Thoughts on Theses: Oren Tirschwell ’25

Q: To start off, could you tell me a little bit about yourself?

A: I’m a senior, and I am a double major in statistics and an interdisciplinary major called Applications of Data Science to Public Policy …  combining the complete statistics major, three-quarters of the math major, half of the computer science major, and a little bit of poli[tical] sci[ence]. I have four advisors that sit on my thesis committee … I’m also a cellist …Those are sort of my two big things: stats and cello.

Q: Nice! So did you create your second major

A: Yeah.

Q: How does that work

A: The [way the] interdisciplinary major process [works] is essentially you need to propose a program of study that’s not currently accounted for by any existing major. So in my case, we don’t offer a data science major. The course of study that I was charting was really a lot more than just … statistics and math, because data science is really about how you blend those skills together. I was spending a lot of time in my various classes focused on [this] … In the fall of my junior year, I wrote a proposal, … I collected the advisors who are willing to support this endeavor, sent the proposal off to the interdisciplinary major committee, and they approved it.

Q: Cool. And what’s your thesis about?

A: I’m writing about gun violence and specifically about access to firearms in the home. Research has shown for decades that having a firearm in your home makes your home less safe. It increases risk of firearm homicide, firearm suicide and domestic violence. But survey data consistently shows that the American public doesn’t believe this, and in fact, they believe that … having a gun in their home makes them more safe. At the end of the day, it’s really buying into the N[ational] R[ifle] A[ssocation] … slogan of “the only way to stop a bad guy with a gun is a good guy with a gun.” I was very interested in this disconnect between what the … research was showing and what … the American public believes.

The purpose of my thesis is threefold. One is to conduct novel research into the public opinion side of things, so that’s going to include building on and probably refuting some prior analyses … [using] some machine learning techniques, and then analyzing some pre- and post-pandemic data, because there’s also a massive shift in the sort of demographics of gun buyers pre- and post-pandemic. Then there’s the empirical side, where, again … it may involve recreation and updating of existing analyses … It may involve some data mining … [to] allow for a much more robust analysis. One of the biggest challenges with gun violence data is the United States does not systematically record who purchases a gun, and it’s only one state, California, where that information is stored, and so it’s really hard to know from an individualized standpoint whether buying a gun is leading to a potential firearm homicide or suicide. In California, we can do some of that, but not really anywhere else, and so we have to rely on other statistical techniques to try and understand this.

And then the third piece is to … try and create some form of interactive, public-facing dashboard that a general audience member who doesn’t have formal statistical training and who isn’t … well researched on this topic would be able to engage with … and understand the data and the logic behind what a lot of people see as counterintuitive, that having a firearm in your home makes your home less safe. And to do that, I’m going to be relying heavily on tenets of behavioral psychology, using … Daniel Kahneman’s famous book, “Thinking Fast and Slow,” to try and think about, how do we use a combination of data and real stories … to persuade people? Because we know that a lot of research has shown that giving people numbers is not going to … [work], but providing them with anecdotes, with stories, with something that they can identify with themselves, is the most effective way to change minds and to teach people something new.

A lot of the statistical research that’s out there right now is extremely academic, extremely inaccessible, and so it’s [a] clear and simple way for a general audience member to understand.

Q: Wow, that’s really interesting. I was also interested in what you said about demographics of gun owners being very different pre- and post-pandemic. Would you be able to talk a little more about that?

A: Absolutely. It’s really, really important and really alarming. In 2019 it’s estimated that just over 13 million guns were sold in the United States. In 2020, that number jumped to almost 21 million. So we’re seeing … a  60 to 70% increase. And that trend continued the following year … It’s almost 60 million guns that were sold between 2020 and 2022. These are record high numbers. But it’s not only that more guns were being sold. It’s also that the demographics of who was purchasing the guns and what their ideologies are was a lot more extreme, and there are more and more surveys coming out now that [show] that belief in political violence is much, much higher among post-pandemic gun owners. Belief in conspiracy theory is much higher among post-pandemic gun owners … And multiple other surveys … are showing greater inclination towards violence. So we’re still very recently into this post-2020 gun owner era … There are all sorts of questions that we’re only just getting the data [for].

Q: Could you talk a little bit about how you chose the topic and what led you to it?

A: I knew that I wanted to be focused on data science, and I wanted the application piece to be a social justice related issue that I’m passionate about … I came down to two that I was particularly interested in: gun violence and educational inequality … Every single Amherst College student that went to school in the U.S. knows what to do in a lockdown drill … because guns are a massive epidemic in this country. And I think [the 2017 school shooting in] Parkland, [Florida] was a big moment for me … Those were people my age … It was something that … caused me to worry for myself, for the state of the country, and so it’s been something that I have thought a lot about since then … There’s a pressing nature to this matter that feels very physical … And I do believe that gun violence has been an underfunded and under-researched space for a long time, and that’s starting to change, but the NRA pushed really hard for this to not be considered a public health crisis in the ’90s, and so there was a massive reduction in funding [for research] for the decades to follow, and this [presidential] administration is the first that’s really pouring massive amounts of money back into it, and has established an office of gun violence prevention, which is wonderful … It’s still a relatively young field in terms of academic research. … That makes it more accessible in a certain sense, for an undergraduate student, because I can basically do a full literature review of all of the studies that have estimated the increased risk of homicide rates if you have guns in your home, because there aren’t actually that many of them over the past three decades.

Q: Could you talk more about what the advising process for an interdisciplinary thesis is like, and how having four advisors works?

A: It’s definitely … a little bit unusual, but I’m really lucky to have four super supportive and super smart advisors. My advisors are Professor [Jonathan] Obert, who’s in poli[tical] sci[ence], and he’s sort of serving as my primary advisor when it comes to thinking about gun violence and the poli[tical] sci[ence] side of things. Professor Kate Moore in the math department, who’s a wonderful mathematician, and her research focuses on machine learning and specifically clustering algorithms …Professor [Amy] Wagaman in the statistics department, who’s sort of serving as my primary statistical advisor for the thesis, does a lot of research into machine learning … And then Professor [Pamela] Matheson, who’s sort of my official primary advisor for the [interdisciplinary] major, and has … been extremely helpful in supporting me through all of that, as well as in thinking through statistical …[and] methodological questions.

Q: Sounds like you’ve got a really good team backing you up. That’s awesome. Would you recommend doing an interdisciplinary thesis for future students?

A: It is a process that I have really enjoyed. It is very difficult to get an interdisciplinary major approved at Amherst. My understanding is that it has gotten more difficult in recent years; that’s not backed in empirical fact, and maybe deserves some fact checking, but that has been something that I heard. I’ll also say, I was told “no” through a lot of this process … I’m really grateful to have advisors that believed in me and that were willing to support me through this process and help me figure out how to make it happen, because I was really passionate about it … I believed it was a reflection of the work and the academic path that I had taken at Amherst. I would say to any future student, if you find something that you’re passionate about and you don’t feel like it is contained within the current majors, absolutely. I have found that this major has pushed me, allowed me to grow as a scholar and an intellectual in a way that I just don’t think would have been possible if I was writing a thesis … even [in] two individual departments.

Q: I’m glad it worked out for you after all that. Okay, going back to the writing of the thesis, what has been the most exciting or interesting part of researching for your thesis so far?

A: So I started this process last spring. I did a special topics course on “Guns in American Politics.” I was following the syllabus of the official course that’s run in the same name, but supplementing it with additional readings related to my thesis … And then this summer, I was doing an internship where I spent some time thinking about gun violence, where I was also working with a lot of people who are experts in gun violence and had an opportunity to talk to a lot of them. And then this fall, I have really dived into the research process. I think the most interesting points so far have been conversations that I’ve had with random people, because one of the things that I love about this topic is it’s something that anyone can understand. Over the summer, I played the cello at a wedding in Boston, and was talking to a wedding guest afterwards, who was saying … he thinks it’s important to consider the impact of personalized characteristics in how that influences people’s thought process. And I was like, ‘That’s such a great point.’

Q: That sounds really interesting. My parallel question is, what has been the most challenging aspect of your research so far?

A: A lot of poli[tical] sci[ence] research is done by people who aren’t statisticians … So one of the challenges that I’ve come across is a lot of statistical methodology that isn’t particularly statistically rigorous. And if I’m not convinced by the methodology, then I question the findings, but I simply don’t have time to redo all of these analyses and perform methodology that I believe is more statistically valid. And so there’s been something of a balancing act … The other main [challenge] is in accessing data sets … There are some data sets that I’ve come across that are really interesting where the data isn’t publicly available, and so I have to go on my own journey to reach out to the authors and see if they’d be willing to let me access the data.

Q: Have the authors been willing?

A: Some. The biggest one that I’m really interested in is, so I mentioned earlier that California has a systemic record from since, I believe, the [19]90s, of every firearm transaction that’s occurred in the state, and they’re the only state that does that. And in addition [they have] the residential address … and name. And once you have that, you can go and map back two important things: voter registration data, which then includes basic demographic information about the individual, and CDC or California state issued mortality data. So these researchers went through a five year long process of mapping all of this data together and creating this massive data set … the closest thing to a truly individualized study of the impact of firearm ownership, and they find that firearm ownership dramatically increases risk of homicide and suicide. It’s a really interesting data set … that I’m hoping to be able to access. But I didn’t hear back from the first researcher, and I’m going to be reaching out to a second shortly.

Q: Do you see any of the research and work you’re doing now extending into your postgrad plans? I know it’s early in the year, so if you don’t have post grad plans yet, that’s fine.

A: For sure. This past summer, I was working at the UChicago Crime Lab and UChicago Education Lab … it’s a data science research nonprofit affiliated with UChicago doing gun violence, criminal justice reform, and educational inequality research. I really, really liked the work [and it was] … the type of place that I would really love to go back to after college. It’s possible that I might end up at grad school for a year, but … [regardless] I definitely want to jump into work. And I think that the most logical area for me to start is something gun violence research related.

Q: Best of luck with all of that! Now I have my fun question, which is, what is your favorite place on campus to work on your thesis?

A: I think right now, my answer is still my room. I’m in a thesis dorm, so I’ve got an extra long desk, and I got a standing desk converter so I can, just like, pull a little lever, and then I can do a standing desk or seating desk … And I brought a monitor from home for my room, so especially when you’re doing coding and reading papers and comparing numbers and whatever, it’s very helpful to have two screens. That’s probably my number one. Definitely. The Green Room of Arms is a common work spot for me too.

Q: Nice. And what’s your favorite thesis study snack?

A: I have always been a huge fan of Trader Joe’s dark chocolate covered peanut butter cups.

Q: Nice! Do you have any advice for students who are thinking about writing a thesis or starting out writing a thesis, especially anyone who’s thinking about doing an interdisciplinary major?

A: Yeah, I mean, to me, this thesis has been the most exciting and intellectually stimulating part of my academic journey at Amherst so far. I think if there’s something that you’re really passionate about, getting to spend hours and hours every week thinking about it, working on it, researching it, is incredibly rewarding … And so my biggest piece of advice is find something that you really care about … I think the other thing that’s been really helpful for me, I’ve heard a lot of thesis students talk about how hard it is to not have any deadlines and to just have this sort of open-ended space to work in. So what I did is … Block out my semester into two-week chunks and make myself goals for each of those two-week chunks on what I want to get done. And I think that’s just been really helpful in letting me structure my time effectively and keeping me from feeling totally overwhelmed by everything that has to get done.

Q: I think that’s very helpful advice. And then finally, what is the biggest thing you want readers to take away from your thesis? Or from interacting with your dashboard?

A: I have two major goals. The dream for a general public user who’s looking at my dashboard is for them to come away with their preconceived notions of what owning a firearm does for your safety challenged. I’m not really under the illusion that I’m going to be able to change minds with a dashboard like this, but I do hope that it makes people think [and] makes people question … their preconceived notions, and maybe it inspires them to dive deeper … For my written thesis, I do hope to be able to come away with findings that will impact policymakers and future researchers in the fields … that might actually give policymakers an idea of what the most important things to be focused on right now are.