Bachelorthesis

Before reading the post I recommend at least opening by thesis so that you can look at my cool plots hehe: Link to Thesis

In my Bachelor thesis, I analyzed the positive framing in sociological publications using the theory of academic capitalism. The theory of academic capitalism would expect an increase in self-promotion. I looked at whether this is reflected in scientific abstracts through increasingly positive language like “favourable,” “promising,” or “unique.”

For this, 121,789 English-language abstracts of sociological articles from the Web of Science database (1993–2023) were analyzed with two dictionaries (to increase validity and to teach me how to use dictionaries scientifically 😁). I found no evidence for a substantial increase in positive framing.

My Highlights

For me, my Bachelor thesis was not only my goodbye to my social science bachelor but also to my two semesters of physics I studied. How? By using error propagation and monte carlo simulation I used techniques I learned in physics which were not covered in my social science studies. But more on that later.

• My first highlight is extended Coleman boat on page 10. I structured the theory part in a way that represents different parts of the diagram. Being able to highlight the main aim/concept of my thesis in the diagram was nice.

• Page 19 has a simple partial derivative and cites the book (Tipler) we used in physics :D!!

• My irrational ideological obsession for heatmaps started with the plot on page 22. So much information in one picture, and it is easy to read as well.

• … also: have you seen a more beautiful plot showing Bradford’s law than the one on page 23?

  • “If scientific journals are arranged in order of decreasing productivity of articles on a given subject, they may be divided into a nucleus of periodicals more particularly devoted to the subject and several groups or zones containing the same number of articles as the nucleus, when the numbers of periodicals in the nucleus and succeeding zones will be as $1:n:n^2$…” - Bradford, 1985, p. 178
    • My plot shows the exact relation perfectly.

• Pages 29 and 30 show the difference of the interval plots of the pages before. This is where error propagation allowed me to draw the error bars to investigate where the difference between the interval plots was systematic and not random.

Because of the large sample size, the only linear regression evaluation metric we learned during my sociology bachelor was useless ($R^2$ will always look super good for large sample sizes). To solve that, I used Monte Carlo simulations to calculate linear models on smaller subsamples to see how the p-value behaved. Since the linear regression was only significant for super large subsamples, I concluded that the effect was negligible. The effect which was tested was whether the relative amount of positive framing increased over time btw.

I used R for the whole thesis, except for the linear model and Monte Carlo simulations, which I ran in Python. For the huge plots on pages X and XI in the appendix, I used Julia (much cooler to see the plot in an instant when tweaking the design).

The Printing Incident

When I wanted to print my thesis, I needed to wait for three additional hours before I could retrieve it. The police had 12 cars and dogs standing in front of the store because someone was attacked with a knife and hid in the building where the print store was located. Luckily noone got hurt and I was still able to retrieve the thesis that day. This was especially important because I wanted to move out the following day to make it to Sweden by the end of the week for the start of my Master in Computational Social Science at Linköping University. I remember this moment as Gießens own way of telling me goodbye :D

Conclusion

… very fun! Thank you Prof. Elmar Schlüter for the fantastic help during my time at the Justus-Liebig-University Gießen and the supervision of my thesis. Also, a big thank you to Tim Schmidt, who pointed me in the direction of the Monte Carlo analysis and my dear social science friends Anna, Alex aswell as my friends from physics jonas and simon who were not only listening to my progress-reports over lunch/dinner but also being a source of inspiration to bounce ideas off.