Abstract
Large-scale military aggression is argued to damage the international image of the aggressor and mobilize global public opinion against it. Previous cross-country research also finds that negative views of the aggressor are usually limited to the government and do not extend to the citizens of the invading country. Our article provides micro-level evidence on attitude change toward Russia as a country, the Russian people, and the Russian government after its invasion of Ukraine. We use data from a survey conducted between the morning of 21 February 2022 (3 days before the Russian invasion of Ukraine) and the night of 28 February 2022 (5 days after the invasion) in the United States to evaluate how the Russian invasion of Ukraine affected attitudes toward the country, its people, and the government. We also conduct a subgroup analysis to explore the magnitude of attitude change across sociodemographic and political subgroups after the invasion. Our findings show fairly significant damage to the image of Russia as a country as well as the Russian government. However, the reputational damage of the Russian people is minimal. The results also suggest that Republican and religious subgroups had the largest attitude change on Russia and the Russian government.
Introduction
Large-scale military aggression is believed to damage the global image of the aggressor. Especially, military invasions without a UN mandate often mobilize global public opinion against the invading country. The most recent well-known example is the United States invasion of Iraq in 2003 which sparked worldwide protests against the war and fueled anti-American sentiments across the world (Baxter and Akbarzadeh, 2012; Chiozza, 2009; Ruzza and Bozzini, 2006). Previous cross-country studies also find that critical views of the United States after its invasion of Iraq were mostly confined to the government. While positive views of the American people declined after the invasion, the magnitude of the decline was not substantial (Holsti, 2009; Katzenstein and Keohane, 2007). Past literature has mainly relied on cross-national and cross-regional data (Holsti, 2009; Katzenstein and Keohane, 2007) to explain attitude change as a result of military aggression. Previous micro-level research (Chiozza, 2009) also provides evidence on sociodemographic profiles of individuals who are more likely to hold anti-American attitudes. Nevertheless, we still have little micro-level empirical assessment on how military aggression affects foreign publics’ opinions and whose attitudes change more radically after aggression.
Our study contributes to the existing research on war and public opinion by providing micro-level evidence on attitude change toward Russia after its invasion of Ukraine. We use data from a survey conducted between the morning of 21 February 2022 (3 days before the Russian invasion of Ukraine) and the night of 28 February 2022 (5 days after the invasion) at Arizona State University, one of the largest public universities in the United States, to evaluate the magnitude and scope of reputational damage to Russia as a country, the Russian people, and the Russian government. We regard the Russian invasion of Ukraine—which took place amidst our survey—as an exogenous impact that created a randomized environment for our sample to explore attitudes on Russia. We conduct balance checks and a logistic regression analysis with relevant control variables in order to mitigate the potential confounding bias in our analyses.
Our study makes two major contributions to the existing literature. First, we use individual-level data right before and after the Russian invasion of Ukraine. This provides us an opportunity to estimate the causal effect of the conflict on public attitudes. Second, we conduct a subgroup analysis to explore the magnitude of attitude change across demographic, socio-economic, and political subgroups in order to understand whose attitudes shifted more radically. By classifying individuals into groups on the basis of shared sociodemographic and political features, the subgroup analysis provides a more nuanced and detailed understanding of attitude change as a result of military aggression.
Our findings suggest that the Russian attack on Ukraine has imposed significant reputational costs on the country and the government. While 23% of the survey respondents expressed very unfavorable opinions of Russia in the pre-invasion sample, 39% of them expressed that they hold very unfavorable views of the country in the post-invasion sample. The reputational cost is even higher for the government: while 56% of the respondents in the pre-invasion sample held very unfavorable opinions of the Russian government, the post-invasion sample shows that 81% of the respondents expressed very unfavorable views of the government. Nevertheless, our results demonstrate that the negative attitudes toward the aggressor government do not necessarily translate into negative attitudes toward their population. The findings suggest that the respondents did not change their attitude toward the Russian people after the invasion. Our results also suggest that Republican and religious subgroups underwent the largest attitude change toward the country of Russia and the Russian government.
Besides the empirical results, our findings have implications for understanding interactions between domestic politics and international relations, especially in democratic countries where public opinion plays a crucial role in shaping foreign policy preferences (Page and Shapiro, 1983; Tomz et al., 2020; Wlezien, 1996). This is even more important in the age of social media in which active citizens have increasingly become powerful in mobilizing public opinion. Strong pressures from public opinion could not only influence political leaders, but they could also exert influence on businesses that constantly need to protect their reputation. Our findings imply that as international public opinion grows unfavorable, aggressor states could potentially face unexpected pressure from governments as well as international corporations. In the case of the Russian invasion of Ukraine, the war has not only resulted in unprecedented economic sanctions against Russia by the European Union, the United States, and their allies but has also led to unilateral moves by foreign companies to stop doing business with Russia. While some of them might be motivated by the politically risky environment of Russia, some corporations pulled out of the Russian market to simply avoid reputational harm (Siemaszko, 2022). In other words, the reputational cost of military aggression is likely to translate into the economic and political cost for the aggressor as a consequence of global public opinion mobilization against the war.
Research design
Our study uses data from a survey conducted at Arizona State University between the morning of 21 February 2022, and the night of 28 February 2022. During this period, a total of 502 respondents participated in the research. The survey was meant to collect attitudes toward a country other than Russia but also included questions on other countries, including Russia, in order to conduct a comparative analysis. 1 However, given the beginning of the Russia–Ukraine conflict on February 24, we focused on Russia to understand how the invasion affected the public attitude toward the aggressor. Our survey included three questions on the degree of participants’ attitudes toward Russia. More specifically, we used the exact wording of Pew Global Attitudes survey questions to measure respondents’ attitudes toward Russia as a country, the Russian people, and the Russian government. The time during which the survey was implemented provides a unique opportunity to evaluate the impact of the Russian full-scale military invasion of Ukraine on participants’ attitudes toward Russia. The invasion started on 24 February, 6 a.m. Moscow time (23 February, 8 p.m. Mountain Standard Time), almost 3 days after the beginning of the survey. Of the 502 respondents, 275 (54.7%) took the survey before the invasion and 227 (45.3%) completed it after the invasion. The time difference between pre-invasion (time1) and post-invasion (time2) provides an opportunity for an analysis of the extent to which the Russian military operations in Ukraine impaired the image of the country, the Russian people, and the Russian government.
It is important to note that the possibility of self-selection bias after the invasion is unlikely in our survey because the respondents had no idea about the topic of the study. The invitation email received by the students never mentioned any words on the study being related to Russia (or any other countries). The students received an invitation email about participation in a research study without being told anything about the topic. They only knew that they would participate in a study but they had no prior knowledge about the topic. Therefore, it is unlikely that the invasion motivated the students to participate in the study.
Our study includes three dependent variables: attitudes toward the country of Russia, attitudes toward the Russian people, and attitudes toward the Russian government. We ask the respondents whether they have a very favorable, somewhat favorable, somewhat unfavorable, or very unfavorable opinion of Russia, the Russian people, and the Russian government in three questions. We employ a 4-point forced scale to measure individuals’ attitudes (Supplementary Table A1 in the Supplementary Appendix displays the three questions). We focus on the “very unfavorable” responses as they indicate extreme negative attitudes to explore to what extent the invasion changed individuals’ attitudes. The variable is coded 1 if the respondents express a very unfavorable opinion of Russia, the Russian people, and the Russian government, 0 otherwise. We also provide an alternative analysis in the appendix by coding “somewhat unfavorable” and ‘very unfavorable” responses as 1, 0 otherwise.
The running variable is time and is coded based on whether an individual completed the survey after the Russian invasion. We coded all the individuals who took the survey after 10 a.m. MST, 24 February 2022, as 1. Individuals who responded to the survey before that were coded as 0. We chose February 10 a.m. (14 h after the beginning of the conflict) as our treatment cutoff because it usually takes a few hours for the news to spread, especially if the event happens at night.
We include several control variables to account for potential confounding bias. First, we control for party identification. The variable Republican is coded 1 if the respondent identifies themselves as Republican, 0 otherwise. Likewise, the variable Democrat is coded 1 if the respondent identifies themselves as Democrat, 0 otherwise. Second, we control for political ideology. Variable Conservative is coded as 1 if the respondent chooses a number between 6 and 10 in the ideology scale, ranging from 1 to 10. Next, we control for Race. Variables White and Black are coded 1 if the respondents identify their race as white and black, respectively. We also control for Political Knowledge which is coded 1 if the respondent answers the two questions on the president of Russia and the current secretary of state correctly, otherwise it is coded 0. We then control for Income. The variable coded 1 the respondent reports their family income $75000 or greater. Next, variable Gender is coded 1 if the respondent identifies themselves as woman, and 0 otherwise. We also control for Religiosity which is coded 1 if the respondents say they are very or rather religious, 0 otherwise. Next, variable Military Family is coded 1 if any member of the respondent’s immediate family is a member of the United States military. Finally, variable National Pride is coded as 1 if respondents say they are very or quite proud of being American, 0 otherwise.
While we acknowledge that there could still be unobserved variables that are not incorporated into the model, we believe that we have included a fairly large number of potential confounding variables in our analysis. Furthermore, in order to account for potential confounding bias, we run a t-test of difference in means for pre- and post-invasion samples for all of the variables (reported in the appendix) to check for balance between the pre-invasion and post-invasion samples. The results show that there is no statistically significant difference between variables in the pre- and post-invasion samples.
Results
Percentage of very unfavorable views in the pre- and post-samples along with T-test of means.
Note: Means and p-values presented in parentheses, respectively.
However, the results for the Russian people variable do not indicate any meaningful difference between the pre- and post-invasion samples. Only 1% of the respondents in the pre- and post-invasion samples expressed a “very unfavorable” view of the Russian people, indicating no reputational damage to the Russian citizens. This suggests that people do not significantly change their views of a country’s citizens for belligerent actions of their government. This finding is consistent with the observed behavior toward Russian citizens around the world since the beginning of the Russia–Ukraine conflict. While there have been some reports on violence and vandalism against Russian private individuals and businesses, 2 violent incidents against Russian citizens have been limited amid the ongoing war. Yet, the possibility of social desirability bias on the negative attitudes toward the Russian people should not be dismissed. Negative judgments toward a broad population of Russian citizens may be perceived as being associated with bigotry and racism, especially among younger generations who usually hold more liberal and egalitarian values. Therefore, the results on unfavorable views of Russian citizens need to be interpreted with caution.
We present the coefficient plot of the regression analyses in Figure 1. The regression analysis shows that the negative attitudes toward Russia and the Russian government increase significantly after the invasion. Put it simply, individuals who took the survey after the invasion are significantly more likely to express a “very unfavorable” view of the country of Russia and the Russian government. Logistic regression coefficients with bars indicating 95% confidence intervals.
In addition to the independent variable, two control variables also demonstrate a significant association with the outcome variable: conservative and national pride. According to the regression results, conservatives are significantly less likely to express a “very unfavorable” view of the Russian government compared to liberals. Also, individuals with a high level of national pride are significantly more likely to hold a very negative view of Russia as a country. Given the increasing geopolitical rivalry between the United States and Russia, it is not surprising that national pride is associated with negative views of Russia. Other control variables do not demonstrate a statistically significant association with the dependent variable. 3
The significant change in attitude toward the country and the government after the invasion is evident in the predicted probability plot presented in Figure 2. The probability of a very unfavorable view toward the country and the Russian government increases by 16 and 26 points in the post-invasion sample, respectively.In order to understand the magnitude of attitude change across different demographic, social, and political subgroups, we present the percent of individuals per subgroup with a very unfavorable view of Russia as a country and the Russian government in Figure 3. In general, there is a clear difference between the pre- and post-invasion samples across all subgroups. Yet, the magnitude of attitude change shows some degrees of variation. With a 24-point increase in the very unfavorable opinion of Russia, Republican and religious subgroups show the most radical attitude change toward the country. The percentage of a very unfavorable view of Russia for the religious subgroup in the pre-invasion sample was less than the average (the lowest in the pre-invasion sample) but it exceeds the average number in the post-invasion sample. The least attitude change happens in the military family subgroup with an 8.5 point increase. However, this finding should be understood with respect to the fact that the military family subgroup had the highest percentage of very unfavorable views of Russia in the pre-invasion sample. In other words, individuals with a military family member had already developed a very negative view of Russia before the invasion. Predicted probability toward Russia and the Russian government before and after the invasion. Very unfavorable attitudes toward the country and government before and after the war. Note: The left and right dashed lines in both plots show the mean number for the pre- and post-invasion samples, respectively.
For the very unfavorable view of the Russian government, the Republican subgroup exhibits the largest increase from the pre-invasion to the post-invasion sample. Nevertheless, it is important to note that Republican respondents had the least percentage of the “very unfavorable” response in the pre-invasion sample. Also, with a 32.4-point increase in the very unfavorable attitude, the religious subgroup demonstrates a meaningful attitude change after the invasion. Similar to the country graph, the military family subgroups had the highest percentage of the very unfavorable view before the invasion and at the same time, demonstrate the least change after that.
The general pattern shows that subgroups with a pre-existing negative opinion of Russia demonstrate higher levels of stability in their attitude, although there are some exceptions. For instance, the proud American subgroup (individuals who are proud of their American identity) had a pre-existing negative view of Russia as a country (around 5% higher than the average of the pre-invasion sample) but their attitude become significantly more negative after the invasion (around 7% higher than the average of the post-invasion sample). The proud American subgroup has the second most negative attitude toward Russia in the post-invasion sample in the country graph.
A similar pattern happens with the Democrat subgroup in the government graph. The percentage of a very unfavorable view of the Russian government is 63.5 (7.5% higher than the average of the pre-invasion sample) in the Democrat subgroup and it increases to 82.8% in the post-invasion sample (2.2% higher than the mean). Overall, attitude change is quite meaningful after the invasion, although the magnitude varies from subgroup to subgroup.
Conclusions
Our study provides empirical evidence on the reputational cost of military aggression by using survey data that was collected amidst the Russia-Ukraine conflict. Our study offers a nuanced and more complex understanding of war and global public opinion. The findings demonstrate that unjustified military aggression does create negative global public opinion on the aggressor party but the negative attitude is mostly directed at the government and the country. The results show that the ordinary citizens of the aggressor government remain immune to the reputational damage.
While our study offers a nuanced picture of war and international image, several cautions should be noted. First, although our survey started 3 days before the attack, public discussion on the possibility of a Russian attack had already begun a few weeks before the attack. Especially, reports from the United States intelligence agencies on the likelihood of a Russian invasion prompted a wide range of speculations among political pundits as well as the public before the attack. Therefore, attitudes on Russia might have started to harden even before the invasion. Therefore, it is possible that our study underestimates the full magnitude of attitude change.
It is also important to emphasize that our study uses a student sample. While using student samples is a common practice in social science research, it is also argued that they are less externally valid compared to the samples drawn from the general population. We compared our results on unfavorable attitudes toward Russia to those of PEW and Gallup polling conducted in 2020 and 2021, respectively (the graph is reported in the appendix). Overall, our pre-invasion sample shows less negative attitudes toward Russia compared to the PEW and Gallup samples (5% less than PEW and 10% less than Gallup). Given the difference, it is possible that our sample somewhat underestimates negative attitudes toward Russia.
Furthermore, it is critical to highlight that the survey was carried out in the United States. While we have witnessed a fairly large number of rallies in support of Ukraine in Europe and North America (Schwartz, 2022), evidence for broad anti-Russian sentiments in the rest of the world is not quite as strong. It could be due to the difficulty of collective action in authoritarian environments (especially in countries with strong ties to Russia) or because of widespread anti-Western sentiments in a large number of countries in the Global South (Aydin, 2007; Lewis, 1993) where many citizens might perceive the war as a confrontation between Russia and the West. Given the presence of anti-Western sentiments in the Global South for a variety of historical reasons (e.g., colonialism or past interventionist policies), the findings need to be interpreted with caution. While the results may show us the magnitude of negative attitudes toward Russian in the United States or perhaps most of Europe, they do not necessarily travel to the Global South.
Supplemental Material
Supplemental Material—The reputational cost of military aggression: Evidence from the 2022 Russian invasion of Ukraine
Supplemental Material for The reputational cost of military aggression: Evidence from the 2022 Russian invasion of Ukraine by Peyman Asadzade and Roya Izadi in Research & Politics.
Footnotes
Acknowledgments
We would like to thank the editor and the three anonymous reviewers for their time and helpful feedback. We would also like to thank Kim Fridkin, Director of the Political Science Experimental Laboratory at Arizona State University, and Trudy Horsting, the lab coordinator, for their help with recruiting participants and data collection.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
