Chapter 8 focuses in on showing how the political geography of electoral autocracies shapes the partisan homogeneity of social networks. The chapter begins by outlining a process for measuring political geography. It then uses the measure to show that political geography is an excellent predictor of the partisan homogeneity of partisan networks, even controlling for other key factors. Third, it provides evidence that the relationship between geography and networks is not simply an average treatment effect. Finally, it uses the survey data to illustrate what happens to people who move between stronghold types: The social networks of people who move to new strongholds appear to change partisan identities to match their new location.

Chapter 9 uses both original survey data from Cameroon and cross-national data from the Afrobarometer to provide evidence for the argument that political geography affects nonpartisan and cross-partisan political beliefs. It first demonstrates that people in different party strongholds describe themselves using categorically different kinds of adjectives, reflecting localized understandings of citizenship shaped by political geography. It then turns to the importance of understanding the effect of political geography on public opinion more broadly: Using Afrobarometer data from five different electoral autocracies, it reveals not only that public opinion is systematically different between party strongholds, even controlling for partisanship, but that even the beliefs of ruling party partisans change depending upon where they live. Finally, using Afrobarometer data from Uganda and Ghana, the chapter shows, first, that the development of party strongholds is not endogenous to preexisting political beliefs, and, second, that these patterns are, indeed, unique to electoral autocracies and do not hold in a democratic context.