As we can see, the question quality for questions from different categories (from figure 1C) reveals a huge difference. That means some questions might be easier to answer so that the answer quality score will be more centralized. On the contrary, some questions are hard to answer, so you can answer it badly but could also answer it in a perfect way.
When analyzing the effects of political polarization on answer quality, I think you could run analysis by question types (that is, explore heterogeneity). Since a certain group of people will probably tend to answer a certain type of questions so there exists selection bias.
As we can see, the question quality for questions from different categories (from figure 1C) reveals a huge difference. That means some questions might be easier to answer so that the answer quality score will be more centralized. On the contrary, some questions are hard to answer, so you can answer it badly but could also answer it in a perfect way.
When analyzing the effects of political polarization on answer quality, I think you could run analysis by question types (that is, explore heterogeneity). Since a certain group of people will probably tend to answer a certain type of questions so there exists selection bias.
Thanks!