Update 1 on Political Discourse Analysis

It’s pretty late in the summer, I know, but great progress has been made. All of the data has been collected, I’ve analyzed it for power and solidarity, and what remains is to do the statistics and type up the report.

I am analyzing the speech of Democratic and Republican senators from 2009 and 2017, on the subject of American healthcare. In 2009, the Democrats were in power. In 2017, the Republicans were in power. I analyzed the data for instances of power and solidarity, to see if there were differences between the parties or the speech depending on what their power status was.

The bulk of the work was in simply transcribing the data. Writing down what the speakers were saying took considerable time and effort, especially since I gave myself more work in the process! While I was transcribing, I chose to do it in a CA style based on Atkinson and Heritage 1986. While I think I produced a good transcription that described the dialogue well, it simply did not fit my research purposes. All I needed were the words – in standardized English. However, at the time of transcription, I was not completely certain of what I was looking for, and also transcribed for extraverbal data. This ended up just hindering my analysis, and it took more work to edit the transcription to make it standardized English. Next time, I will think more carefully about what I am analyzing for before simply trying to do it in the “right” way.

While I have transcribed over 20,000 words, I have realized that with the work I am doing, that is still a very small number. In my statistical analysis, I’m finding that some instances of power and solidarity are showing up only 2 or 3 times within the entire set of data – not terribly helpful for statistical analysis. Linguistics is an interesting field – it’s not quite a science, nor an art. The data I have collected is subjective, and a lot of it is based on my opinion, yet I am still trying to make substantial claims.

The next step was coding. How can I define “power” and “solidarity” beyond the gut feeling I get? Some of it I did just code based on “feeling” – but I also divided up power and solidarity into categories – citation of numerical evidence and of authority for power, and use of humor, second person, reference to a senator’s own state, uncertainty, and a few other categories for solidarity. I am still combing through to see if I can add more categories to sort the data more effectively.

What I have found so far is that Democrats have not changed too much in their displays of power and solidarity between 2009 and 2017. In Republicans, the change was much more noticeable. The data I collected from the 2017 Republicans was totally different from the Republicans of 2009, and from the Democrats of 2009 or 2017. They use a large amount of rhetorical devices, evidence-based claims, and emotional appeals. They also used much more casual language – second person, idiomatic speech, and so on. Overall, there have been some noticeable differences between Democrats and Republicans – Republicans seem to utilize more solidarity – centered language, while the Democrats appeal more to power. Both have to maintain the balance, however, to remain the party in power.

I believe the bulk of my conclusion will be based on the numerical data I have collected, but then I will dive deeper into specifics of what exactly those numbers mean. Breaking down complex rhetorical devices and structures into numerical data has been difficult, and I’m sure I could have done it better. Before I do my analysis, I will do another run through of the data and code again, based on what I have learned as I have coded.


I am in for a few long days.


Speak Your Mind