Student Spotlight: Myles Ellis, University of Maryland, Baltimore County


Tell me a bit about who you’re working for, what work you’ve done for them so far, and how skills at REU might help confront questions that have come up over the course of your work so far? 

I am working for Professor James Heckman at the Center for the Economics of Human Development. As a research assistant, my role at the center is to find, compile and organize data using computer programming languages such as Latex, Stata, and Python. 

The variety of skills I was able to learn at REU had a tremendous positive impact on my ability to efficiently carry out my role on the research team that I work with. Specifically, I was able to attend John Eric Humphries’ seminar about basic research skills for Stata, where we learned about good practice and strong foundational coding approaches. These skills will enable us to effectively analyze important information within datasets to use in our daily research. I am currently using what I learned in Stata to help analyze and make code more efficient for cleaning various information material. 

What sort of state is data in when you receive it? Are you merging multiple existing data sets? What makes it challenging to parse for modeling and examination?

The information we are currently using is mostly longitudinal data acquired through census statistics and surveys. However, most real world data is not perfect. As a matter of fact, a major challenge in our case is parsing the survey data for modeling and examination, because our survey participants may not give accurate information 100 percent of the time. 

In one specific instance, we sought to determine and log the school, location, and other pertinent details through surveying disparate age groups ranging from adolescents to middle-aged adults. However, we suspected that a considerable amount of older, middle-aged adults could not correctly recall where they attended school in the early stages of their childhood. In turn, this prompted us to prioritize the task of validating the data in terms of what school they were most likely to attend, among other details.   

How does this compare to what you usually work on in class? What’s it like to be working with real-world data instead of problem sets? 

I would say that working with real world data definitely contrasts starkly to working with problem sets in the classroom for this reason. I think acquiring real world data poses challenging, yet solvable issues that requires a bit of creativity to resolve. On the other hand, problem sets in the classroom seem a lot more refined and perfected. It is pretty exciting having the chance to learn about how research is conducted in this sense, and to have a new perspective on what it takes to successfully answer the many questions that research entails.  

–Mark Riechers