Where the apple fell from the tree

We all know the phrase, "the apple doesn't fall far from the tree," but what would happen if we were daring enough to venture far from that tree? I believe that we will find we have more courage than we knew had, we will see ourselves in a new light, and we will have experiences we never dreamed about.

As a rural student, I remember being utterly terrified to attend a university 2.5 hours from my hometown. I occasionally still feel that same fear when I do things outside of my comfort zone. This week has reminded me of the importance of moving forward with courage and trying things that make you uneasy. This last week I took a deep dive into Python, created a preliminary dashboard with a teammate in Tableau for my practicum, and had a conversation with a peer that beforehand made my stomach twist in knots. In each instance, I gained valuable skills or knowledge and I built a connection with others because I asked for help.

In 2015, I interned with the North Carolina Department of Public Instruction through the Marian Drane Graham Scholarship. I had the opportunity to complete a capstone research project of my choice and I was nervous about completing a statewide project. As a student from rural NC, I was curious if college-going behaviors were different for rural versus urban students in NC. As a sophomore in college, I had never heard of ANOVA, Multiple Linear Regression, Python, or Tableau. When I first completed this project, I completed all of my analysis in Excel! Figure 1 shows a short snippet of my work when I was running everything in Excel.

From my experience, I expected to see acceptance and enrollment rates were different for rural and urban students, but this hypothesis was proven wrong. Ultimately, the most important conclusion was that enrollment is related to how far the student lives from the university. The closer a student lives, the more likely they are to attend that university. Figure 2 shows an example of enrollment at UNC System Universities compared to the number of hours the university was from the administrative building for the school district.

We can see that for Bertie County, the closer the student was to the university, the higher their enrollment percentage in that university. For each university, these plots looked fairly similar.

This project originally took me 6 weeks and I was able to redo it in a weekend with Python. Admittedly, I did not have to download data from the UNC dashboard or create my own data on distance and time from the universities. However, my first time around I had to select 8 school districts from North Carolina as a representative sample to complete my analysis in a timely manner. This weekend I was able to examine all 116 NC school districts and their UNC System college-going behavior.

I was able to confirm that acceptance and enrollment rates do not differ for rural and urban students. However, rural and urban students make up different proportions of universities. Figures 3 and 4 show the enrollment percentages at Appalachian State University and North Carolina State University by urbanicity.

These plots illuminate some of the differences in university attendance. ASU has a larger percentage of rural students enrolling compared to NCSU with 50% of students enrolling from a city. Generally, universities located in more rural locations in NC had larger proportions of rural students attending.

I also confirmed that distance from the home is associated with UNC System enrollment. Generally, the closer a student lived to the university, the more likely they were to enroll. I created bins based on quartiles of distances from each university to complete this analysis. Figures 5 and 6 show enrollment at Elizabeth City State University and North Carolina State University by distance from the university.

At ECSU, the students most likely to attend, lived the closest to that university. However, NCSU seemed to have a sweet spot. Many students that lived closest to NCSU attended, but even more students who lived between 123.5 and 177 miles away, or quartile 3, attended NCSU.

These findings are critically important because it means that students attend universities close to home. It means the apple literally doesn't fall from the tree. A student living in Manteo or Murphy is extremely far from universities like NCSU or UNC-CH compared to other counties like Johnston or Yadkin. This leaves these students with fewer options in the UNC System if they follow the pattern of most students.

I am left wondering, how can we better support students in a university that is far from home? How can we encourage students to attend a university a little farther from their hometown? How can we help students who travel farther for college be successful in building a support system so far from home? How can we encourage people to venture far from the tree and step outside of their comfort zone to see what they are truly capable of?

As I reflect on my analytical work in the last 5 years, and more specifically the last 5 months, I am truly amazed by what we can do with analytics and tools like Python, R, or SAS. We can illuminate discrepancies and show similarities. We can identify areas for growth and areas where we are doing well. More importantly, we can do all of this in a timely manner and build a more complete picture by using larger amounts of data.

I am truly excited by what I have learned at the Institute for Advanced Analytics and I cannot wait to use my knowledge to answer questions and provide a clearer picture of reality. It is humbling to look back at how little I knew when I used to introduce myself as, "Ashley, the girl from the middle of nowhere." It is even more encouraging to see what that girl from the middle of nowhere has accomplished and how much I have pushed myself outside of my comfort zone.

Each time I have stepped outside of my comfort zone I have discovered I have more courage than I knew had, I see myself in a new light, and I have experiences I never dreamed about. I encourage you to take a little time this week to venture far from your tree.

For more information about this project like p-values and fun graphs, check out my code at this link.