work to uniquely identify it.

And I think you get it when I think you get it with name and where it happened and you're close enough like like you could figure it out from there. That's I think.

But I mean, you know, at some point.

Isn't that a credibility issue also that at some point, you know, you find a 90 percent match and then you say, okay, this is who the person is. But that turns out to be, you know, the Black Swan problem and you have the 10 percent, which is the wrong one. And then now people start saying, well, these guys are just reporting the wrong guys and you're done, you know?

So.

I just think it where the initial report stage like that's an issue at trial or that's an issue or an evaluation, but it's an initial report like you could even just say, hey, looks like 90 percent certain. This is the. This is the person. And then at that point, follow up for additional information. It just seems like it just seems like it initial collection, relying on Facebook weeds out. So many people like it reads out so many people that it that it just seems like if all you had was name and even like zip code, I bet you could get it. Like it's not like there's that many. Exactly the same names in know in a specific geographic location.

Yeah. And it depends a lot on the university's Hoover small schools. It's unlikely that somebody would have the same same first and last name. In the case of a large university, you could have like 20 different Bobby leaves and you might not know the last name. People are probably able. Well, although I guess if you can find the Facebook, you're all you can probably find a way to get out. But, yeah, I think it's I think it's worth considering. I mean, you guys bring up really good points.

Hi. Um. Yeah. I have a question about just I mean, obviously, as were technologies becoming embedded in our lives, like we're all evolving from old school methods of reporting about what not into kind of like a new new school models. And as that transition takes place, um, I'm wondering what you're doing to kind of work with the groups on the ground that are maybe doing the old school version of these things. And I'll give you an example so that someone dealing with these issues is not getting mix. You know, there's not confusion as to where do I go first? What do I do? For example, in New York City, if a woman experiences sexual assault. You are correct with the statistics that the probability probability of them going to the police is pretty small. But there's like an organization called Safe Horizon that has a hotline and they have a lot of a lot of people calling in and putting complaints. And they do a lot of advocacy with like going to the police and kind of acting as a liaison. So and then you see a lot of like, you know, now you're seeing like complaint forms speak becoming into like you can do now through text messaging, or you can do it through like online confidential online portals.

So as there is like these organizations that are embedded in communities that people trust and people kind of see as a place to go to for like assistance with sexual assault or other issues. And then you have these new models coming in, like how do you plan to kind of both, um, you know, kind of get the buy in of like some of these traditional models that are taking place that are, you know, very trustworthy. And also because it's it's a really amazing thing that you guys have built, but also not confuse the person as to like where I'm experiencing this. Where do I go? Do I call the hotline at Safe Horizon? Because I know they'll help me with the police. And like, you know, other things that do I go on this Web site and do I submit a complaint? You know, how how do you negotiate that as like the months, you know, future?

Yeah. And I think it's one thing that we always struggle with when we have this reporting explorer putting options, which lays out of the things that our official reports. What are your different options? And as much as we can't explain how they're work, how they're different from each other. It's a lot to sift through because there's different definitions of sexual assault. There's different things that happen once you report. And then we have a sort of a get support part, which is sort of more about what our hotlines you can call very different types of support, including legal support you can get. And that's customized to the local area. And we often partner with organizations in the local area to let them know that this is a thing, so that if they have a survivor coming in and talking about Calista mentioning Questo to them, they know what it is and to make sure that the information that we have and get support is accurate. So, for example, in some cases, if you access certain types of support, they'll call the police. So if you want a forensic exam or rape kit done in San Francisco, they will call the police unless you ask them not to.

And you don't have to talk to the police. The police come. But they'll probably call in the case of around the L.A. area, around Pamona. You can't get a rape kit done without filing an official police report. So some of the important things about support is how is confidentiality work with us politically if you're under age? Particularly if you're worried about the police being called when you don't want them to. Uh. And that can have a big impact on whether or not people want to connect with those. Right now, we kind of the bottom of every page. We have sort of a hotline number to call. One thing that's definitely come up for us is a Web site is not a person, for better or worse. And going through this process of filling out this form can be really traumatizing. And, um, having someone with you, either when you're filling out the form or going to the police or whoever, it can be really helpful. So.

Mm hmm, hmm hmm hmm, yeah. So our hope is make sure these front grine front line groups know about it, see if they're comfortable using it. We could also see it being used in the case of when I come in in person to the to the school or I mean, the police would be nice that they could sit down with me and fill out this form with me as a way of starting my initial intake. You could also do that specifically with the matching. So you could skip the whole writing it down, documenting it. And if what the police really care about it, does this trigger a match or what school really cares about, then go straight to that. Just fill in that information.

Reporting is hard. It's not just hard because it's re triggering. It's hard because it's not just hard because you're not sure of an action will get taken. It's hard in part because it is an isolated activity or you have to put yourself in a position of vulnerability to an authority of power. And so part of what's at stake is you're trying to navigate this, is that you're trying to create these moments where you have some bond where like, you know, you're not alone in order to create that solidarity, but because of a whole set of reasons, you're not actually connecting these people. And I understand why I'm struggling with it, because I'm struggling with what happens when, like, you know, you're part of something, but you still experiencing it individually, you're still experiencing the act of reporting individual. You're slowly experience the act of filling this out individually. And I'm thinking back in the early 90s, I went to Brown University and one of the things that happened on Brown's campus was that there was an art, an art school, and it's called list. And inside the women's bathrooms and list, people started writing the names of men who had abused them on campus. And they would fill them in, they would add plus one and they would add so many just different notes to what was happening about how this did this.

And it was an interesting moment, how it created a sense of solidarity amongst a lot of women who then were willing to fight. They started talking enough to find one another to find that confidence come forward. At the same time, it created a huge other ramification on campus. Men threw hissy fits about who was being accused. There was all of this public outcry, and because it was anonymous reporting ended up happening, of course, was that the sentiment on campus was that men were wrongly accused and that will end because the women were not identified. And it was a hetero normative situation because the women weren't identified. What ended up happening was that they ended up finding each other and then going more into silence. So part of what I'm trying to get through in my head is how do you give the people who take the action to report enough of the support structure around them that the act of even when they're matched will give them the sense that they know they can go forward and they know they won't get screwed? And that's what I'm struggling with as I hear you account for this.

Yeah, I we face the same struggle. I don't have a clear answer. I think there are ways of trying to create support outside of this. So one of the things that Kat's working on is report a girl and it sort of storytelling. It's how can you share stories among survivors in a way that ideally protects their confidentiality so that you can reduce that feeling of isolation. Some campuses have groups you can go to as a student that can be confidential advisers and that can work with you through the process. So Pomona College has that. They have a group of students, many of whom are survivors themselves, who have been trained on how to go through this and and are meant to be your advocate. But it's not available on all campus. And you don't necessarily want to go to your fellow student who knows who you are and probably knows your assailant in order to help you out. So your ideal situation is probably somebody employed by the college who knows what they're doing, who's meant to be a confidential out of advocate. Maybe most schools don't have that. Most schools right now, everyone's a mandated reporter, meaning that anyone you go to, including our aides and tell about your salt, has to report it officially. So you can't unless you're ready to report, you can't get that confidential support you need unless it's through your fellow students or through the Internet or through another sort of support service you're contacting. So it's it's a real challenge, definitely.

So I was wondering about your planned expansion, you're in two schools now. What do you see as the biggest challenges to convincing other schools to sign on? Or is it just a matter of time before all college campuses have this?

I like to think it's a matter of time.

I think when I think there's a few things happening with schools, I think when they don't want it, I think some of its liability. So it's worrying about a third party is now holding really sensitive data about my students. I think some of its concern about actually increasing the number of reports right now, suddenly this is blowing up on college campuses. Suddenly they're employing people to work on this. Those people have very small budgets and not enough staff. Um, and often they're finding it challenging to handle the number of cases that they do get, even though that number is relatively small. So saying we'll increase your number of reports when they don't have the capacity to take in those reports is like not a great value proposition for them necessarily. Um, so I think that those are a few of the things going on. Um, and and the economic incentives aren't super great. So. Sure, schools are far more likely to be sued by somebody they expel than they are by a survivor of sexual assault. So expelling more students doesn't necessarily help their bottom line in the short term, even if in the long term it might decrease assault, which would then make everybody happier.

Hi, um, I'm just gonna wanna go back to the Facebook issue, um, just really quickly, I'm wondering whether or not, um, as you push this forward, whether or not you could see if there are other consequences of that. So I could imagine somebody going to a Facebook. You are all seeing all the friends they have in common with this person and that actually deterring them from coming forward because they can see very easily how deeply their life will be affected. Um, if they do come forward. So as you move like as you start doing this, trying to see whether or not that actually decreases rates of reporting, uh, would be really interesting. Yeah, I think that's a great point.

Another concern that we have about the Facebook, you're all it's Facebook feels public like even if this isn't gonna push to my wall. Right. Just putting in a Facebook your can feel public. And so, um, I. I'm also concerned that. Yeah. Being asked for that as the thing first, uh, might might worry people. Oh. What are you going to do with this information. Um.

Ed, you had a follow up question about the Facebook thing. First of all, mistaken identity, because so many people on Facebook and the same name, you might not actually be looking at where they are geographically. But second of all, you talked about this being a deterrent, just the fact that there was a reporting system. But once it becomes known that the Facebook Eurail is part of it, you know, what's the chance that people just take down their Facebook pages? Mm hmm.

Yeah, it could definitely happen. And I think right with Facebook, you're out. The nice thing about Facebook is you have name and you have social contacts and you have school. And so that's a lot of data about somebody to make you pretty sure it's the person, even if you're not Facebook friends with them. Um, but part of the reason why at at, uh, Pamona we have more than one identifier is because we're like, well, more is probably better and we want to see how much having multiple types of identifiers means that people don't fill out both. Um, and. And as we get more schools will have more flexibility to think through what are different identifiers that can be used in different settings. What work? Well, what don't. And over time, this will change. Right. Like Facebook. I don't know that college students will always have Facebook accounts for many reasons, not just because they don't want to be reported.

My. We can do one more.

Um, great presentation. Uh, I'm really curious about the deterrent idea. And how can you potentially measure culture beliefs campus wide and how Callisto is potentially shifting, shifting that? Yeah.

It's hard. Uh, it's really hard to do anything causal there. You can measure rates over time. And campus climate surveys are this thing they're doing.

Being done now more and more on campuses and might be mandated soon where our colleges are sending out surveys to their student bodies. Um. Most of those are just about have you experienced this? But some of the new ones that are being created by awesome researchers and open sourced also include questions about perpetration and attitudes. And so you could try to see shifts over time on those. Um, obviously, it's really hard to say if that's due to Callisto or other factors, but you might have enough of them. Actually, that one thing I would love to do at some point is, um, of schools who are all using the same survey methods, uh, randomly select which ones we promote Callisto to. Not all of those will adopt it, but, uh, they'll all have data using the same measurement techniques to do the same surveys. So you kind of can have like a cluster randomized sample kind of.

Somebody help me.

Thank you so much just for coming and talking with us today. It's super interesting. Um, let's all thank.

You're not to be.

Sonix is the world’s most advanced automated transcription, translation, and subtitling platform. Fast, accurate, and affordable.

Automatically convert your mp3 files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.

Sonix has many features that you'd love including automatic transcription software, automated translation, share transcripts, advanced search, and easily transcribe your Zoom meetings. Try Sonix for free today.