![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Much of the existing literature on the gender gap in support for war has focused on average differences between men and women. Little research has focused on variations in men’s and women’s support due to differing aspects of the situational context, such as regime change, economic interests, threat of terrorism, or humanitarian crisis. Previous research has suggested that gender differences are reduced when the objective of a military intervention is to promote human rights and stop a humanitarian crisis (Brooks and Valentino Citation2011; Eichenberg Citation2016, Citation2017), but the underlying reason for this finding remains elusive. In this article, I seek precisely such an explanation—values. Using original data from an experiment, I find that different values, such as militaristic and humanitarian values, mediate the gender gap, depending on the reason for intervention. These results add to the accumulation of knowledge on foreign policy attitudes, the gender gap, and values.
Notes
Acknowledgments
Many thanks to Leonie Huddy, Stanley Feldman, Howard Lavine, Kira Sanbonmatsu, and Richard C. Eichenberg, as well as to the reviewers and editors for their extremely valuable feedback.
Notes
1 Participants were not asked their country of origin or citizenship status. The university where the study was conducted has a varying percentage of international students of 18 percent or less of the undergraduates being considered international students.
2 This type of sample has been used before in public opinion research (Kam Citation2012; Malhotra, Margalit, and Mo Citation2013; Berinsky, Margolis, and Sances Citation2014; Lizotte Citation2016). Survey Sampling International randomly selects panel participants with survey invitations when deploying a survey resulting in a diverse national sample including hard-to-reach groups (e.g., ethnic minorities and seniors). In addition, internet-based panels, including Survey Sampling International samples, have been shown to be more representative than Mechanical-Turk samples (Berinsky, Huber, and Lenz Citation2012).
3 First, Gambia does not have oil or any exports of economic importance that could realistically threaten US economic interests.
4 Because the Department of Homeland Security might be associated with the presidential administration, I chose to change the source of the information to the nonpartisan group Council on Foreign Affairs.
5 Age bias is of particular concern with internet samples, and the mean does not give an accurate indication of whether or not the sample is skewed toward younger individuals. This is not of concern in this sample, because 85.03 percent of the respondents are under the age of 65, whereas 14.97 percent are 65 and older. This is quite similar to the NES 2008 data, which has 83.01 percent of the sample under 65, and 16.99 percent are 65 and older.
6 The 21-percentage point gap in response to the terrorism condition is much larger than the typical gender gap. This is likely because one disadvantage of experiments is that the use of fictional situations requires respondents to respond to a hypothetical question asking support for military action. This can lead to greater levels of support than is found in response to real conflicts.
7 Party identification and ideology are correlated at .60, which does not pose a multicollinearity concern.
8 This is an example of moderated mediation (for more information about moderated mediation, see James and Brett Citation1984 and Preacher, Rucker, and Hayes Citation2007). The effect of the mediator humanitarian values on the dependent variable support for war is moderated by the independent variable experimental condition. As a result, the conditional indirect effect of humanitarian values on support for war is dependent on the experimental condition. In other words, I only expect to find humanitarian values increase support for intervention in the humanitarian condition.
9 The Sobel-Goodman mediation test in STATA uses regression, which is inappropriate, given the dichotomous nature of my dependent variable. Therefore, the percentage mediated is not completely accurate.
10 To investigate moderation, I ran separate models for men and women in an analysis not shown. Humanitarian values were significant in the men-only and women-only models. This indicates that men and women did not differentially rely on humanitarian values when deciding whether or not to intervene in the humanitarian condition. Militant values were only significant in the men-only model, meaning that men relied on militant values whereas women did not when considering intervening in the terrorism condition in the student study but not the adult study. In the adult study analysis, there were no differences in the significance of party identification, age, or ideology between the men-only and women-only analyses.
11 Age is a significant predictor of opposition to war in the analysis of humanitarian and terrorism condition. This may be because older individuals are more likely to have negative memories of past military interventions, including the Iraq War and military maneuvers in Latin America during the Cold War.
Additional information
Notes on contributors
Mary-Kate Lizotte
Mary-Kate Lizotte is an Associate Professor of Political Science in the Department of Social Sciences at Augusta University. Her research focuses on gender differences in public opinion, voting, and party identification. Her work has been published in various journals and edited volumes. Her book, Gender Differences in Public Opinion: Values and Political Consequences, with Temple University Press was published in March 2020.