Fascinating paper and thank you for providing a talk! In thinking through your paper, there seems to several important caveats in the framing and measurement of political polarization.
1. Measure of Polarization
From what I gather in the paper, you define polarized edit groups as those being politically heterogeneous, using the variance of their alignment scores. Conversely, edit groups that are politically homogeneous are not polarized. As a matter of framing, this seems to cut against the more common colloquial approach of characterizing politically balanced groups as non-polarized and those consisting of predominantly or entirely one political group as polarized. A bipartisan coalition is the bridging group in a Congress with two polarized parties rather than the group showing polarization. As #62 points out, political diversity is not the same as political polarization and there are different levels at which we can evaluate the presence or absence of these qualities.
Methodologically and perhaps more obvious to this point, polarized distributions have a specific bimodal form with the majority of proponents clustered at the poles versus the center. Consider Figure 1B in your paper, which does not show evidence of polarization. This is apparent if compared to Figure 1 in Fiorina and Abrahms 2008, which displays a schematic of a polarized versus non-polarized distribution. Figure 1 B matches the non-polarized distribution versus the polarized one. Scholars like McCarty, Poole, and Rosenthal or Bonica have more about the exact methods of defining polarization. See also work by DiMaggio.
2. Nuances of Framing
In the political science literature (also discussed in Fiorina and Abrams), polarization typically refers to distributional assumptions of political ideology (Conservative-Liberal) not distributions of party identification (Republican-Democrat). A separate strain of literature discusses "party polarization" or "partisan polarization" to distinguish it from the former phenomenon. Further clarification in the framing and methods would be useful to better outline your presentation of polarization.
Fascinating paper and thank you for providing a talk! In thinking through your paper, there seems to several important caveats in the framing and measurement of political polarization.
1. Measure of Polarization
From what I gather in the paper, you define polarized edit groups as those being politically heterogeneous, using the variance of their alignment scores. Conversely, edit groups that are politically homogeneous are not polarized. As a matter of framing, this seems to cut against the more common colloquial approach of characterizing politically balanced groups as non-polarized and those consisting of predominantly or entirely one political group as polarized. A bipartisan coalition is the bridging group in a Congress with two polarized parties rather than the group showing polarization. As #62 points out, political diversity is not the same as political polarization and there are different levels at which we can evaluate the presence or absence of these qualities.
Methodologically and perhaps more obvious to this point, polarized distributions have a specific bimodal form with the majority of proponents clustered at the poles versus the center. Consider Figure 1B in your paper, which does not show evidence of polarization. This is apparent if compared to Figure 1 in Fiorina and Abrahms 2008, which displays a schematic of a polarized versus non-polarized distribution. Figure 1 B matches the non-polarized distribution versus the polarized one. Scholars like McCarty, Poole, and Rosenthal or Bonica have more about the exact methods of defining polarization. See also work by DiMaggio.
2. Nuances of Framing
In the political science literature (also discussed in Fiorina and Abrams), polarization typically refers to distributional assumptions of political ideology (Conservative-Liberal) not distributions of party identification (Republican-Democrat). A separate strain of literature discusses "party polarization" or "partisan polarization" to distinguish it from the former phenomenon. Further clarification in the framing and methods would be useful to better outline your presentation of polarization.