This article measures the degree of political alignment, p, as the probability of contributing to red articles, and measures political polarization as the degree of spread of editors across the liberal-conservative spectrum: they measure "polarization of any given group of editors by the variance of their alignment scores." However, in measuring polarization, variance might not be the optimal choice, and I think it would be better if kurtosis is also considered (DiMaggio et al 1996), to better capture the conceptual attributes of political polarization.
Patterns of disagreement and conflict depend on the patterns of relations of people. If people from various identities and ideologies participate, the variance of the wikipedia page is higher. But, this does not refer to a polarized state, where a field is comprised of two poles that are apart from each other; rather, it refers to political diversity, where people with diverse ideas participate. Some extreme ideas might exist, but they would be mitigated by diverse, moderate participants. (The positive relationship between political diversity and reasoning has been investigated by some scholars (Klar 2014; Muts 2002)).
To measure polarization, we can think of two levels.
First, in the micro-level (where the unit of analysis is each Wikipedia page), a page is polarized when the ideational distribution of participants is bimodal (DiMaggio et al 1996). In other words, when participants of both very high or very low alignment index (thus, very liberal and very conservative) participates, a page is polarized. This can be measured by kurtosis. When measured solely by variance, it’s hard to distinguish between a bimodal (thus polarized) group, and a diverse group.
Second, in the macro-level (where the unit of analysis is Wikipedia page itself), Wikipedia is polarized when each page has small variance and high average alignment (Baldassarri and Gelman 2008; Della Posta et al 2015). In other words, Wikipedia would be polarized when more pages become internally homogeneous and externally heterogeneous; when more pages are edited by strong partisans with the same side. In this case, interestingly, local polarization actually lowers general polarization (Axelrod 1997)
This article measures the degree of political alignment, p, as the probability of contributing to red articles, and measures political polarization as the degree of spread of editors across the liberal-conservative spectrum: they measure "polarization of any given group of editors by the variance of their alignment scores." However, in measuring polarization, variance might not be the optimal choice, and I think it would be better if kurtosis is also considered (DiMaggio et al 1996), to better capture the conceptual attributes of political polarization.
Patterns of disagreement and conflict depend on the patterns of relations of people. If people from various identities and ideologies participate, the variance of the wikipedia page is higher. But, this does not refer to a polarized state, where a field is comprised of two poles that are apart from each other; rather, it refers to political diversity, where people with diverse ideas participate. Some extreme ideas might exist, but they would be mitigated by diverse, moderate participants. (The positive relationship between political diversity and reasoning has been investigated by some scholars (Klar 2014; Muts 2002)).
To measure polarization, we can think of two levels.
First, in the micro-level (where the unit of analysis is each Wikipedia page), a page is polarized when the ideational distribution of participants is bimodal (DiMaggio et al 1996). In other words, when participants of both very high or very low alignment index (thus, very liberal and very conservative) participates, a page is polarized. This can be measured by kurtosis. When measured solely by variance, it’s hard to distinguish between a bimodal (thus polarized) group, and a diverse group.
Second, in the macro-level (where the unit of analysis is Wikipedia page itself), Wikipedia is polarized when each page has small variance and high average alignment (Baldassarri and Gelman 2008; Della Posta et al 2015). In other words, Wikipedia would be polarized when more pages become internally homogeneous and externally heterogeneous; when more pages are edited by strong partisans with the same side. In this case, interestingly, local polarization actually lowers general polarization (Axelrod 1997)