This summer, I assisted Professor Elissa Berwick on a project about sub-state nationalism and independence referendums, examining in particular the cases of Quebec and Scotland. We aimed to use a regression discontinuity design (RDD) methodology to assess the impact of participating in a referendum on an individual鈥檚 long-term political identity, combining this with survey data and qualitative interviews. I was interested in this work because I have previous experience in quantitative analysis for the Social Sciences, primarily in Statistics, Econometrics, and R programming language, and wanted to apply these academic skills to a research project. My objectives were to gain hands-on experience in research with the supervision of a professor and on a specific project. I was also interested in the uniquely Canadian context of Quebecois nationalism in particular, having taken several Canadian politics classes, and wanted to dive into this topic more deeply. I was tasked primarily with conducting a review of the relevant literature, researching the methods to be used in the study, and creating and presenting descriptive statistical analyses using public opinion data.
Some highlights included visualizing and creating representations of data through graphs, plots, and charts. Though seemingly simple, there were many creative decisions behind the data visualization that would each contribute differently to the simplicity, clarity, and informativeness of the representation. Other highlights included the opportunity to deep dive into a particular research methodology, conducting a review of the literature and better understanding the challenges and opportunities of the RDD design. I did an analysis of statistical power in RDD designs and learned about the applications of them across different disciplines. In doing so, I studied not only the theoretical mathematical models behind power calculations, but also the practical challenges associated with implementing them with real data, which is often messy and imperfect.
Challenges primarily concerned data cleaning and merging. While often densely detailed, survey data often has inconsistencies due to updates and tweaks made by the researchers over time. In my case, over twenty years of Scottish public attitudes data was very thorough but had also undergone changes to coding practices and variable naming. While the motives for these are understandable 鈥� for example, the need to comply with EU data collection privacy laws 鈥� they made the task of combining and cleaning datasets more difficult. However, it presented an opportunity to learn and understand the ins and outs of the data collection process. I also deepened my familiarity with the programming language that many academics use in conducting research.
I would like to thank the donors of the Faculty of Arts Internship Award. This experience in research methodology, collection, and presentation has given me invaluable insight into research in academic fields, which I have learned is not static but ever-evolving. Methodologies, for example, are not set in stone procedures but rather tools that each present their own opportunities and obstacles. The financial support of the ARIA award allowed me to commit to this project over the summer and spend the necessary time to be thorough in my learning and research. I am better equipped for post-graduate work in Social Science research and hope to pursue more of it in the future.
