close× Call Us
close×
Title Fear and Politics in Divided Societies: Assessing the Foundations of Political Behavior in Lebanon.
Post date 09/07/2017
C1 Background and Explanation of Rationale

Much contemporary research on political behavior in the Middle East and other developing regions emphasizes the role of clientelism in structuring elections and politics. Yet there is good reason to think that other factors beyond clientelist transactions shape political behavior in important ways. In particular, we focus on the political effects of fear in the context of instability, violence and the erosion of political order - a set of conditions that is unfortunately increasingly common throughout the Middle East and North Africa and in other developing regions. In this context, millions of Iraqis, Syrians, Libyans, Yemenis and Lebanese, among others, live in a state of fear and uncertainty. What impact does this state of affairs have on political attitudes and behavior?

At the same time, new political movements and initiatives indicate that citizens across the region are craving an alternative to clientelist politics. In this research, we aim to move beyond the current prevailing conceptualization of politics in developing countries as short-term clientelist transactions to consider a range of alternative or additional material and nonmaterial factors that may affect political behavior. The relative importance of fear and promises of protection against threats by violent extremists is then compared to a range of other possible motivations driving support for politicians, including the receipt of distinct levels and types of clientelist handouts and programmatic policy preferences, among other factors.

C2 What are the hypotheses to be tested?

H1a: Low-value, short-term clientelist distribution is more likely to affect electoral behavior than willingness to take part in demonstrations, while high-value, more continuous clientelist distribution should affect both outcomes.
H1b: In general, clientelist distribution is more likely to sway citizens with low socioeconomic status, but high-value clientelist distribution has broader effects.
H1c: Material benefits are less likely to influence core co-partisans but are more likely to move marginal activists and even non-partisans.
H1d: Core co-partisans are more likely to be moved by high-value material benefits.
H2a: Respondents are more likely to vote for candidates with a detailed plan to address salient policy issues (such as job creation or waste management).
H2b: High SES respondents, who are less likely to be swayed by and integrated in clientelist networks, are more likely to vote for candidates with a detailed plan.
H3a: Respondents favor candidates promising in-group protection over clientelist distribution, especially of low value goods and services.
H3b: Respondents favor candidates promising in-group protection over programmatic issues.
H3c: Respondents who express less tolerance for out-group members favor candidates promising in-group protection over programmatic and clientelist distribution, especially of low value goods and services.
H4a: Treated respondents, or those who have been primed to feel fear, express lower support for candidates who promise clientelist distribution, especially of low value goods and services.
H4b: Treated respondents express more support for candidates who pledge to protect co-religionists.
H4c: Treated respondents express no or less support for candidates who pledge protection for out-group members, especially when the out-group has antagonistic relations with the in-group.
H4d: Treated respondents are more likely to vote for and take part in demonstrations for co-religionist candidates.
H4e: Treated respondents are less likely to support female candidates.
H5a: Respondents are more likely to vote for and take part in demonstrations for co-religionist candidates.
H5b: People who identify more with their religious community are more likely to vote for and take part in demonstrations for co-religionist candidates.
H6a: Respondents who report higher levels of piety and religious observance are more likely to support pious candidates.
H7a: Co-partisanship has a bigger effect on political support than co-religionism.
H7b: Co-partisanship has a higher effect among core co-partisans than more marginal supporters or non-supporters on political support.
H8a: Respondents will prefer candidates with higher education.

C3 How will these hypotheses be tested? *

This research employs a conjoint experimental design (Hainmueller & Hopkins, 2014), in which voters are asked to choose between two hypothetical candidates. Certain aspects of each candidate's background are randomized to identify the causal effect of each attribute on respondents' willingness to vote for the candidate and to attend a rally sponsored by the candidate. In addition, for those respondents who are assigned to a version of the survey which combines a priming experiment with the conjoint experiment, we will randomly prime half of the respondents in the sample to think about what makes them fearful about their current political situation in order to test how living in a volatile political environment shapes preferences for political candidates. After each task (pair) in the conjoint experiment, we will ask respondents [in a randomized order] 1) which candidate they would vote for 2) how likely they would be to attend a rally or demonstration organized by each candidate (on a scale 1-7).

We will estimate the AMCEs by running regressions of the outcome variables (vote choice and willingness to attend candidate-organized rallies) on sets of indicator variables measuring the levels of each attribute. We will also compare the effects of candidate attributes among individuals in the fear prime control and treatment groups to explore how the estimated AMCEs differ in distinct political contexts (i.e., safe versus volatile). Specifically, we will show the results of our tests when divided in separated groups by priming treatment, as well as the differences in AMCEs for each attribute level between the two experimental (activity control/fear prime) conditions. Finally, to test our hypotheses about in-group identities, we will also run analyses including a dummy coding whether the respondent belongs to the same religious sect as the candidate. We will also estimate the AMCEs by running regressions on the theoretically relevant subsets of respondents as suggested by our hypotheses. Specifically, we will estimate AMCEs for high SES and low SES individuals; religious and nonreligious respondents; respondents tolerant and intolerant of out-groups; and non-partisans, marginal activists, and core supporters.

C4 Country Lebanon
C5 Scale (# of Units) 2400
C6 Was a power analysis conducted prior to data collection? No
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number RB17-0058
C9 Date of IRB Approval 5/26/2017
C10 Will the intervention be implemented by the researcher or a third party? Statistics Lebanon
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? not provided by authors
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? not provided by authors
C13 JEL Classification(s) D72