Embedded Files
Assisting McCombs professors with their research project
Project Overview
Project Overview
From the Spring of 2023 to now, I have worked with Professors Laura Starks and Janet Dukerich on their research project. This project involves comparing reputable journals within the fields of Finance and Management to look at how they have accepted submissions over the past three decades. The professors had concerns that the journals could potentially have biases within their selection process of certain articles. For example, the affiliations of authors, such as whether or not they went to a top research institution, could potentially impact the likelihood of an article being published, rather than only taking into consideration the quality of the article.
Research Steps
Research Steps
Data Gathering: To start the project, I gathered data from 6 reputable journals, 3 in the field of management and 3 in the field of Finance. This data was exported into Excel from Web of Science and I brought it into Python to create useable data frames. The data frame below is the starting point of all of the data that I was able to gather across the journals (showcasing the first 5 rows).
Data Exploration: I explored the data to look for any important statistics or trends that could be useful in a regression or time series model. I looked at the universities that had the most authors affiliated with them, the articles that had the highest number of cited references, the number of articles published every year, the number of articles published per journal, and various other statistics. One of the first regressions I created was to look at the influence of the article by looking at what variables impacted the number of times an article was cited. With cross-sectional data, the r-squared value tends to be lower due to the nature of the data, but other aspects of the regression are incredibly useful. The regression below shows that holding other variables constant, a one-unit increase in Affiliation Count is associated with an average increase of 290.72 in Cited Reference Count. This suggests that articles with more affiliations tend to receive more citations on average. Similarly, a one-year increase in Publication Year is associated with an average increase of 1.53 in Cited Reference Count. This suggests that newer articles tend to receive more citations on average.
Next Steps: After creating various regressions and exploring the data, it was important to compile a list of the top 100 research institutions globally to look at the impact of these institutions on the number of articles published that were affiliated with them. For this, I compared the top 100 research institutions that the University of Texas at Dallas created as well as U.S. News and I put this against the list of universities that I had in the original data and found that there wasn't much overlap, indicating that the top research universities did not have that large of a presence in my data set. Something else that the professors wanted to look into was the number of yearly submissions to each of the journals, that way we can look at rejection rates which would greatly aid our research process, however a lot of this information is private. Even after contacting the authors of all of the journals, only one of the six journals openly gave us data to work with. We are currently in this step of the data analysis process, and we continue looking at different angles to try to work towards our final result.
Page updated
Google Sites
Report abuse