This chapter establishes the empirical and historical context in which protest data were collected and describes the methods used to collect it. It also provides a first-order analysis of aggregate patterns of protest. Preliminary findings suggest that organization type, party alliance, and electoral cycles all matter for protest strategies.

MUNICIPAL GOVERNMENTS AND PROTEST

In order to test the theoretical hypotheses proposed in Chapter 1, the data must meet several criteria: (1) cover an adequate number and variety of organizations (to measure the impact of organizational type); (2) include a variety of organizational alliances, as well as organizations not allied to political parties; (3) cover periods of at least one alternation in power (to measure the impact of party allies in and out of power); and (4) cover these organizations across a sufficient span of time to determine whether behavior changes according to cyclical patterns (to measure whether electoral, budgetary, or administrative cycles matter).

The choice to focus on municipal governments addresses these criteria admirably. The particular municipal governments I examine administered large urban areas. I do not mean to suggest that rural protest is unimportant; indeed, movements like the Landless Movement (MST) in Brazil and the zapatistas in Mexico have had a significant impact on their respective nations' politics. Nevertheless, large urban areas tend to produce a wider variety of organization types and party alliance choices, permitting maximum variation on two key independent variables. These areas also offer the pragmatic advantage of multiple local media covering protest.