Being a Researcher

Background:

This post covers my experience as a fresh researcher during my undergraduate degree working with $2$ psychologists and a statistician at the university I was attending for my Finance degree (at the time at least). We were trying to develop a model for explaining and demonstrating something known as the SCHOOL-TO-PRISON PIPELINE. It’s basically what the title says. The main question under study was basically:

How and what factors cause youth and adolescents in America to be incarcerated as juveniles?

Super depressing in hindsight, but I wasn’t really laserd in on the overall outcome, as you’ll understand later on.

Structural Equation Modelling:

If you’re paying attention, you’ll notice I made a post around 2022 demonstrating what Structural Equation Modelling (SEM) could do. When I first saw and implemented a demonstration of what SEM could do, in like 2020 while wokring on a research project during my undergraduate degree, it amazed me. I didn’t really understand how you could just - excuse my language - put a bunch of disparate variables together and a hopefully statistically-significant result occurs. Literally blew my mind. While working on that project with my colleagues - college professors if you’re curious - it was mostly a labor of data cleaning and coming up with code to create the dataset needed to model. I think in that $1$ month I learned more about R and how to clean data properly than my entire time in undergrad. I had to work quite a bit just to understand what the source problem was and read a lot of literature around proper standards - if that’s even a thing.

To this day, I still try to think of ways at my job to utilize the learnings from my time as a researcher, but there is sadly yet to be anything of usefulness.

Learnings:

Being a researcher, especially in undergrad, is not as simple as I thought it would have been. I used to want to be a researcher (still kinda do) when I grew up. However, there’s a lot of hats that needed to be worn and deadlines that needed to be met, especially from the journals themselves, that can weigh on a person’s psyche. There are an obvious number of issues with American publishing practices, but that’s not really the subject for today’s review.

1. Always have a solid understanding of the problem, before you begin coding away.

I made the mistake of just throwing my limited experience at the problem of modelling the topics of interest and it didn’t land very well with the professors. There were a multitude of early mistakes that could have been easily avoided had I understood the assignment a little better.

2. Ask questions.

This one seems obvious to me now, but it certainly wasn’t to me as a $22$ year old. I believe, looking back, that all of my struggles could have been more easily solved had I just simply asked for help. For example, I couldn’t figure out how to code a particularly tricky feature that we were trying to use within our model, so I just spent like $2$ days frothing about, to no avail. Please don’t do that!