Jeffrey Conroy-Krutz—Dissertation Project

Photo: Mural outside Electoral Commission, Kampala, Uganda

Dissertation Overview: “Political Information and Electoral Behavior in Sub-Saharan Africa”

My dissertation research focuses on how African voters access information about politics, and how they use different types of information in their electoral decision-making processes.

I am developing a framework that, drawing upon established theories of information collection and processing in U.S. electoral contexts, posits that voters in Africa—where rich information about politics tends to be costly and scarce—rely on a number of “shortcuts” as they gather information to make electoral decisions. These shortcuts can include candidates’ ethnic identities and cues that candidates’ widespread distribution of relatively low-valued goods (e.g., food, clothing, small amounts of cash), which I term "petty patronage," during campaign periods provide. These are all relatively cheap sources of information about candidates’ potential preferences and capabilities.

In the empirical sections of the dissertation, I examine how support for co-ethnics and distributors of low-value patronage (“petty patronage”) and levels of communal voting co-vary with their access to political information from mass media sources, such as television, newspapers, and, especially, FM radio.

In order to test the empirical implications of my theory, I am employing a number of research strategies, including a survey experiment (conducted in the Districts of Kampala and Kayunga, Uganda, in January 2008), and econometric analysis of survey and local-level electoral data from a number of African countries.

For more on the Ugandan survey experiments on political information and voting behavior, including data, a codebook, sampling procedures, and survey instruments (in English and Luganda), go here.

General comments welcome [e-mail: jkk2003@columbia.edu]

Working Papers, Selected Confererence Papers, and Associated Research Projects

"Information and Voting Behavior in Africa:  Results of a Survey Experiment in Uganda" [pdf]
                Revised working paper title:  "The Ethnic Shortcut:  In-Group Voting and Information Availability in African Politics"

Abstract: This essay argues that the marginal effect of ethnicity on vote choice in Africa is variable according to broader informational contexts. I present a theory positing that part of the utility of ethnicity for voters in Africa lies in the information that demographic cues provide regarding electoral competitors’ potential preferences and capabilities. In turn, I hypothesize that, in situations where broader types of political information are more available, reliance on the ethnic cue will decline, with the observable implication being decreased incidence of straight-ethnic voting. In order to test this hypothesis, survey experiments were conducted in two districts of Uganda with 370 participants, and subjects were asked to make selections in seven separate two-candidate races for local government positions, in environments where varying amounts and types of information about the hypothetical candidates was presented. I find evidence that, while support for co-ethnics is extremely high in settings where ethnic identity is the only distinguishing fact about candidates, this support declines markedly when certain kinds of richer information are presented. These results represent preliminary yet significant evidence that information environments can impact African citizens’ propensities to vote a straight ethnic line.

Version presented at American Political Science Association Annual Meeting (Boston, August 2008)

Comments welcome  [e-mail:  jkk2003@columbia.edu]

"The Informational Face of Patronage:  Reputation Building in African Electoral Contexts" [pdf]

Abstract: This paper offers a theoretical argument in favor of the conceptualization of the distribution of goods by electoral competitors in Africa as, in part, a strategy for communicating costly signals to voters. First, I highlight some of limitations that exist in the current literature on distributional linkage strategies in African electoral contexts, primarily in terms of ongoing difficulties in explaining the ubiquitous, campaign-time distribution of relatively low-value items, such as T-shirts, bags of food, bars of soap, and phone cards, as attempts at loyalty- or vote-buying. Next, I argue that African voters derive certain informational cues from witnessing the distribution—or non-distribution—of goods of varying volumes and values, which they use to make electoral decisions. Voters can generate assessments of a distributor’s generosity; commitment to redistribution, rather than personal hoarding; attitudes towards recipients’ ethnic group, village, or region; capacity; and electability through observations of such distributions. Following on this conceptualization, I hypothesize that in situations when other, possibly higher-quality sources of information about electoral competitors are available, voters will tend to weigh the cues they receive from observations of distribution less, vis-à-vis these alternatives. Empirically, therefore, I expect that we will observe that the marginal utility that electoral competitors derive from the distribution of one unit of patronage varies, across individuals, within and between countries, and over time, according to voters’ abilities to gather information from other sources, such as mass media and social networks. Finally, the paper presents a research strategy—involving survey experiments, focus groups, and local case studies in Uganda—for collecting data that will allow for testing of some of the observable implications of the model.

Presented in African Politics Conference Group panel at African Studies Association meeting (New York, October 2007)

Comments welcome  [e-mail:  jkk2003@columbia.edu]

“How Common is Communalism? Diversity, Information Access, and Community Political Cohesion in Uganda”

Abstract: Communal voting along ethnic lines is widespread throughout much of the developing world, and in Sub-Saharan Africa, in particular. However, despite the apparent ubiquity of such political behavior, there is substantial variation within countries in local levels of political cohesion, even within populations that are rather homogeneous ethnically. What factors, then, explain this variation in levels of communal voting at the local level in Sub-Saharan Africa? I argue that, within countries, certain structural realities, such as ethnic demographics, access to political information via mass media, land-tenure régimes, education levels, and standards of living, are likely to constrain local élites interested in building and maintaining high levels of political cohesion. This paper attempts to answer some of these questions, through examination of the case of the 2001 presidential election in Uganda.

Earlier draft presented at Midwest Political Science Association National Conference (Chicago, April 2007)

Undergoing revision.

“Whither the Angry Voter? Evaluation, Information, and Party Support in Sub-Saharan Africa”

Abstract: Since 2000, elections in a number of Sub-Saharan countries—including Bénin, Ghana, Kenya, Madagascar, Mali, Mauritius, Sénégal, and Sierra Leone—have resulted in an ouster of ruling parties. Analysts have often attributed these turnovers, and the significant levels of support for opposition movements in other countries such as Ethiopia and Zambia, to widespread dissatisfaction with incumbents’ performance, particularly in the economic arena. Yet while these developments have encouraged new interest in individuals’ electoral decision-making processes in Sub-Saharan Africa, our understanding of the extent to which voters there can be considered retrospective is still quite limited. Anecdotal literature often portrays some sets of voters—particularly rural-dwellers—as prone to rather blind support for co-ethnics or patrons, and others—such as urban youth—as perpetually disenchanted, and hence, opposition-supporting. In other words, some African voters seem to incorporate evaluations of past government performance more readily into their electoral decision-making calculi than others. In this paper, I seek to identify more formally the individual-level and environmental factors that are likely to affect individual Africans’ propensity towards retrospective voting. Particularly, I hypothesize that individuals who have regular access to formal mass media will be more likely to convert negative retrospective evaluations into opposition to the incumbent, due to the role that media play in conveying information about linkages between government policy and macroeconomic outcomes, and opposition alternatives. I test these expectations with data from the second round of the Afrobarometer survey, which was conducted between 2002 and 2004 in sixteen countries.

Earlier draft presented at Boston University African Studies Graduate Conference (Boston, March 2007)

Undergoing revision.

Return to Jeff Conroy-Krutz’s Home Page