The final question at my Ph.D. defense was by prof. Richard Gill. He saw an easy improvement to the data visualization in the introduction of my Ph.D. thesis.
Not that I was too happy about it myself. The main point of the figures was that sampling distributions and p-values depend a lot on the sample size (the number of heart attacks n, in the example) – you can only visualize them in this way given a sample size or stopping rule. That was still clear from the figure, but it was very distracting and counterintuitive that the distributions in the figure shrank, for which I had to add a long caption below Figure 1 of my Ph.D. thesis. It was October 2021, and I had set the deadline for my Ph.D. thesis to November 1st.
I simply did not see the easy fix.
Thanks to Richard, my Ph.D. now has better figures. Also thanks to Nicos Starreveld, who wrote a very nice article about my Ph.D. thesis that reuses the improved figure.
In this blogpost I also post the improvements to Figure 2 and Figure 3 that follow.
Fri, 20 January