Blog #2: Voting Behavior Amongst Young Adults: An Analysis of Youth Nonvoters and how Behavioral Economic Concepts can be Applied to Increase Young Voter Turnout.

Since I am presenting my full research paper in December, I have had to act quickly and efficiently throughout the research process so far. Allocating time on my schedule was crucial to allowing my research to move forward. I attended meetings with my faculty mentor once a week to discuss what I have done and what needed to be done. I always thought of myself as an independent worker, but these weekly meetings acted as mini-deadlines, which kept me on track.

Writing my literature review was a challenge for me because I had to filter through past research and then critically analyze and write about it. I revised it a few times before settling on a final version of that specific part of my paper. Next, I looked for data and found it rather quickly because my topic of voter behavior is popular. I decided to use data from the U.S. Consensus Bureau because it had an abundance of quality data. However, the problem was that not all of the data was useful. In my paper, I specifically analyze voters and nonvoters from ages 18 to 24. This made consolidating data to fit my research almost impossible. Luckily, another professor caught wind of my dilemma and referred me to the data sets available from the Current Population Survey. This data allowed me to include the data I needed and exclude the data that I didn’t.

Since all of my data derived from this survey, another problem of using a lot of dummy variables arose. Using a lot of dummy variables could negatively affect my results. Luckily, I had a few continuous variables, such as age, income, and duration of residence, which offset the use of multiple dummy variables. Next, I had to regress my data using an application, called Stata, in order to determine what variables were significant. When analyzing my results I found that almost all of my variables were significant, but my data only explained 10% of the variation of my y variable. Although this isn’t necessarily a bad thing, I brainstormed new variables that could be included. When I regressed my data with three new variables, my data explained 25% of the variation of y. I was happy to see the increase. Initially, I was only focusing on data from 2016, but my faculty mentor suggested I do the same regression for prior years so that I can compare my results.

After I felt comfortable with my data, I wrote my first draft of my whole research paper. Now, I am awaiting feedback and revisions from my faculty mentor. Once I receive that, then I will have a lot of revisions to make and possibly brainstorm other ways to improve my study if time allows.


Leave a Reply

Your email address will not be published. Required fields are marked *