Al Qaeda Versus Big Brother: Anxiety About Government Monitoring and Support for Domestic Counterterrorism Policies
Explanatory models of attitudes toward U.S. domestic counterterrorism policy routinely incorporate individual concern over terrorism, but uniformly disregard concern about the government’s use of domestic surveillance. Indeed, one of the most prominent works of this kind explicitly argues that ordinary Americans will not perceive that government monitoring targets people like themselves and thus domestic surveillance programs will not generate anxiety. We question this assumption on theoretical and historical grounds. Our research uses a unique probability sample survey to demonstrate that significant portions of ordinary Americans feel anxious about domestic government monitoring. Moreover, the results show that anxiety about government monitoring negatively relates to attitudes toward domestic counterterrorism policies. Although never included in previous models, and even plainly dismissed as irrelevant, felt anxiety about government monitoring importantly predicts attitudes about domestic counterterrorism policies.
KeywordsAnxiety Domestic counter-terrorism policies Government monitoring Attitudes about terrorism Public opinion
In part because public opinion has been decisively important for the passage of U.S. domestic counterterrorism policies (Mueller 2009; Domke 2006; Grandy 2003), political scientists have sought to understand the sources of these opinions. The dominant explanatory approach predicts support for domestic counterterrorism policies using cognitive assessments of terrorist threat and or affective reactions to the prospects of terrorism (e.g., Davis and Silver 2004; Huddy et al. 2005). This terrorism-centered approach coincides with the widely held axiom that people generally become deferential to a newly invasive state when threatened by terrorism. However, as we will describe, this literature explicitly assumes that individuals do not perceive the U.S. government’s domestic actions countering terrorism likely targets ordinary Americans (e.g., Huddy et al. 2005). Consequently, felt anxiety stemming from broad based U.S. government monitoring has been excluded as a potential explanatory factor in models predicting support for domestic counterterrorism policies.
We propose that the U.S. government’s domestic policies designed to reduce terrorism make many citizens anxious about a loss of liberty similar to the way that terrorism makes people anxious about a loss of security. In our view citizens generally want both security and liberty and are anxious when either value is at risk. Therefore, the balance of anxiety about both terrorism and domestic government monitoring should powerfully dictate individuals’ support for domestic counterterrorism policies. The paper’s empirical analysis predicting attitudes toward domestic counterterrorism policies generally supports this assertion.
The framework used in this paper, which simultaneously considers two distinct anxieties, also helps comprehend and resolve a key empirical puzzle in the literature. As we will describe, past work consistently finds that anxiety about terrorism fails to independently predict attitudes about domestic counterterrorism policies. This persistent result is surprising in that Cognitive Tendency Theory suggests that individuals anxious about terrorism would become risk averse and therefore support domestic public policies that would reduce the risk of future terrorism.
Attitudes Toward Domestic Counterterrorism Policies
Political scientists have endeavored to understand why individuals support or oppose various domestic counterterrorism policies, including government surveillance of telephone and email communication, development of national ID cards, the use of torture, and spending on Homeland Security. Although important work predicting attitudes regarding domestic counterterrorism policies has shown the significance of such factors as authoritarianism (Hetherington and Weiler 2009; Cohrs et al. 2005) and ethnocentrism (Kam and Kinder 2007; Kinder and Kam 2009), a core component of nearly all explanatory models include measures of individual concern about terrorism (following Davis and Silver 2004).
An important early study by Davis and Silver (2004) finds that the greater individuals’ terrorism threat level, the more likely they are to support domestic counterterrorism policies. Multiple other studies confirm the connection between cognitive perceptions of the likelihood of another terrorist attack and support for numerous domestic counterterrorism policies (Hetherington and Weiler 2009; Huddy et al. 2005, 2007b; Cohrs et al. 2005; Merolla and Zechmeister 2009a; Kam and Kinder 2007; Joslyn and Haider-Markel 2007). In addition, the relationship between terrorism threat and support for domestic counterterrorism policies appears robust across both observational and experimental studies. Whereas most of this empirical work relies on survey data, Haider-Markel et al. (2006) use controlled experiments to demonstrate that exposure to threatening terrorism news frames increases support of domestic counterterrorism policies. In sum, the results from this body of literature are consistent: terrorism threat positively associates with support for domestic counterterrorism policies.
The Importance of Anxiety About Terrorism
Huddy et al. (2005) make a critical contribution to this literature with their insight that the way people emotionally process the cognitive perception that a terrorist attack is likely to occur importantly influences their attitudes toward foreign and domestic counterterrorism policies above and beyond the cognitive assessment alone. Drawing on Cognitive Tendency Theory (an extension of Cognitive Appraisal Theory) from psychology, they argue that those who experience anxiety in threatening situations become increasingly risk averse as negative events becomes more imminent because anxiety heightens the amount of risk people perceive from a threat. This link between anxiety and risk aversion has important implications for the study of attitudes toward both foreign and domestic counterterrorism policies (Eysenck 1992; Lerner and Keltner 2001; Lerner et al. 2003).1
As Huddy and colleagues (2007a, b) show, those who experience anxiety in response to terrorism threat are more likely to be risk averse and thus against taking aggressive military action against terrorists, as this action could lead to more terrorism. Yet the central concern of this paper is not with attitudes toward foreign counterterrorism action but instead with the role of anxiety on attitudes toward domestic counterterrorism policies designed to reduce the risk of terrorism. Using Cognitive Tendency Theory’s identification of risk aversion as a central tendency of anxiety, Huddy and colleagues note that individuals anxious about terrorism should be supportive of domestic counterterrorism policies even as they oppose aggressive foreign military responses because domestic counterterrorism policies “are designed to minimize the future risk of terrorism” (2005, p. 596).
Nonetheless, once controlling for perceptions of the likelihood of a terrorist attack, terrorism anxiety does not relate to attitudes toward domestic counterterrorism policies (Huddy et al. 2005). That anxiety about terrorism does not positively associate with support of domestic policies designed to reduce the risks of terrorism goes against the core expectations of Cogntitive Tendency Theory with its focus on risk aversion.2 This unanticipated null finding is not an artifact of their sample; other studies using terrorism anxiety and terrorism threat to predict domestic counterterrorism policies also find a null relationship between anxiety and support for domestic counterterrorism policies (Haider-Markel et al. 2006; Huddy et al. 2007a, b). Only those studies that do not include a cognitive assessment of terrorism threat show the expected significant positive relationship between terrorism anxiety and domestic counterterrorism policies (Joslyn and Haider-Markel 2007; Schildkraut 2009).
The Potential for Anxiety from Domestic Counterterrorism Policies
In contrast to the large volume of literature that deems the level of concern about terrorism worthy of serious consideration as a determinant of support for domestic counterterrorism policies, no empirical study has used concern about the government’s domestic actions used to fight terrorism as an explanatory factor. We argue that this is a glaring omission. Both safety from terrorism and liberty from an overly intrusive government are relevant to many ordinary citizens when evaluating domestic counterterrorism policies.
Certainly the relevance of an overreaching government is not contentious to students of American history. Since the founding of the United States, individuals have clashed over establishing the proper balance between ensuring civil liberties from the government while effectively providing security against foreign and domestic threats. Madison and the other framers of the U.S. Constitution, with the recent memory of both King George III and the armed citizen rebellion in Massachusetts led by Daniel Shays, often grappled with the question of how to give the overhauled federal government enough power to both protect the citizenry against violent domestic threats while at the same time protecting the liberty of citizens from government itself (Madison et al. 1987). These early American thinkers relied heavily on liberal theorists such as Locke who recognized the importance of government to maintain security and property but who also deeply considered the interrelated problem of an overly intrusive state for the well being of a polity (Locke 2003). More recently, Posner refers to this tension as the Security-Liberty Frontier (Posner and Vermeule 2007). The problem with the Security-Liberty Frontier is that it is similar to the Pareto frontier identified in welfare economics. In both cases, there is no optimal point along the frontier. Security from foreign and domestic violence and liberty from inappropriate government interference are both desired but increases in either security or liberty usually means decreases in the other.3 Individuals are therefore often forced to make a tradeoff between two cherished values.
Not only have theorists long considered the problem of an overreaching state, but the types of policies currently employed in the domestic fight against terrorism have been seen as particularly anxiety inducing. Theorists as diverse as J.S. Mill and Foucault argue that inappropriate government monitoring and interference with citizen communication will create anxiety among the citizenry (Mill 1991; Foucault 1979). Empirical work as well known as Almond and Verba’s (1963) classic, The Civic Culture, even attributes the relative unwillingness of the citizenry in post war Italy and Germany to express political views in public to these countries’ legacies of closely monitoring even routine political activity. And while it is inaccurate to attribute fascist style surveillance to the practices found in the USA, the 1975–1976 Senate Church Committee detailed nearly one half century of questionable domestic surveillance by the FBI, Pentagon, and CIA (Donner 1980; Cunningham 2004). And upon later inspection, the vast majority of Counter Intelligence Program (COINTEL) files were found to be kept on citizens with no violent history or even a criminal past but instead compiled based on their political associations or beliefs (Donner 1980; Cunningham 2004). Whatever the actual reach of recent domestic counterterrorism policies, the twentieth century historical precedent certainly provides ordinary citizens a recent related experience from which to draw conclusions.
The mainstream media also has offered a more recent gauge of the relevance of these domestic counterterrorism policies to ordinary Americans. Revelations starting at the New York Times have shown that peaceful activities of U.S. citizens have been monitored through an NSA program conducted in secret without court oversight or approval (e.g., Lichtblau 2005). The PBS show Frontline detailed the story of a whistleblower from AT&T who testified in a federal trial that AT&T cooperated with the Federal government to attach their Internet backbone to government controlled data-mining super computers designed to monitor the entire data stream (Frontline 2006). After 2005 these types of stories have become commonplace; a recent one involves a Justice Department investigation demonstrating that the F.B.I. improperly classified a range of non-violent groups and political activists as terrorists in order to justify surveillance (Markon 2010).
Many scholars considering related lines of research have reached similar conclusions. Gibson suggests that “a reasonable view of public policy in the current period is that freedom has been restricted…[and] an appreciable threat exists that more draconian restraints on political freedom will be put in place in the future” (2008, p. 97).4 And though Cho et al. make no claims about whether these policies pose any genuine threat to innocent citizens, they do argue that these counterterrorism policies have the potential to create anxiety (2006). Our study also is not interested in discerning the extent to which these counterterrorism programs intrude on ordinary citizens’ lives. Instead it is enough that well-known theories, historical precedents, and recent events make quite plausible this paper’s expectation that government monitoring creates significant levels of anxiety in the U.S. public.
Contrary to our proposed framework, anxiety about government monitoring has been excluded from consideration as an explanatory factor in all extant empirical models of support for domestic counterterrorism policies. Implicitly, this suggests that the scholarly community rejects the relevance of concern about government monitoring in the U.S. context. Some of the best known work has been even more explicit, contending that individuals are unlikely to feel anxiety because Americans view these policies “as aimed at guilty others, not at them” (Huddy et al. 2005, p. 595). In this study we disagree with this assumption and argue that citizens generally want both security from terrorism and liberty from government. Therefore, given the nature of contemporary domestic counterterrorism policies, we expect many ordinary individuals to feel anxious about government monitoring.
Many individuals feel anxious about U.S. government monitoring.
Moreover, consistent with Cognitive Tendency Theory that identifies risk aversion as a core tendency of anxiety, we expect that felt anxiety about domestic counterterrorism should decrease support for these policies. Individuals anxious about government monitoring should become risk averse and generally will not support expanded domestic surveillance programs that exacerbate the risk of inappropriate government monitoring. More formally, this perspective leads to a hypothesis concerning the emotional reaction of anxiety and support for domestic counterterrorism policies.
All else equal, the level of felt anxiety about government monitoring will negatively relate to attitudes toward domestic counterterrorism policies.
Data and Key Measures
As part of a larger project designed to investigate political attitudes and behaviors during the summer of 2007, the Center for Survey Research and Analysis at the University of Connecticut administered a RDD telephone sample of adult, non-institutionalized residents of the contiguous United States. From August 15 through September 12, 2007 telephone numbers were dialed daily. Interviewers attempted up to 8 calls to contact potential respondents and used a Computer-Assisted Telephone Interviewing (CATI) system to administer questions and record responses.5 Only the key measures receive a detailed discussion below; the Appendix contains the question wording of the other measures.
Felt anxiety about U.S. government monitoring represents the key independent variable in the analysis. Anxiety, according to Cognitive Appraisal Theory, is a common but not automatic affective reaction when individuals cognitively perceive a negative event is likely to occur (e.g., Lerner and Keltner 2001). To measure anxiety about monitoring, the survey primes respondents about government monitoring: “Now, we would like to know how you feel when you think about the government monitoring the activities of people like you.” We, then, rotate randomly three separate questions using one of the following emotions: anxious, worried, and scared. “How_____does it make you feel—very_____, somewhat_____, a little_____, or not at all _____?” Past work demonstrates that this technique elicits specific emotions rather than general emotions (e.g., Marcus et al. 2006; Huddy et al. 2007a, b). In other words, this technique differentiates anxious people generally (i.e. trait anxiety) from people who get anxious when considering a specific circumstance or target (i.e. state anxiety). Following Huddy and colleagues who use similar indicators6, we expect that the anxious, scared and worried responses should form a monitoring anxiety dimension (2005; see also Best and Krueger 2011). The three anxiety measures are combined using factor analysis (pcf, in STATA). The mean of the anxiety factor loadings equals .87, with a low loading of .86 (see Tables 3 and 4 in Appendix). Although the extracted factor score for monitoring anxiety is used in the analyses to reduce measurement error, a simple additive scale reveals a Cronbach alpha of .83.
Cognitive Appraisal Theory and Cognitive Tendency Theory connect specific types of cognitions with particular emotions, which have various attitudinal and behavior tendencies (e.g., Lerner and Keltner 2001). Though the theory centers on the special behavioral tendencies of various emotions, the cognitions themselves often have an independent influence on attitudes and behaviors (e.g. Huddy et al. 2005). For example, the literature often considers both egocentric and sociotropic cognition about terrorism as potentially important independent factors, though sociotropic perceptions are the most salient and powerful cognitive predictor in models of counterterrorism policy attitudes (e.g. Joslyn and Haider-Markel 2007). We consider cognition in our context by including measures for sociotropic and egocentric perceptions about the likelihood of government monitoring. To measure perceptions of the likelihood of personal government monitoring, we ask respondents: “How likely is it that the government monitors your activities?—very likely (16.7%), somewhat likely (26.0%), not very likely (29.2%), or not at all likely (28.2%).” To tap perceptions of the likelihood of sociotropic government monitoring, we ask: “How likely is it that the government monitors the activities of ordinary Americans?—very likely (47.8%), somewhat likely (35.3%), not very likely (12.8%), or not at all likely (4.1%).” The dissimilar frequency distributions for these monitoring cognitions suggest substantial independence, while the correlation between the two items (.44) confirms that these are related but distinct cognitive indicators. Unlike the terrorism cognitions, these cognitive perceptions about the likelihood of monitoring should not automatically be considered a perceived threat. For some, perceiving that the government’s domestic counterterrorism programs are likely used in part for monitoring ordinary Americans (or their own activities) may be viewed as a necessary and justifiable response to the contemporary environment.
Attitudes about domestic counterterrorism programs are the dependent variable. Given the various media revelations demonstrating that the government, in their search for potential terrorists, frequently monitored law abiding Americans’ electronic communications with various data-mining monitoring techniques, the survey questions assess attitudes toward this key part of the contemporary debate—whether to allow the federal government to use broad based monitoring procedures in the fight against terrorism. Interviewers asked about a series of different surveillance programs including a general question about government monitoring and five specific questions about the respondent’s attitudes about potential broad based government monitoring programs. These programs include wiretapping, viewing phone contacts, reading emails, instituting a national ID program, and the use of facial recognition software. The exact question wording is displayed below followed by the distribution of responses.
Generally speaking, do you support or oppose the government monitoring the activities of ordinary Americans? Strongly oppose (40.8%), oppose (17.9%), neither support nor oppose (4.9%), support (21.1%), strongly support (15.4%).
Do you support or oppose the government wiretapping telephone conversations between ordinary Americans without a court order? Strongly oppose (59.2%), oppose (13.6%), neither support nor oppose (2.7%), support (12.8%), strongly support (11.7%)
Do you support or oppose the government looking at the telephone numbers dialed by ordinary Americans without listening to the content of their calls? Strongly oppose (28.9%), oppose (15.4%), neither support nor oppose (4.0%), support (26.0%), strongly support (25.8%)
Do you support or oppose the government reading ordinary Americans’ email messages without a court order? Strongly oppose (54.7%), oppose (13.1%), neither support nor oppose (4.3%), support (16.1%), strongly support (11.8%)
Do you support or oppose the government requiring ordinary Americans to carry a national identification card at all times to show to a police officer upon request? Strongly oppose (27.8%), oppose (10.5%), neither support nor oppose (3.5%), support (22.5%), strongly support (35.7%)
Do you support or oppose the government using facial-recognition technology to scan ordinary Americans attending public events? Strongly oppose (20.8%), oppose (12.6%), neither support nor oppose (7.7%), support (28.8%), strongly support (30.0%)
Because our theory does not distinguish between types of broad based surveillance practices these six measures are combined into a government monitoring index using principal components factor analysis (pcf, in STATA), which is used as the dependent variable in the regression analysis. By combining these measures into a factor score, this approach also has the advantage of reducing measurement error for attitudes toward domestic government counterterrorism monitoring. The Appendix includes the factor loadings for each item. Each of these items load highly onto one dimension, with a mean score of .75 and a low of .60. A reliability analysis reveals a Cronbach alpha of .84. Combined these tests support the appropriateness of combining these measures into one attitudinal dimension.
Methodology and Modeling
To test the first hypothesis we use a simple descriptive analysis of the felt anxiety about government monitoring variable. Although we do not expect nearly all individuals to feel anxiety about government monitoring, we do expect that non-trivial proportions of U.S. residents will feel this negative emotion. This descriptive characteristic is a necessary but not sufficient condition for our framework. Next, we initiate the multivariate analysis by running two ordinary least squares regression models using only the standard factors previously found to predict domestic counterterrorism policy preferences. These two shortened models, which exclude the monitoring items, serve to demonstrate the comparability of our data with previous studies. Finally, to test the second hypothesis we add the monitoring variables to the multivariate model. In this fully specified model, we also display the standardized regression coefficients as an indicator of the relative magnitude of each independent variable’s effect size on attitudes toward domestic counterterrorism policies.
Descriptive Results: Felt Anxiety About Government Monitoring
Frequencies for felt anxiety about government monitoring
Anxious about monitoring
Worried about monitoring
Scared about monitoring
Not at all
Multivariate Results: Predicting Domestic Counterterrorism Policy Attitudes
OLS models of support for domestic counterterrorism policies
Coefficient (standard error)
Coefficient (standard error)
Coefficient (standard error)
Sociotropic monitoring likelihood
Egocentric monitoring likelihood
Prob > F
The bivariate correlations help comprehend these results. The terrorism anxiety measure correlates with terrorism threat (.21) and terrorism threat positively associates with attitudes toward domestic counterterrorism policies. Therefore, in models without a terrorism threat variable, terrorism anxiety’s coefficient increases because it picks up some of the positive influence from the excluded terrorism threat measure. But terrorism anxiety, in this misspecified model, merely captures the effect of terrorism threat. When perceptions of threat from terrorism are included, the coefficient for terrorism anxiety is not statistically significant at conventional levels.
With the literature’s findings for terrorism replicated, the third column in Table 2 shows the results of the fully specified model. Demonstrating that substantial portions of Americans feel anxiety about government monitoring does not establish that this emotional condition independently predicts attitudes toward domestic counterterrorism policies. The weight of terrorism, with its bearing on the very lives of Americans, may trump any feelings of anxiety about government monitoring when shaping attitudes toward domestic counterterrorism programs.
The estimates from the fully specified regression model reported in Table 2, column 3 strongly supports Hypothesis 2. The estimates indicate that we can have very high confidence that monitoring anxiety negatively relates to attitudes toward domestic counterterrorism programs. Moreover, this effect appears large. The standardized coefficients reported to the right of the third column in Table 2 show that monitoring anxiety’s effect size is considerable (β = −.326). This effect size is larger than the two terrorism indicators combined (terror threat β = .129 & terror anxiety β = .100). And monitoring anxiety’s β clearly exceeds the Beta’s of other well-known predictors of attitudes toward domestic counterterrorism programs such as authoritarianism (β = .102) and ethnocentrism (.130). Altogether these estimates suggest that when individuals feel anxiety about government monitoring we can expect considerably lower support for domestic counterterrorism policies compared to those not anxious about monitoring.
Standardized coefficients are useful in that they offer a comparative statistic for an independent variable’s magnitude of influence. Yet, the overall effect of monitoring anxiety on aggregate public opinion also depends on the relative proportion of the population that is anxious about terrorism and monitoring. In other words, monitoring anxiety may powerfully predict domestic counterterrorism attitudes but only for a comparatively small portion of the public. Relative to anxiety about government monitoring, many more U.S. residents are anxious about terrorism; 43% of the public are at least somewhat anxious about terrorism, whereas only 28% of the public feel comparably anxious about government monitoring. Another way to think about this is that monitoring anxiety has a fairly strong negative influence on domestic counterterrorism attitudes for about a quarter of the population, whereas terrorism anxiety has a moderate positive influence for nearly one-half of the population.
One other finding in Table 2 deserves some attention. Once controlling for monitoring anxiety7 (Table 2, column 3) the terrorism anxiety coefficient becomes statistically significant at the p < .01 level. This represents the first study to show a statistically significant positive influence of terrorism anxiety on attitudes toward domestic counterterrorism policies once controlling for terrorism threat. This finding is contrary to Huddy and colleagues (2005) empirical results but does support their original theoretical expectations of a positive association. Further considering the interconnection between the two types of anxiety clarifies this outcome.
Past empirical work demonstrates that individuals anxious about one circumstance tend also to feel anxious about dissimilar circumstances (Berenbaum et al. 1995). Therefore, in an environment where terrorism is possible, we suggest that it is likely that the same individuals will tend to be anxious about both terrorism and government monitoring associated with counterterrorism programs. The .19 correlation between terrorism anxiety and monitoring anxiety supports this expectation that those anxious about terrorism also have a propensity to be anxious about government monitoring. Yet, while these two types of anxiety positively relate, they exert opposite effects on support for domestic counterterrorism policies. It seems then that Huddy and colleagues’ (2005) null findings for terrorism anxiety may be explained by the fact that their single anxiety measure captures terrorism anxiety but critically also captures anxiety about government monitoring. Because anxiety about monitoring has a strong negative association with attitudes toward domestic counterterrorism policies, a measure of terrorism anxiety entered alone in a model would attenuate toward insignificance.
Numerous studies use concern about the prospects of terrorism as the central explanatory factor when predicting support for domestic counterterrorism policies. Yet, despite theoretical, historical, and contemporary grounds suggesting the related importance of government monitoring to ordinary citizens, to date none of these studies’ explanatory frameworks incorporate individuals’ degree of anxiety about the government’s domestic use of surveillance. In addition, some of this body of work that focuses on terrorism as the central mechanism to explain support for domestic counterterrorism policies explicitly argues that ordinary Americans will not be anxious about government monitoring because individuals will perceive these programs only target guilty others. In this paper a unique probability sample survey is used to demonstrate that significant portions of ordinary Americans feel anxious about domestic government monitoring. As important, the results show that anxiety about government monitoring independently negatively relates to attitudes toward domestic counterterrorism policies. Although never included in previous models, and even explicitly dismissed as irrelevant, felt anxiety about government monitoring importantly predicts support for domestic counterterrorism policies.8
Given its importance predicting attitudes toward domestic counterterrorism policies, closely related areas of research suggest other venues in which anxiety about government monitoring may offer explanatory power (e.g. Haider-Markel et al. 2006; Merolla and Zechmeister 2009b; Campbell 2003). Future research may fruitfully consider whether anxiety about government monitoring influences individuals’ level of participation in politics as well as the intensity and types of political criticisms expressed to other citizens face-to-face, over the internet, and in correspondences with elected officials. Monitoring anxiety could alter voter’s decision-making processes such as the weights given to candidate qualities (i.e. trustworthiness, leadership) when evaluating politicians. And, of course, scholars could reconsider the link between monitoring anxiety and domestic counterterrorism policy attitudes using experimental manipulations to exogenously produce anxiety about government monitoring.
The implications of this research go beyond the identification of a previously unrecognized explanatory factor that substantially increases the variance explained in attitudes toward domestic counterterrorism policies9. The framework used in this analysis helps resolve a key puzzle in the extant literature. Despite original theoretical expectations that terrorism anxiety should positively associate with support for domestic counterterrorism policies, past work fails to demonstrate any relationship (Haider-Markel et al. 2006; Huddy et al. 2005, 2007a, b). Our work suggests that the source of this unexpected null result is not located within the theory that links anxiety about terrorism with a desire to reduce risk by supporting domestic counterterrorism programs, but instead with a questionable assumption about the irrelevance of government monitoring in the contemporary U.S. context. Our empirical estimates for anxiety about terrorism support the original theoretical expectations; the key is controlling for the positively associated but countervailing force of anxiety about government monitoring. We suggest that many U.S. residents value both security and liberty and therefore anxiety will occur when either value is at risk. And because monitoring anxiety negatively relates to attitudes about domestic counterterrorism policies, a measure of terrorism anxiety entered alone in a model of support for domestic counterterrorism policies would be pushed and pulled in opposite directions and as a result tend toward statistical insignificance. Only once also controlling for anxiety about government monitoring does anxiety about terrorism display the originally expected positive association with attitudes toward domestic counterterrorism policies.
Anxiety has been perhaps the most widely studied emotion, forming the core of Affective Intelligence Theory (Marcus and MacKuen 1993; Marcus et al. 2000) as well as many other prominent studies on affect (e.g., Huddy et al. 2007a, b; Brader et al. 2008). Yet, these studies do not use different types of anxiety in the same model to predict political attitudes or behavior. The implicit assumption is that only one type of anxiety is relevant for shaping attitudes about a particular political issue. We diverge from this common approach and demonstrate that distinct anxieties may arise from the same political context, display only modest positive correlations, and have divergent effects on political attitudes. Perhaps more than anything else, our work suggests that those interested in the connection between anxiety and political behavior would benefit by considering the many contexts where individuals should simultaneously experience different types of anxiety.
This approach is consistent with Zaller’s (1992) well-known framework, which argues that competing and even contradictory considerations, including emotional reactions, are the norm rather than the exception when individuals formulate political attitudes. “People may have one reaction to an issue that would cause them to favor it and another that would cause them to oppose it, but—and here is the heart of the argument—for most people, most of the time, there is no need to reconcile their contradictory reactions to events or issues. Each can represent a genuine feeling, capable of coexisting with opposing feelings” (Zaller 1992, p. 93). Unlike the persistent practice employed in studies about emotions and politics, we would expect that opposing types of anxiety would “coexist” in people’s minds for a wide range of other issues. People anxious about the budget deficit are also apt to be anxious about high taxes. Those anxious about high inflation are also likely anxious about unemployment. Individuals anxious about Wall Street risk taking are probably also anxious about government over-involvement in the economy. And people anxious about environmental degradation should also tend to be anxious about economic stagnation. Because the public policy solution to one of these concerns often is accompanied by the exacerbation of the second concern, the distinct types of anxiety will likely moderately covary but have opposite effects on political attitudes toward government policies intended to deal with these interconnected issues. The results from our analysis suggest that in these cases where tradeoffs must be made between two cherished goals or values, incorporating these related but distinct anxieties is crucial to validly estimating statistical relationships.
As Huddy et al. (2005) note, the focus on threat without also considering anxiety leads to unstable findings. Whereas studies of domestic counterterrorism routinely show a positive association between terrorism threat and support for these policies, related studies considering foreign military action find unstable results, with some studies linking threat to more supportive attitudes and others finding no relationship (e.g. Arian 1989; Brody and Shapiro 1989; Tir 2010). The approach taken by Huddy et al. (2005), which takes seriously the role of anxiety, helps resolve the micro foundation of this puzzling inconsistency. They find that when controlling for anxiety, threat has a direct positive association with support for aggressive military action that strikes out at the perceived source of the threat (Huddy et al. 2005). However, when individuals experience anxiety about terrorism their support for retaliatory military policies wanes because the aggressive retaliatory action may increase the future risk of terrorism. Even when also threatened, anxiety leads individuals to support passive foreign policies that are perceived as less risky. This framework helps explain why terrorism threat at times does and does not generate support for military action against terrorists, with the key factor being the degree of associated anxiety (Huddy et al. 2005).
Though Huddy and colleagues “had, in fact, suggested that anxiety could increase support for such policies because they might reduce the risk of terrorism…the coefficients for anxiety are all positive, …but none are statistically significant at conventional levels, and none of the effects are substantively large. Perceived threat increases support for heightened surveillance policies but anxiety does not” (Huddy et al. 2005, p. 604). It appears that while the consideration of anxiety's associated tendency of risk aversion helps resolve the puzzle of the inconsistent relationship between terrorism threat and support for foreign counterterrorism policies, their approach opens a new puzzle involving terrorism anxiety and domestic counterterrorism policies that defies original theoretical expectations.
Sniderman et al. (1996, p. 244) make a similar argument, exercising liberty “unavoidably collides with other values” causing a “contestability of rights” where people’s commitment to liberty clashes with other values, in this case safety and security. Of course, some of the most important political philosophers, starting with Hobbes, note that individuals can only achieve liberty in an environment where the government or leviathan protects their security.
Wolfendale (2006, p. 764) concurs, “[c]urrent counterterrosim measures pose a very real threat to the lives and well being of innocent citizens…this legislation threatens security from the state itself—security from being investigated, detained, controlled, and placed under surveillance without ones knowledge.”
The response rate was 36.2% (using AAPOR Response Rate 6), which yielded over one thousand respondents. Although not as high as some national surveys such as the National Election Studies or the General Social Surveys, research has shown that this response rate is now common and does not necessary yield lower quality data (Keeter et al. 2000).
For their context specific anxiety dimension they use the terms nervous, scared and afraid.
The statistically significant results for terrorism anxiety can be attributed to monitoring anxiety rather than the two cognitive monitoring items. In an alternate model that includes monitoring anxiety but excludes the two monitoring cognitions, the coefficient for terrorism anxiety is nearly identical to that found in the baseline model. However, terrorism anxiety is not significant when monitoring anxiety is excluded and the two cognitive measures are included.
The results also suggest a reconsideration of the role of anxiety during the George W. Bush administration. In Mueller’s (2009) recent book, Overblown: How Politicians and the Terrorism Industry Inflate National Security Threats, and Why We Believe Them, he argues that during this period anxiety about terrorism was an especially effective tool for increasing public support for domestic counterterrorism programs such as the Patriot Act and NSA wiretapping. Our research implies that while the independent influence of terrorism concern does conform to this conventional storyline, it is incomplete without a consideration of the interconnected concern about the executive branch’s domestic surveillance programs.
The R2 increases from .33 in the baseline model to .42 in the model that adds the monitoring items (Table 2).
- Almond, G., & Verba, S. (1963). The civic culture: Political attitudes and democracy in five nations. Princeton: Princeton University Press.Google Scholar
- Cho, W. K., Gimpel, J., & Wu, T. (2006). Clarifying the role of socioeconomic status in political participation: Policy threat and Arab American mobilization. Journal of Politics, 68, 977–991.Google Scholar
- Cunningham, D. (2004). There’s something happening here: The new left, the Klan, and FBI counterintelligence. Berkeley: University of California Press.Google Scholar
- Donner, F. J. (1980). The age of surveillance. New York: Knopf.Google Scholar
- Eysenck, M. W. (1992). Anxiety: The cognitive perspective. London: Lawrence Erlbaum.Google Scholar
- Foucault, M. (1979). Discipline and punish: The birth of the prison. New York: Vintage.Google Scholar
- Frontline. (2006). Spying on the homefront. PBS. http://www.pbs.org/wgbh/pages/frontline/homefront/
- Huddy, L., Feldman, S., & Cassese, E. (2007a). On the distinct political effects of anxiety and anger. In R. Neuman, G. E. Marcus, A. N. Crigler, & M. MacKuen (Eds.), The affect effect: Dynamics of emotion in political thinking and behavior (pp. 202–230). Chicago: University of Chicago Press.Google Scholar
- Kam, C. D., & Kinder, D. R. (2007). Terror and ethnocentrism: Foundations of American support for the War on Terrorism. Journal of Politics, 69, 318–336.Google Scholar
- Kinder, D. R., & Kam, C. D. (2009). Us against them: Ethnocentric foundations of American opinion. Chicago: University of Chicago Press.Google Scholar
- Lichtblau, Eric. (2005). F.B.I. watched activist groups, new files show. New York Times, December 20. Google Scholar
- Locke, J. (2003). Two treatises of government and a letter concerning toleration. Binghamton: Vail-Ballou Press.Google Scholar
- Madison, J., Hamilton, A., Jay, J., & Kramnick, I. (1987). The Federalist Papers. Penguin Classics.Google Scholar
- Marcus, G. E., MacKuen, M. B., Wolak, J., & Keele, L. (2006). The measure and mismeasure of emotion. In D. Redlawsk (Ed.), Feeling politics: Emotion in political information processing (pp. 31–46). New York: Palgrave Macmillan.Google Scholar
- Marcus, G. E., Neuman, W. R., & Mackuen, M. B. (2000). Affective intelligence and political judgment. Chicago: University of Chicago Press.Google Scholar
- Markon, J. (2010). FBI probes were improper, Justice says. Washington Post. September 20th.Google Scholar
- Merolla, J. L., & Zechmeister, E. J. (2009a). Democracy at risk: How terrorist threats affect the public. Chicago: University of Chicago Press.Google Scholar
- Mill, J. S. (1991). On liberty and other essays. Oxford: Oxford University Press.Google Scholar
- Mueller, J. (2009). Overblown: How politicians and the terrorism industry inflate national security threats, and why we believe them. New York: Free Press.Google Scholar
- Posner, E., & Vermeule, A. (2007). Terror in the balance. Oxford: Oxford University Press.Google Scholar
- Sniderman, P., Fletcher, J. F., Russell, P., & Tetlock, P. E. (1996). The clash of rights liberty, equality, and legitimacy in pluralist democracy. New Haven: Yale University Press.Google Scholar