As juniors, students complete their Junior Paper (JP) over the course of the academic year, meeting regularly with their Junior Independent Work group and faculty advisor to discuss research interests and learn about common tools that are essential to research in economics.
While the prospect of a major independent research project can be daunting to many students, the department is committed to providing students with a wealth of resources to inform and improve their projects. From their faculty advisor to the research librarians, students will find there’s lots of help along the way.
Below, we talk to a few juniors about their 2020 JPs and what advice they have for those just beginning the process. And if you’re a rising junior with more questions about your JP, don’t forget to read the Junior Independent Work Handbook.
In her Junior Paper, Sara Hailu looks at gaps in access to services for veterans across U.S. states and finds that the presence of a County Veterans Service Officer (CVSO) and having more than one CVSO increased county VA expenditures per veteran, albeit by a small fraction of the geographic spending disparity.
What motivated you to study this issue?
At the time that I was exploring potential topics for my independent work, the Department of Veterans Affairs was in the wake of another scandal surrounding its hospitals and spending. The widespread issues of veteran homelessness, unmet mental and physical health needs, mismanagement of funds, and insufficient educational and vocational civilian rehabilitation motivated me to read countless articles, when I stumbled across the finding that VA per-veteran expenditures in counties across the country had wide disparities which the VA claimed could be mitigated by counties opening more CVSOs to increase outreach and assist veterans to access their full benefits and services. Despite this, limited research has been conducted regarding VA county expenditures, and even less has been done on CVSOs, so I was eager to investigate further.
What was the biggest challenge you faced when writing your Junior Paper?
The biggest challenge I faced was in my attempt to extend my main models to a case study of Texas counties to further test if having a CVSO increases county expenditures per veteran by comparing the 2018 per-veteran expenditures in counties that opened a CVSO after 1996 (the year the VA county expenditure dataset began) to the 2018 per-veteran expenditures of counties that never opened a CVSO. My goal for this difference-in-differences analysis was to determine if counties that chose to open CVSOs were otherwise different from non-CVSO counties and if those other omitted variables accounted for expenditure differences in my main models. I contacted several Texan veteran and archival organizations, but none had records of when CVSOs were opened, so I then contacted individual CVSOs by telephone to compile their opening years. However, I learned that most of the offices opened prior to 1996 and therefore were not suitable to the outcome variable’s dataset. Although this challenge was a little time-consuming and presented many obstacles along the way, it gave me a great opportunity to develop and implement a research strategy, explore multiple avenues of problem-solving, and consider the hypothetical implications had I been able to complete the analysis.
In his Junior Paper, Devin Sun explore the effects of the ride-sharing economy on wages for low-skilled occupations in the U.S. and finds that Uber’s entrance into a city is positively correlated with and reasonably leads to an increase in wages of workers in low-skilled occupations, relative to those of workers in high-skilled occupations.
What was the biggest challenge you faced when writing your Junior Paper?
The largest issue that I encountered during my junior independent work was ensuring that my data and subsequent analysis upheld a rigorous test against internal validity. First and foremost, I needed to ensure that my dataset was comprehensive and provided a representative sample for the population that I wished to study. Bobray Bordelon in the Princeton University Library, my advisor, Adrien Matray, and my graduate assistant, Simon Schmickler, were instrumental in this process of finding, vetting, and then cleaning my data sources. Using a combination of American Community Survey (ACS) and Bureau of Labor Statistics (BLS) data, I was able to accomplish that goal. Second, I aimed to use a method of empirical analysis that would accurately test my hypothesis. Simon and Professor Matray, along with Bruno Baranek, an economics data consultant, were great resources in helping me to refine a difference-in-differences model with various fixed effects considerations to better isolate for my desired explanatory variables. With a robust empirical model, I could continue with my research reassured that my results were understandable and accurate.
What advice do you have for rising juniors tackling their first independent research project?
Aside from topic ideation, the biggest lesson that I took away from this entire process was that Princeton and the Department of Economics have a bounty of available resources and that it does one well to take advantage of it all. I often prefer doing work by myself, but the University does a great job at offering assistance at nearly every step along the way. This means that you should reach out to the research librarians in the library because they have a wealth of information that they would love to share. Keep your advisor and graduate assistant in the loop at every stage because they genuinely want to see you succeed and produce good work. And frontload your work in the fall semester before winter break; when you return in the spring, you want to have set yourself up for a manageable work schedule and aren’t subject to unexpected last-minute scrambles.
In his Junior Paper, Jack Aiello investigates whether the International Monetary Fund’s (IMF) effect on income inequality in developing countries has changed in recent years. He finds significant redistributive effects of IMF programs. On average, IMF programs help redistribute income one year after program years across the sample period.
What motivated you to study this issue?
I’ve had an interest in the IMF and economic development since high school and I hoped my JP could help me learn about both of these fields. As I dug more into the literature, I realized the intersection between IMF programs and income inequality was exciting, important, and not exhaustively researched. As I read more, pieces of an interesting research question and a feasible methodology began to fit together in a coherent and compelling way.
What advice do you have for rising juniors tackling their first independent research project?
Pick a few possible topics you are interested in and read as much as you can find as early as possible. In my experience, the most important step in the research process is reading. Professor Zaidi also gave us very good advice when he encouraged us to stay ahead of the pack or ahead of the curve when it comes to making progress through the research process. Try to set deadlines for yourself a week or two ahead of the official deadlines so that when parts of the process inevitably take longer than you expected, you still have plenty of time to do good work.
Finally, do not get discouraged or intimated by math! Econometrics did not come easily or naturally this year, but you can learn so much if you stay calm, read and re-read methodology sections, and ask for help from your advisor, professors, teaching assistants, preceptors, and peers. Professor Farber’s ECO 313 class in the Spring was also a tremendously positive and helpful experience which I would recommend to everyone.
In her Junior Paper, Alexandra Rice uses data from a 15-year longitudinal Fragile Families and Child Wellbeing Study to examine how prior or current homelessness affects mental health. She finds that experience of homelessness in early childhood, specifically, has a significant effect on standardized depression scores at age 15.
What motivated you to study this issue?
I was always very interested in mental health, given the role it has played in many of my friends’ and family’s life. I thus found the subject to be both fascinating and relevant. I also was incredibly interested in the impact of early childhood experiences on later-in-life outcomes, especially after reading a lot of Professor Janet Currie’s work in my Public Economics class, taught by Professor Bogan. After going through the literature and realizing the correlation between homelessness and depression, I knew that this was a potential area of research for me. And, after realizing that there hasn’t been a lot of studies done on depression in one’s teenage years and prior homelessness, specifically, I knew that I could make a contribution to the literature with my research on this topic.
What was the biggest challenge you faced when writing your Junior Paper?
I think the biggest issues I encountered were (I) trying to figure out which variables to use as controls, and (I) attempting to get at a causal answer to my research question without having mental health data over a period of multiple years. Since I only had one observation on depression for each individual, I was limited in the types of analyses I could do. For instance, I couldn’t do a fixed effects model to remove time-invariant heterogeneity. To solve both of these issues, I really had to understand the narrative of my paper—what was it I wanted to show?
With regard to issue (I):
In my preliminary analyses, I found that early-childhood experiences of homelessness had a statistically significant effect on depression at age 15, while later-in-life experiences of homelessness did not. After talking about it with my advisor, Professor Noonan, I realized that given these findings—and my hypothesis that earlier-in-life experiences have a bigger impact than do later-life experiences on mental health outcomes—it made the most sense to control for characteristics from the baseline survey, when the children were just born, rather than controlling for characteristics of the teen at age 15.
With regard to issue (I):
I really just had to work within the limitations of the data. To make my results more credible, I used different techniques- linear regressions, probit, ordered probit- and looked at whether they gave similar results. I also noted and explained the limitations of my dataset, and gave ideas for further research on this topic.