Content Analysis Protocol (Draft)

The protocol for the content analysis portion of my dissertation will remain similar to the originally proposed protocol: Statements by political elites during election campaigns will be analyzed, with careful attention to specific issue dimensions, broken down into multiple questions, each placed along a linear “either-or” five-point scale. If necessary, those can be compressed into a three-point scale.

The various dimensions are:

1. Economics: This includes attention to statements related to general attitudes towards neoliberal economic policies (on a scale from pro-neoliberal to anti-neoliberal). Additionally, I consider questions on more specific either-or questions related to: a) privatized v. statist economic models, b) support v. oppose direct foreign investment, c) export v. industrialization.


2. Drug War: This involves two very specific questions: One, whether elites speak strongly in favor or strongly opposed to coca eradication policies. Similarly, whether elites speak very favorably or very negatively about cocaleros. I include this new dimension primarily because the drug question is an important one in Bolivian politics, especially with the rise of a powerful, organized cocalero political party (MAS).


3. The “National Question”: This dimension evolved from the previous research model’s emphasis on the 1952 Revolution effect on political culture. I will still consider elite statements on a five-point scale related to whether they speak favorably or negatively about the general scope of the 1952 Revolution. Additionally, I pursue a few specific questions about different answers to what I describe as “the national question” in Bolivian politics. These include: a) whether elites use the campesino or indigeno to describe rural peasant farmers, b) degree of support for agrarian land reform, c) use of “pluricultural” rhetoric (essentially, arguing on behalf of ethnic differences v. national similarities), d) appeals to regional v. national politics and interests, e) degree of anti-Chilean (or anti-foreign) rhetoric.


4. Inter-Elite Attitudes: The bulk of this section remains similar to the previous research design in my proposal. This dimension takes each elite, identified as a member of a specific political party, and considers his or her statements related to other elites. These include: a) attitudes toward other parties, b) attitudes towards the incumbent government coalition, c) attitudes towards the current president’s party, d) attitudes towards the party system in general, e) attitudes towards the electoral system. The bulk of the content analysis falls on the analysis of inter-elite attitudes, since it involves a larger number of paired possible scales (e.g. MNR-ADN, MNR-MIR, MNR-UCS, etc.).

The basic process for statement coding involves identifying a statement, identifying the applicable dimensions and questions, and coding the statements. The unit for analysis is identified as any individual, identifiable statement made by a single elite in a given space and time. In other words, if the UCS presidential candidate gives a campaign speech in Cochabamba, all reports of that one speech are jointly treated as one event. While this reduces the number of observed units, it eliminates possible selection bias if some events/statements receive greater coverage than others. I will also, of course, keep track of the number of reported instances of each statement, and may return to use that data.

Obviously, some statements may only involve one dimension, or only a few possible scale issues. I will not be measuring anything other than clear statements along at least one of the existing scales (reserving the possibility, of course, to include more questions into the analysis if some issues are frequently repeated in elite political discourse). Dimensions or questions not addressed in any single elite statement will remain unscored and blank.