Blog Post 6: Preliminary Analysis
Hello everyone, and welcome to my blog post! Last week, I finished cleaning my articles, preparing myself for the analysis portion. I decided to do basic keyword and phrase analysis to begin my analysis. Through this, I wanted to identify an association between the news source and the type of framing language used.
The first step in this process was developing lists of common left-leaning and right-leaning terms associated with each of my five topics. I then used Python to scan all 15 articles and count the number of times each phrase occurred. At first, my results were skewed; the majority of results showed up as 0, which showed me that some of the terms I used were too specific. After refining the keyword lists with more realistic and widely used terms, the analysis became much more balanced and meaningful.
With these updated counts, I added the total number of left- and right-leaning phrases used by each outlet across all articles. To determine the significance of this data, I used a Chi-square Test of Independence, with my null hypothesis being “There is no association between the news source and the type of framing language used. (They are independent.)” The results of this test include a chi-square value of 13.97 and a P-Value of 0.0009. Since the P-Value is far less than the standard threshold of 0.05, we can reject the null, and the results of the test provide convincing evidence that there is an association between the news source and the type of framing language used.
This statistical confirmation reinforces what many people intuitively sense — that different outlets frame issues in politically distinct ways. Next week, I’ll continue this analysis with a sentiment analysis to see whether tone and emotion also vary by source.