Introduction
Government adherence to international human rights law, especially when human rights obligations fail to coincide with national security interests, is presumably driven in part by the constraining effect of public opinion.Footnote 1 Yet findings are mixed on how much international law affects public opinion. Some show that the public is highly sensitive to international normative understandings.Footnote 2 Others suggest the public is more than ready to subordinate legal and normative concerns to strategic interests, especially when national security issues are at stake.Footnote 3 Much IR scholarship in this area references or uses public opinion polls to gauge how much political attitudes towards the use of force change when respondents are primed to think about international law.Footnote 4
In an effort to expand this research agenda beyond the debate over how much international law ‘matters’, we focus critical attention to the relationship between public opinion surveys and the ways in which international law and norms are disseminated to the public in the first place. IR literature on norm diffusion stresses the role that international organisations and advocacy NGOs play in shaping preferences and normative understandings, with public opinion data often used as a dependent variable.Footnote 5 However, no study we know of has considered the ways in which citizens are socialised into certain understandings of international law through participation in and consumption of public opinion surveys themselves.
We connect IR literature about norm effects with public opinion literature about survey effects to develop a new testable hypothesis: that public opinion polls may be structured so as to not simply measure citizen attitudes, but also to shape respondents’ understandings of international laws and norms. We ground this discussion empirically by analysing a new dataset of post-9/11 survey questions on two US national security policies at odds with international human rights norms: ‘enhanced interrogations’ and ‘targeted killing’. Each in different ways, these two issues offered pollsters chances to engage or shape respondents’ understanding of the relevant laws or legal debates around important national security issues.
By comparing the discursive structure of poll questions to what IR scholars know about international norm socialisation, we show that polls around both these issues were structured discursively in ways likely to discount or weaken, rather than strengthen, respondents’ understanding of the essential character of important international laws and legal debates. This preliminary analysis suggests the need for experimental research on survey framing and international law socialisation in the national security area. If our hypothesis is borne out – that researchers measuring attitudes towards national security issues could be shaping respondent knowledge of the law applicable to those issues – it would have methodological implications for those conducting survey experiments on international norms, substantive implications for scholars of norm dissemination, and policy implications for norm advocates.
International law, public opinion, and the political power of surveys
An emerging literature in IR explores the relationship between international norms and public attitudes towards the use of force.Footnote 6 This is an important line of inquiry given the assumption that state compliance with international norms, expressed through treaty law, results from the constraining effect of domestic and international public opinion. Presumably, a socialisation process occurs whereby prohibition norms, once codified, are used by domestic constituencies to either to hold their own political elites accountable,Footnote 7 or to ‘name and shame’ recalcitrant third party states,Footnote 8 or both in tandem – the ‘boomerang effect’.Footnote 9 With public opinion so mobilised, politicians are thought likely to consider the ‘electoral implications’ of their foreign policy actions.Footnote 10 The reputational cost to policy elites for violating international legal agreements is said to be a linking mechanism between public opinion and foreign policymaking, and a raison d’être for norm-building efforts by NGOs.Footnote 11
Current research on the impact of international law on attitudes towards international security issues exhibit mixed findings. Many studies find a significant effect of international law on public attitudes.Footnote 12 Sarah Kreps and Geoffrey Wallace's research on drones suggests that American citizens are actually more persuaded by arguments that invoke international legal principles than efficacy.Footnote 13 James Walsh argues the promise of sanitised warfare that accompanies drones and precision weapons may result in a higher aversion among American populations to civilian casualties.Footnote 14 Wallace's findings on support for torture suggest that international law will affect public opinion even more in cases where external pressures to engage in the policy are at their highest.Footnote 15
However, other experimental studies cast doubt on these findings: Adam Chilton and Dustin Versteeg find no statistically significant decrease in support for torture when respondents are primed to consider international legal norms.Footnote 16 Some find that Americans are more likely to utilise an interest-based ‘logic of consequences’ when weighing support for foreign policy options; interest-based considerations that appear to offset international legal concerns include the level and context of US military fatalities,Footnote 17 the expectation of victory,Footnote 18 force protection,Footnote 19 and military utility.Footnote 20 Some findings are ambiguous.Footnote 21
Rather than resolve this debate, we reframe the question and consider whether the act of responding to and consuming information from public opinion surveys themselves might affect intersubjective understandings about the law. Most scholars who study the socialisation of political elites and mass publics into international normative understandings posit either formal avenues of statecraft, media, or global civil society as the primary socialising forces.Footnote 22 For example, in the area of human rights law, according to David Davis, Amanda Murdie, and Coty Garnett Steinmetz, INGOs are the ‘makers and shapers’ of domestic public opinion.Footnote 23 Beth Simmons argues that human rights organisations alter the preferences of domestic audiences by transmitting knowledge of international law through activities such as shaming campaigns,Footnote 24 and Sally Engle Merry argues international law is ‘translated’ into domestic practice by state and intergovernmental bureaucracies and NGOs.Footnote 25 Similar findings exist in the area of international security and international political economy.Footnote 26
We concur that the cumulative impact of norm leadership among government officials and INGO/NGO engagement can impact public awareness of international norms, which can then encourage greater restraint on the part of governments, even in least likely cases like national security. That this effect is stronger in established democracies than in non-democracies confirms that politicians are relatively sensitive to public opinion when re-election depends on it.Footnote 27 We suspect, however, that the ability of state norm leaders and global civil society to inform public understandings of international law may be overstated relative to other ‘everyday sites’ through which ordinary citizens learn about international politics – including both information and misinformation about international legal obligations. This may be particularly true when norm entrepreneurs must compete with norm anti-preneurs – states and/or INGOs who contest the content or validity of existing international laws and norms.Footnote 28
Public opinion polls may be one such site for the construction of, rather than the mere communication of, political preferences. Political psychologists have long observed that survey methods have the potential to create the very ideas they are trying to measure.Footnote 29 Justin Lewis argues in Constructing Public Opinion that a symbiotic relationship exists between the technology of polling, the news media that report the polls, and people's sense of what they want.Footnote 30 Building on Blumer, Lewis posits that polls – ostensibly neutral instruments – instead exert an agenda-setting effect similar to the news media by telling people not what to think, but rather what to think about, what others are thinking about, and how to think about what their opinions might be, through framing and structuring the discursive process by which answers are reported and recorded.Footnote 31 Similarly, Lisbeth Lipari conceptualises surveys as conversational settings, albeit where the structure of the communicative process is controlled by the researcher, and analyses the discursive structure of polls to show how they presuppose certain assumptions and constrain political expression, generally in the service of status-quo arguments.Footnote 32
John Zaller and Stanley Feldman have also challenged the notion that public opinion polls are even capable of measuring preferences.Footnote 33 The authors suggest that there is no such thing as the ‘true attitude’ among respondents – rather, survey answers reflect thoughts that are ‘the most accessible in memory at the moment of response’.Footnote 34 This moment of interacting with the survey question may actually ‘decide’ respondent attitudes.Footnote 35 Other public opinion scholars focus more heavily on the effects and conditions of elite framing on issues as diverse as a Ku Klux Klan rally,Footnote 36 affirmative action,Footnote 37 and the 2003 Iraq War.Footnote 38
Dissemination of survey top-lines through the media may also impact the wider public's understandings of the law and its policy preferences regarding government compliance. Lipari writes, ‘Although the participants in poll interviews themselves are relatively few in number, the symbolic resonance of poll discourse is amplified in the interpretations of pollsters, policymakers, newspaper readers, TV audiences, and so forth.’Footnote 39 Indeed, as Lewis shows through various case studies, political elites count on this: the structure of specific poll questions produces results that appear to support mainstream political narratives, and when the media reports those results, it further entrenches or influences ‘opinions’, which are then measured recursively.Footnote 40 The way in which media outlets portray the information can affect how citizens internalise political knowledge.Footnote 41 Not only does the media signal which issues are important to mass publics,Footnote 42 it may co-actively construct policy preferences, understandings, and outcomes alongside political leaders and citizens.Footnote 43
One of the best applications of this hypothesis in the international law literature is a recent study by Sarah Kreps on drone strikes.Footnote 44 Kreps shows that most public opinion polls on drones incorporate contested assertions, such as that drone strikes target terrorists rather than civilians. Moreover, she shows, polls are silent on certain questions such as whether or not drone strikes are legally authorised under US or international law. As such, the form of the questions not only predetermine the answers but also constrain the terms of debate. Using survey experiments, Kreps shows how when researchers shift the terms of the debate through reworded survey questions, support for drone strikes turns out to be much lower than previously reported, and Americans are indeed sensitive to legal concerns.
Building on Kreps's theory of survey effects, and Lipari's understanding of surveys as conversational settings, we hypothesise that the characterisation of international law in the discursive structure of national security poll questions constitutes an under-studied site for citizen socialisation into understandings of its content and applicability, particularly in cases where the law is unfamiliar or contested in the public debate around an issue. Indeed, the literature on norm socialisation, diffusion, contestation, and decline suggests it is precisely forms of ‘communicative action’ like this where intersubjective understandings of norms as relevant social facts get reinforced or contested, and where norm socialisation (or counter-socialisation) can potentially occur.Footnote 45 Given what is known about how international legal norms get affirmed or contested as information about them is disseminated in conversational settings or political speech, we expect four discursive properties of survey questions, if present, would be particularly likely to shape substantive perceptions, or misperceptions about international law – even as opinions about policies that violate those laws are presumably being objectively measured.
First, as Kreps notes, whether or not international law is mentioned at all in the preface to a survey question on a national security issue would signal in a discursive setting whether or not law is a relevant consideration in a particular policy debate.Footnote 46 This is consistent with a constructivist understanding of norm socialisation (and counter-socialisation) as a series of speech acts in which ethical norms and international laws are affirmed, contested, or ignored. For example, Nina Tannenwald looks for evidence of an emergent nuclear taboo in the discourse of policymakers, who first began incorporating ethical concerns into their discussions of strategic gains and military necessity, and eventually prioritised a certain understanding of ethics over interests entirely in their post-Cold War decision-making.Footnote 47 Conversely, in situations where international law and norms are not even mentioned as relevant considerations, scholars of norms would view this as not only a signal of a weakening norm, but a speech act with a recursive effect on such weakening.Footnote 48
Second, assuming that a survey question does even mention international law, its phrasing can signal which international laws may matter. In cases where multiple international instruments could be relevant, where the applicability of specific rule-sets is contested, or where there are multiple avenues of legal debate, ‘grafting’ the legal debate to one set of laws rather than another can be a politically significant move.Footnote 49 For example, as Jutta Joachim shows, when women's rights activists reframed women's rights as human rights, it enabled them to address gender-based violence through a much sounder set of norms prohibiting bodily integrity rights, whereas previously a weaker regime had been established for women.Footnote 50 Charli Carpenter's work on humanitarian disarmament campaigns shows how campaigners’ political success depended on linking claims about new weapons to specific arguments in humanitarian law rather than human rights or UN Charter law.Footnote 51 Ian Hurd shows how questions about the legality of humanitarian intervention depend on which parts of international law one invokes; and elsewhere shows a recursive relationship between practices (including how we speak of rules) and the rules themselves – a postulation made as well by scholars of ‘norm evasion’ and ‘norm decline’.Footnote 52
Indeed, whether and how actors mention rules and laws in discursive situations to which they are designed to apply is not only an indicator of norm strength but can be used as a tactic designed to contest norms or justify norm violations. Alexander Cooley, Ron Diebert, and Patrick Merloe show how powerful illiberal actors use such strategies to contest human rights norms.Footnote 53 Betcy Jose shows how efforts to recast a practice as outside the bounds of existing norms (or even as new ‘norms’ themselves) involve framing contests about the applicability of international law, with human rights groups often pushing back.Footnote 54 If debates about certain policies are largely debates about which laws matter, then survey questions that assume a certain law applies may bias respondents in favour of actors on that side of the debate, as well as constructing understandings of what type of ‘issue’ a policy problem is – exerting an agenda-setting effect.Footnote 55
Third, in questions that do mention international law and signal which set of laws matter, survey prefaces may explain to the respondent (correctly or incorrectly) what international law says: that is, what is prescribed or proscribed by the law. When this occurs, the survey not only primes a respondent but also potentially educates (or miseducates) the respondent about the law itself. Much norm socialisation work is centred on disseminating the relevant law to citizens and policymakers, ‘teaching’ them norms:Footnote 56 in the laws of war, for example, a considerable portion of the ICRC's work involved ‘dissemination’ of the law.Footnote 57 These efforts matter, alongside others, because norms are theorised to diffuse through various such teaching moments, and an informed public is seen as a mechanism to encourage norm adoption/compliance.Footnote 58 Thus any descriptions of what the law says in a survey question have a potential norm-diffusing effect to an American public uneducated in international law – or miseffect, if communicated incorrectly.
Fourth, and perhaps most problematically for absolute prohibition norms such as rules against torture, survey questions sometimes prime respondents to describe under what conditions they would support violating an international norm/law. This can occur even when the question omits international law entirely. Being asked to imagine when it is permissible to violate a norm can affect understandings of the substance of the norm, because as Vaughn Shannon explains in his seminal piece on the social psychology of norm violation, the content of international legal norms includes not simply prescriptions/proscriptions but also parameters: conditions under which they are meant to apply or not.Footnote 59 In other words, international laws vary by how many ‘exceptions to the rule’ they may have: the resort to force is legal under some conditions; torture, genocide and rape are prohibited in every situation.Footnote 60 By prompting respondents to decide for themselves when it is legally acceptable to engage in a universally prohibited behaviour, the pollster is implicitly inviting respondents to contest the parameters of the norm itself. This discursive strategy can undermine intersubjective agreement about the absolute nature of the norms in question, implying interpretive loopholes that don't exist, and suggesting that it is up to everybody to decide what a norm means, rather than teaching them what the law says. Such speech acts can contribute to cycles of norm decline by signalling to respondents that norm parameters are weaker than indicated by codified international law.Footnote 61
We do not argue that survey questions inevitably fall into these discursive traps. Nor do we argue that the purpose of scientific surveys should be to advocate for a particular position in a normative debate. Since existing international laws and norms are relevant social facts, it should be possible to communicate them as such to respondents when measuring opinions on policies that violate those norms.Footnote 62
Consider a hypothetical survey question preface that simply informs rather than primes the respondent about the relevant international norms. Such a question would factually acknowledge the substance and parameters of international norm, and then ask the respondent to evaluate government policy against those treaty obligations. For example: ‘Many countries have signed a treaty that states “the civilian population shall not be the object of attack” in war. However, some US officials believe it is acceptable to attack civilians if they believe this would save the lives of US troops. Would you approve or disapprove of the US government deliberately carrying out an air strike against a civilian target such as a hospital or a school?’ Such a question apprises the respondent of two relevant facts – what the law says and what the government might do – and invites him/her to express levels of opposition or support towards that behaviour, with full knowledge of its illegality.
What happens in actual surveys on important human rights issues in the national security context? We analysed an original dataset of 188 survey questions on two highly salient national security policies in both Bush and Obama administrations – ‘enhanced interrogation’ and ‘targeted killing’ – each engaging legal debates and pushback from the human rights community. Our analysis, detailed below, shows surveys on these issues rarely engage international law, and even more rarely mention international human rights law or provide accurate information on relevant international human rights norms. Nor do they invite the public to weigh in on the legal debate between the government and the human rights community. Instead, we find it is the frameworks and assumptions of government elites – those most likely to want to bend or break international norms – that are reproduced in the structure of survey questions. Pollsters either omit or discursively counter the advocacy frames of the global human rights movement – precisely those frames that IR scholars have found are crucial in socialising political actors into adherence to international norms.
As our analysis will show, the most common type of questions in our dataset were actually those that did not mention international law at all, while still priming respondents to imagine situations in which they would support a policy that would clearly violate such an unnamed law. These mechanisms can also work together in unexpected, and seemingly opposite, ways. For example, we identified questions that: (1) mentioned international law; (2) signalled the proper legal framework; (3) accurately explained the content and obligations of the law, while still (4) priming the respondent to consider ‘violating’ or even ‘change’ the law.
In the rest of this article we illustrate this general finding with specific examples through a discussion of the two issues we studied: ‘enhanced interrogations’ and ‘targeted killing’. We first outline the methodology we used for collecting and analysing these survey questions, the way in which we coded them with an eye towards potential norm weakening effects, and then explain how this process took place in the context of these two cases. We conclude with a call for further research on the socialisation effects of public opinion polls.
Human rights vs national security: Examining the discursive power of polls
To explore discursive patterns in national security polls empirically, and develop hypotheses about their impact on respondent understandings of international norms, we analysed the discursive properties of public opinion poll questions across a large dataset of questions related to the torture of detainees (or what foreign policy elites under the Bush administration referred to as ‘enhanced interrogations’) and extrajudicial execution of suspected terrorists (or what foreign policy elites under the Obama administration referred to as ‘targeted killings’). We chose these two cases because both illustrate the tension between national security and human rights, and have been salient enough in the public over recent years to attract significant attention by pollsters.
These two cases are similar and different in ways that help us explore the plausibility and generalisability of our hypotheses. First, they both involve the potential violation of non-derogable prohibition norms – international human rights laws considered universally binding even in national emergencies.Footnote 63 Both cases also involve a similar ‘threat’: (political violence against civilian populations by non-state actors), and therefore involve issues where government elites and the public have a similar propensity to want to bend the rulesFootnote 64 – precisely those kinds of cases where human rights law is needed. Third, in each case, human rights NGOs opposed the policy on legal grounds; in each case, the US government sought to redefine the rules to fit its policy – what Zoltan Buzas calls ‘norm evasion’.Footnote 65 Together these cases both provide useful ground to explore the hypothesis that pollsters may frame questions in ways that privilege one interpretation of the law over another.
At the same time, these cases are different in an important way: they spanned two different administrations, one Republican and one Democratic – enabling us to show that any similar patterns in the discursive structure of polls are not simply driven by political winds. Moreover, these human rights norms, while substantively similar, differ in their relationship to other areas of international law: while torture is prohibited in both human rights law and the laws of war, extrajudicial killing is primarily a peacetime offense, with killing without trial permitted in war under certain conditions. These differences allowed us to examine how far our findings about the discursive properties of polls might be expected to be robust across not only political contexts but also normative contexts.
We collected our public opinion questions from the Roper Center for Public Opinion Research's iPOLL database. We gathered a total of 188 distinct public opinion poll questions, with 137 pertaining to torture, and 51 pertaining to drones. We supplemented the iPOLL findings on targeted killing and drones with 14 additional survey questions from Sarah Kreps's and Micah Zenko's separate databases in order to arrive at our total.Footnote 66 Since our study focuses specifically on US foreign policy and its domestic audiences, we only included public opinion questions that were disseminated in the US context. In order to ensure that we collected as many questions on our two cases as possible, we varied our search terms in the database. These search terms included neutral words like ‘drones’ and ‘unmanned aerial vehicles’, but also more politically loaded terms central foreign policy debates, such as ‘enhanced interrogation’, ‘targeted killing’, ‘torture’, and ‘extrajudicial execution’. We temporally bounded our search terms to no earlier than 2001, in order to focus analytical attention on the post-9/11 ‘War on Terror’. We compiled the resulting data in Excel according to source. Since our unit of analysis is the question itself, in cases of progressive polling within a single survey, we separated its various components into distinct questions to more closely analyse wordings.
In order to analyse this data, we utilised a process of iterative qualitative coding. We first used a grounded theory approach to identify salient themes regarding how the questions addressed international law in each issue area.Footnote 67 We found striking patterns in both cases. In the torture case, prefaces to survey questions almost never informed respondents about status of torture in international law. Instead the structure of these surveys invited Americans to join a dialogue about both the proscriptions in the anti-torture norm (that is, what counts as torture) and its parameters (that is, under what conditions one might justifiably torture, even if torture is generally wrong).
Similarly in the targeted killing case, we observed that many questions about targeted killing fail to point respondents to international law at all. When they did invoke international law, they tended to favour the government's legal interpretation. As with the torture case, the questions ask respondents to comment on the ‘parameters’ of the relevant norm – that is, inviting them to say when it might be applied or bent. In some cases, respondents were invited to appraise the law itself, rather than the government policies in question. In no cases did we observe survey questions accurately explaining the relevant law, objectively describing US behaviour and inviting participants to assess US behaviour in light of its treaty obligations.
To ensure the reliability of our analysis, we uploaded our data into the qualitative data analysis software NVivo.Footnote 68 We created a codebook based on the four mechanisms observed in our initial grounded theory analysis. With these guidelines, each discrete survey question was coded for: (1) whether the question mentioned or alluded to international law or legal concepts at all; (2) whether the question explained the content of the law;Footnote 69 (3) whether the question mentioned or alluded to human rights law or international humanitarian law; and (4) whether the question asks respondents to decide for themselves if the law applies, what the law means, or whether it would be justifiable to violate the law.
We discuss each case in turn below to illustrate the discursive properties of these surveys in each, and to illuminate the four mechanisms in action. Although the cases show some important differences as well as similarities – in particular the ways in which human rights law was skirted in each – we show an importantly commonality to both: that assumptions favouring the government political narratives were embedded in the framing of the survey questions, in ways likely to run counter to norm socialisation efforts by the human rights NGOs.
Indeed, we found no instances where respondents were primed for compliance with the law. Out of 188 total questions on torture and drones, roughly 63 per cent of the questions invite respondents to signal when and how the US government might torture, and how and when it might arbitrarily kill suspected criminals without trial – what we call ‘priming for violation’.Footnote 70 In addition, where international law is invoked at all in these data, it is was most often invoked in order to ask respondents to evaluate the law itself, rather than US political behaviour. We detail these patterns in the sections below.
‘When and how may we torture?’ ‘Enhanced interrogations’ in the US (2001–08)
Torture is a universally prohibited practice. It is illegal in peacetime as well as wartime, and applies to all individuals regardless of their nationality, location, or role as warfighters or civilians. Because torture is a criminal act considered jus cogens, it is subject to universal jurisdiction and supersedes other treaties. Nonetheless, in the early 2000s, a powerful democratic superpower engaged in acts of torture and attempted to justify them as somehow outside the scope of these legal prohibitions. We show here how public opinion polls on torture implicitly lent legitimacy to this policy, while failing to include the country's legal obligations in their measurement of the political debate, and instead inviting citizens to re-evaluate those obligations rather than the actions of their government.
Torture in international law
The ban on torture appears first in the 1948 Universal Declaration of Human Rights (UDHR), which states ‘no one shall be subjected to torture or to cruel, inhuman or degrading treatment or punishment’.Footnote 71 The Geneva Conventions of 1949, as well as the 1977 Additional Protocols, strictly ban torture in several Articles, including Article 3 common to all four Geneva Conventions. Article 7 of the 1964 International Covenant on Civil and Political Rights (ICCPR) strengthened the ban in human rights law. One hundred and fifty-three countries, including the US as of 1992, have ratified the ICCPR. The 1984 Convention Against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment (CAT) refined the definition of torture to:
any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as obtaining from him, or a third person, information or a confession, punishing him for an act he or a third person has committed or is suspected of having committed, or intimidating or coercing him or a third person, or for any reason based on discrimination of any kind, when such pain or suffering is inflicted by or at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity. It does not include pain or suffering arising only from, inherent in, or incidental to, lawful sanctions.Footnote 72
CAT has 83 signatories and 165 parties, including the US, which ratified as of 1994. The ban on torture is non-derogable regardless of national emergencies, including terrorism and armed insurrections. In addition to this robust network of international legal prohibitions, many domestic legal systems prohibit torture as well. The US Constitution guarantees fundamental rights of detained persons under the Fourth, Fifth, Eighth, and Fourteenth Amendments.
Uses and justification for torture by the Bush administration (2001–08)
Shortly after 9/11, President George W. Bush issued a Military Commissions Order that sanctioned the indefinite detention of individuals captured in the war against terrorists.Footnote 73 While this Commission stipulated fair treatment to detainees, the early rhetoric about the ‘gloves coming off’ and working the ‘dark side’ hinted at illegal measures that would come to light only with the Abu Ghraib Prison scandal.Footnote 74 Indeed, the Central Intelligence Agency's (CIA) yet-unsanctioned detainment and interrogation programme had begun days after 9/11.
Throughout this period, US officials sought to frame their acts as something other than torture, while also arguing that the relevant conventions did not apply. When confronted by a reporter about Abu Ghraib, Defense Secretary Donald Rumsfeld stated: ‘I'm not going to address the torture word.’Footnote 75 The detention of Abu-Zubaydah sparked the systematisation of what the CIA called ‘Enhanced Interrogation Techniques’.Footnote 76 In an internal memo, the White House Office of Legal Counsel considered whether the actions of the CIA could be considered torture. On the administration's definition of torture, it states:
Physical pain amounting to torture must be equivalent in intensity to the pain accompanying serious physical injury, such as organ failure, impairment of bodily function, or even death.Footnote 77
Part II of the memo specifically addressed the CAT's definition, finding the list of purposes for which torture is prohibited (including extracting information) as ‘illustrative rather than exhaustive’; it recommended CAT be interpreted as covering only the most ‘egregious’ and ‘extreme’ acts, and ‘other acts of cruel, inhuman or degrading treatment or punishment’ were not banned, but only discouraged.Footnote 78
Amid these exchanges of memos, suspected al-Qaeda fighters were treated as being beyond the protections of the Geneva Conventions, since they were deemed ‘illegal combatants’ rather than prisoners of war.Footnote 79 The 2014 declassified version of the Senate Select Committee on Intelligence's study of the CIA detainment and interrogation programme details and confirms the specific actions undertaken at CIA black site interrogations.Footnote 80 Among the key findings is that the CIA's actions were supposedly more ‘brutal and far worse than the CIA represented to policymakers’.Footnote 81 Such actions, some public knowledge before the report's release, include waterboarding, ‘walling’, rectal feeding for non-medical reasons, locking detainees in small boxes, extreme sleep deprivation, threats of rape and death to detainees’ female relatives, forced nudity, and exposure to extreme cold or heat.
Public opinion polling and survey experiments on torture
In the context of this emerging policy, the American public was exposed to frequent polling – and poll results published in the mass media – on American attitudes towards torture. The Roper iPOLL database contains roughly two hundred instances of polling on ‘torture / enhanced interrogation’ (many repetitive) between the years 2001–11. Americans were polled on torture on average once every three weeks during this period. The polls themselves reached 85,903 Americans; the poll results, often concluding that ‘Many Americans Support Torture’ impacted a large US readership through widespread dissemination in the media.Footnote 82
We examined 137 unique survey questions on torture during the Bush era from the iPOLL database. Our initial search terms were ‘torture’ and ‘enhanced interrogation’ in order to cast a wide net. Through processes of iterative coding, we found striking patterns: prefaces to survey questions almost never informed respondents about status of torture in international law. Instead the structure of these surveys invited Americans to join a dialogue about both the substance of the anti-torture norm (that is, what counts as torture) and its parameters (that is, under what conditions one might justifiably torture, even if torture is generally wrong), a situation we coded as ‘priming for violation’.
In the polls that we identified as ‘priming for violation’, questions specify and even describe techniques such as waterboarding, and then ask respondents (in various ways) if they cross the threshold of constituting torture. For example, a 2011 ABC News/Washington Post poll asks respondents if ‘what American soldiers did to prisoners at the Abu Ghraib Prison in Baghdad amounts to torture’, or if it was ‘abuse, but not torture?’ A follow up question asks the respondents ‘What about physical abuse that falls short of torture? Do you think it's acceptable to physically abuse but not torture people suspected of terrorism in some cases, or do you think the use of physical abuse is never acceptable?’ Taking an even vaguer tack, a 2011 Farleigh Dickenson poll tells its respondents that ‘some people call the harsh interrogation methods enhanced interrogation techniques, while others call them torture’. Respondents are then asked which term is ‘closer to [their] view’. Some surveys examine which acts a respondent ‘considers a form of torture’, listing practices such as ‘forcing a prisoner to stay awake for up to 180 hours’, ‘forcing a prisoner to take an ice water bath’, and ‘threatening to sexually abuse a prisoner's mother’.
In the majority of cases, these questions do not even mention international law, much less point to the definition of torture under international treaties; international law was mentioned only 34 times, and explained only 16 times. Instead, many polls invite respondents to speculate on what torture means. By inquiring into what a respondent ‘counts’ as torture, the survey questions imply the definition of torture is subjective rather than codified in law. These questions mirror and unquestioningly adopt a similar logic to the Bush administration – namely, that the definition of torture is up for debate, and the tactics employed by the US can be interpreted as something other than and less than torture.
Other questions probe respondents to think about whether or not their support for torture is dependent upon the circumstances. Many of these questions assume the efficacy of torture.Footnote 83 Some are one shot questions: a well-circulated Pew Research Institute asks ‘do you think the use of torture against suspected terrorists in order to gain important information can often be justified, sometimes be justified, rarely be justified, or never be justified?’ Others prime in specific ways: Gallup/CNN/USA Today asks ‘Would you be willing or not willing to have the US government torture suspected terrorists if they may know details about future terrorist attacks against the US?’ Other questions utilise progressive scenarios. In one extreme example from a 2009 Fox News Corporation poll, respondents are prompted to consider their opinion on torture in increasingly dire situations. The primary prompt reads: ‘Do you support or oppose allowing the government to use any means necessary, including physical torture, to obtain information from prisoners that would protect the United States from future terrorist attacks?’ The scenario is conditioned progressively by the following questions:
If Osama bin Laden were captured and refused to cooperate, would you favor or oppose using torture to obtain information from him about members of Al Qaeda and other terrorist activity?
What if innocent lives could be saved, then would you support or oppose allowing the government to use physical torture to obtain information that would protect the U.S. (United States) from a terrorist attack?
If there were a possibility that a member of your own family could be saved, then would you favor or oppose allowing the government to use physical torture to obtain information from terrorist prisoners?
If theories about everyday legal socialisation are accurate, respondents and consumers of these scenario-based questions may be taking away impressions (or perhaps legitimising pre-existing beliefs) that torture is justifiable in national security emergencies – a view seriously at odds with codified international law and jurisprudence. It has been hypothesised that the ‘Jack Bauer effect’ had real and lasting consequences on public understandings of torture's efficacy and morality.Footnote 84 The popularity of Jack Bauer is evidenced by various references in post-9/11 speeches to the character's tactics as embodying the ‘ideal model’ for real counterterrorism policy.Footnote 85 The scenario portrayed in the Fox News poll questions is a particularly dramatic example of how surveys can blend fiction and non-fiction. As noted earlier, the consumption of media polls may have similar and additive effects to the consumption of (and saturation with) pop culture products like 24. Footnote 86 By fictitiously holding the respondent's family hostage, individual survey takers are essentially being primed to contest an absolute prohibition norm.
Although most questions ignore international law, we noted a number of questions that do explicitly mention and explain relevant legal obligations in the preface to the question. However, upon closer examination, international law is not typically mentioned so respondents can evaluate US policy against it. Rather, the surveys invite respondents to evaluate the laws themselves, and whether or not US should feel justified in breaking them. There are twenty co-occurrences of these codes – ‘ILaw Mentioned’ and ‘Priming for violation’ – in our torture dataset. With international law being mentioned only 34 times, this means respondents were ‘invited’ to break the law nearly 60 per cent of the time that it was the subject of conversation regarding torture. For example, a PIPA/Knowledge Networks poll gets specific about international law, stating:
The US and most countries in the world have signed a number of conventions – that is, treaties that create international laws – that prohibit certain methods for trying to get information from detainees. Here are some of these prohibitions. For each one please select whether you favor having it or if you think it is too restrictive.
Another form of question in this category may inquire directly into an individual's view of where ‘terrorists’ stand under international law. For example, a PIPA/BBC World Services poll states that ‘most countries have agreed to rules that prohibit torturing prisoners’, and asks whether such rules should be applied to suspected terrorists or whether ‘terrorists pose such an extreme threat that governments should now be allowed to use some degree of torture if it may gain information that saves innocent lives’. In short, these questions as well invite the respondent to not only question whether the substance of international law is unfairly onerous, but also the parameters: whether the rules might be appropriately bent in certain cases, and if so, which.
These public opinion polls on torture did not occur in a vacuum; the questions and their results got taken up and amplified to larger audiences in media stories from a variety of sources. Indeed, many of these polls originated from journalistic sources, which then published the polls for wider audiences through their platforms. For example, a 2014 Washington Post/ABC News Poll that we coded as ‘priming for violation’ was published in a Washington Post news story titled ‘New Poll Finds Majority of Americans Think Torture was Justified after 9/11 Attacks’, as well as in a Washington Post blog entry titled ‘Let's Not Kid Ourselves: Most Americans are Fine with Torture, Even When You Call it “Torture”’.Footnote 87 The former Post story also quotes another poll from our dataset, a 2014 CBS News poll, which reports that while 70 per cent of respondents think waterboarding was torture, half still supported its usage.
In a similar and more recent example, a Reuters story titled ‘Exclusive: Most Americans Support Torture against Terror Suspect’, published a 2016 Reuters/Ipsos poll we coded as containing no mention of international law, and ‘priming for violation’.Footnote 88 These surveys were taken up in non-traditional news sources as well, such as FiveThirtyEight, which published a story titled ‘Americans Have Grown More Supportive of Torture’.Footnote 89 This story was based on two separate polls; a 2011 Associated Press and 2013 Pew poll. Even when survey top-lines are reported ‘objectively’, the publication of such results provides consumers with the impression that many of their fellow citizens think torture is unproblematic, and frames the question of torture as a legitimate topic of public debate.
Such a portrayal of international norms runs counter to the efforts of human rights organisations aiming to disseminate international legal understandings to the public and stakeholders during the same period. Rather than reinforcing the socialising efforts of NGOs by accurately describing the law, the subtext embedded in these questions dovetailed with the narrative of the Bush administration: that the new post-9/11 world requires extreme measures, and that these extreme measures might not necessarily constitute torture under the law; and that even if they do, it is an option for Americans to consider ignoring that law if they wish.
‘Who may decide when/who/how we may kill without trial?’ ‘Targeted killing’ in the US (2008–16)
As stated earlier, the status of extraterritorial drone strikes under international law is a more complicated case than torture in terms of absolute prohibition. Like torture, the arbitrary killing of criminal suspects without trial is considered a non-derogable human rights violation under human rights law, even in the case of a declared ‘public emergency’.Footnote 90 However, many exceptions to the ‘no killing’ rule exist under the parallel law of armed conflict. Thus, when the United States began hunting and executing suspected members of the al-Qaeda network during the later war on terror, it sought to treat these killings as subject to the laws of armed conflict rather than human rights law.Footnote 91 Although this framing was heavily disputed by international lawyers precisely because a state of armed conflict did not exist in many places where these killings occurred, public opinion polls in the US almost invariably echoed the government's framing of ‘targeted killings’ as both lawful and effective.
Extrajudicial execution of criminal suspects in international law
Execution of criminal suspects without prior due process is prohibited under human rights law.Footnote 92 By contrast, killing is permitted in wartime, so long as a state of armed conflict exists; so long as such killings are targeted at members of the combatant armed forces or those ‘directly participating in hostilities’ rather than at non-combatant civilians or captured fighters; and so long as attacks are subject to the principles of military necessity, distinction, and proportionality. Even in war, however, the execution of non-combatant prisoners of war without a trial is considered a grave breach of the Third Geneva Convention. The summary execution of captured enemy civilians is considered a grave breach of the Fourth Geneva Convention. Both of these prohibitions apply to internal conflicts as well as international ones.Footnote 93
Extrajudicial execution by the Obama administration (2008–16)
Although President Barack Obama vowed to end the Bush-era torture programme, within its first term the Obama administration executed far more criminal suspects without trial than its predecessor. While many killings were targeted at specific criminal suspects, others were aimed simply at military-age men who were judged to fit the profile of such a suspect. Targeting decisions for such a ‘signature strike’ required only that the person's ‘pattern of life activity’ is judged as aligning with that of a potential terrorist's, and did not require a place on the so-called ‘kill list’ or executive approval.Footnote 94 While the preponderance of recent attention on targeted killing and extrajudicial execution has been focused on drones, targeted killings can be carried out in numerous ways, including airstrikes from conventional weapons, gunshots, stabbing and poisoning. For example, the US special operations mission that killed Osama bin Laden in Pakistan took place using ground forces with small arms.
US legal justifications for these killings have consistently framed them in terms of law of armed conflict, rather than human rights law, claiming the former superseded the latter. For example, after the 2011 killing of US citizen Anwar al-Awlaki, who was also presumably protected under US constitutional law, the Department of Justice released a White Paper to justify the killing in terms of international law.Footnote 95 The paper argued that such a killing – as well as others like it against non-US citizens – is lawful if an informed, high-level official believes capturing the person is ‘infeasible’, and if the strike is ‘consistent’ with the ‘laws of war governing the use of force’.Footnote 96
Yet the US government's stance – that the 2001 Authorization to Use Military Force Against Terrorists (AUMF) permits the use of armed force wherever al-Qaeda and its associated forces are located – is highly disputed. Numerous international lawyers argue instead that international human rights law is better suited to counterterrorism operation outside of traditional battlefields, which they argue requires a ‘policing’ paradigm rather than a ‘war’ paradigm.Footnote 97 For example, the UN's Global Counterterrorism Strategy focuses on coalition-based law enforcement responses to terrorism.Footnote 98 A report by the UN Special Rapporteur on Extrajudicial Executions and Arbitrary Killing urges the US and other states engaged in international counterterrorism operations to respect both international humanitarian law as well as human rights law.Footnote 99 Amnesty International states that it is ‘gravely concerned’ that the US drone programme is ‘killing outside the bounds of law’ and conducting extrajudicial executions.Footnote 100 Similarly, Human Rights Watch wrote a joint letter with other major human rights litigation advocacy groups urging the Obama administration to subject the policy to both branches of the law.Footnote 101 The extent to which legal agreement exists on drones and targeted killing matters for our article, precisely because it is the parameters of this debate that are not being reflected in the survey questions. Public opinion polls on targeted killing have discursively appropriated the narrative of US political elites, rather than that of human rights NGOs, or any of the multiplicity of international actors that contest this ‘global war’ frame.
Public opinion polling on execution of terror suspects (2009–15)
According to Roper's iPOLL database, at least 42,282 Americans were subject to at least 37 distinct public opinion poll questions on the extrajudicial execution of terrorist suspects between 2006 and 2015. We found polls addressing these topics in Roper's iPOLL database using the search terms ‘drones’ and ‘killing terrorists’; pollsters do not use the terms ‘extrajudicial execution’, ‘assassination’, or ‘targeted killing’. We augmented the iPOLL search results with five additional questions from Kreps's 2014 dataset, and nine from a CFR article by Zenko, for a total of 51 questions.Footnote 102 As with the torture case, broad patterns regarding the invocation of international law and ‘invitations’ to break it emerge across the survey questions on drones and targeted killing.
First, as with the torture case, many questions on targeted killing fail to point respondents to international law at all, much less human rights law. Rather, many of the questions are focused on perceptions of military efficacy, the geography of strikes, or whether the DOD, as opposed to the CIA, should carry them out.Footnote 103 Out of 51 questions, only 6 mentioned international legal concepts at all. A Fox News poll showcases a typical question: ‘What do you think is the better way for the United States to deal with terrorists? Kill them with drones, [or] capture and interrogate them, and then try them in a military court?’ An oft-repeated Pew Research Poll question asks respondents if they ‘approve or disapprove of the United States conducting missile strikes from pilotless aircraft called drones to target extremists in countries such as Pakistan, Yemen and Somalia?’ A Farleigh Dickinson University Public Mind poll first asks its respondents if they generally approve of the US military conducting such killings, while its follow up asks whether they generally approve of the CIA executing the same task. In cases such as the above, the pollsters do not mention international law or gesture to legal concepts relevant to these policy dilemmas.
Elsewhere, pollsters allude to legal concepts – but only those found in the law of armed conflict (such as distinction and proportionality) rather than international human rights law. Indeed, only one question in our entire dataset referred to the possibility of capturing and trying suspected terrorists; war law concepts appeared in three questions. In this respect, the surveys incorporate the highly contested assumption that the law of armed conflict is applicable outside hot battlefields – the US government's position in the debate, rather than that of many INGOs and foreign governments. For example, although the concept of the ‘innocent civilian’ only applies in armed conflict, an Economist/YouGov poll asks: ‘Do you approve or disapprove of the Obama administration using drones to kill high-level terrorism suspects overseas, even if innocent civilians may also be killed?’ A Pew poll also primes respondents to think of civilian loss of life: ‘How concerned are you, if at all, about whether U.S. drone strikes … Endanger the lives of innocent civilians?’ By offering no context on the legal debate, or whether it is right to treat terror suspects as soldiers rather than suspected criminals, these questions are framed in a way that invites respondents to assume it is legitimate to target the terror suspects, if only collateral damage can be minimised, even in a non-conflict zone.
Second, many of the questions build on the assumption that targeted killing itself is unproblematic by asking respondents about the use of drones rather than the execution of suspects without trial. The primer of a Pew Research poll states that the United States military has ‘made increasing use of unmanned aircraft called “drones” to launch aerial attacks in Afghanistan, Iraq and elsewhere’, before asking if the respondents ‘think the increased use of drones by the military is a good thing, or a bad thing’. Questions tend to operate in similar contextual vacuums, such as a Quinnipiac University poll: ‘Do you support or oppose the US using unmanned aircraft or “drones” to carry out missile attacks against suspected terrorists in foreign countries?’ Some pollsters get more specific about the nature of the weapons platform in the primer, before asking about levels of support. An NBC News/Wall Street Journal poll tells its readers that ‘the United States has been targeting and killing suspected members of Al Qaeda and other terrorists in countries such as Pakistan, Yemen, and other countries’, and that ‘many of these killings have been conducted using unmanned aircraft that are controlled remotely, also known as drones’.
Framing the debate as one about specific technology rather than about whether targeted killing is legitimate has implications. While drones might well be capable of being used in legal ways, the ethics of extrajudicial execution do not depend on the weapon used. Conflating the two debates, and focusing on the supposed precision of drones – a factor that Walsh demonstrates as being highly salient to American constituentsFootnote 104 – draws attention away from and may even induce a positive association with the policy it facilitates. When coupled with largely unfalsifiable claims about ‘targeting and killing suspected members of al-Qaeda’, the questions create a narrative of efficacy and precision around drones that distracts a larger and more technical policy discussion on the legality of executing criminal suspects. Because individuals tend to weigh negative information more heavily than positive information when formulating decisions, the lack of sobering context on legality or actual efficacy may be inflating support for targeted killings.Footnote 105
Finally, surveys about ‘targeted killings’, like surveys about ‘enhanced interrogation’, ask respondents to comment on the ‘parameters’ of the relevant norm – that is, inviting them to say when it might be applied or bent, although extrajudicial killing of suspects is a universally prohibited practice in human rights law. We found thirty instances of such priming, making up well over half of our drones dataset. Many of these questions ask Americans about their relative levels of approval of drone strikes in countries the US is not at war with, and contain no further information on the policy or its legal parameters, as in a 2013 Farleigh-Dickinson poll: ‘In general, do you approve or disapprove of the US military using drones to carry out attacks abroad on people and other targets deemed a threat to the United States?’
Such questions inquire about the appropriateness of a policy that violates a non-derogable human right law, with no alluding to the law regulating the practice. The nationality of the targeted individual stands out as the most salient parameter the pollsters identify, as they often distinguish between targeted killings against Americans (where due process concerns weigh heavier in the minds of respondents) and foreigners. A 2013 Reason-Rupe poll calls for respondents’ opinions on the constitutionality of the president of the United States ordering the ‘killing of American citizens who are suspected of being terrorists’. Other questions are more specific about the geographic location of the strike. A Gallup poll asks if the government should ‘launch airstrikes in the US against US citizens’ who are suspected terrorists. Some questions are progressive scenarios, meant to pinpoint which elements of targeting (including identity and geography) make Americans most (un)comfortable. A Fox News poll first informs the respondent that they will be read ‘a series of ways in which the United States might use deadly force against suspected terrorists’, and that they are to indicate their approval or disapproval of US uses of drone to do each of the following:
‘ … to kill a suspected terrorist in a foreign country if the suspect is a US citizen?’
‘ … to kill a suspected foreign terrorist on US soil?’
‘ … to kill a suspected terrorist who is a US citizen on US soil?’
As with the case of torture, and as noted by Kreps, the prompts as well as results from these polls were frequently echoed in news media.Footnote 106 For example, a 2013 Washington Post story titled ‘Most Americans Approve of Foreign Drone Strikes’ relies on two separate questions from a 2013 Pew survey.Footnote 107 We coded the first question in this story as not mentioning international law, while the second was coded as mentioning international law, but ignoring human rights law. A Washington Post blog entry titled ‘The American Public Loves Drones’, is based on three polls from our dataset: a 2013 Washington Post/ABC poll, a 2011 Washington Post/ABC poll, and a 2011 Pew poll.Footnote 108 A 2015 PBS News Hour report offered a similar headline, ‘Americans Approve of Drone Strikes on Terrorists, Poll Finds’, citing a 2015 Associated Press/GfK poll.Footnote 109 A contemporaneously published Al Jazeera America story ‘Poll Finds Strong Support For Drone Strikes Among Americans’, citing a 2015 Pew poll question.Footnote 110 Similarly to the torture questions, less traditional media outlets also published these polls. A headline from The Hill reads: ‘Majority of Americans Support US Drone Strikes, Survey Says’, citing the 2015 iteration of a Pew poll on drones.Footnote 111
This sampling of mass media stories that adopted these polls is meant to indicate that both the premise and results of the surveys were disseminated beyond the small population of respondents. As Kreps points out in her critical analysis of polling on drones in particular, ‘the public's enthusiasm has become part of the political narrative and shaped the boundaries of the conversation.’Footnote 112 The mass media take-up of these survey questions greatly expands their audiences and penetrates the US public discourse, possibly constituting a key linking mechanism between individual polls to wider political understandings. Because of this, we see value in empirically testing the hypothesis we have built in this article – that polls should be understood as conversational settings, and that the discursive nature of poll questions invites or encourages wider, weakened understandings of international human rights norms.
In sum, polling on targeted killing largely renders international law invisible, with very few questions even broaching it as a relevant consideration. When international law is included in a subject of conversation, human rights law is ignored, while the government's preferred legal ‘war frame’ is elevated with no explanation of obligations. This has the effect of levelling out very real, contested international legal debates over drones and targeted killing. Poll questions repeatedly ask respondents to assume the policy is legitimate and effective (a legal argument that can only be made from a war law frame), largely by focusing on weapon technology and the nationalities of those targeted. When considered in conjunction with IR literature on norm socialisation and communication theories on media effects, we suspect that the qualitative examples of priming we've presented may be potentially impactful to respondents’ understandings of the applicability of international norms to national security issues. In the drone case, this may be especially true in regards to attitudes towards due process for criminal suspects.
Discussion and conclusion
The cases vary slightly in terms of the narratives used by political elites in the US government to skirt international legal prohibitions, and also in terms of how ‘settled’ the international legal debates are on the practices. In the torture case, foreign policy elites acknowledged the anti-torture norm, which is heavily solidified and codified, but were particularly keen to reconstitute the definition of ‘torture’ and set their own acts as somehow outside the parameters of that prohibition. In the case of extrajudicial execution, foreign policy elites stressed constitutional law and international law of armed conflict, rather than international human rights law, in order to carve out exceptions to the universal norm against extrajudicial execution. Legal agreement on the US's current usage of drones is certainly less consolidated than in the torture case. But the US's position in the legal debate – that areas like Pakistan, Yemen, and Somalia are subject to the laws of war and can be subjected to extraterritorial strikes – is contested by a multiplicity of international actors. Both cases suggest that public opinion polls, far from simply measuring sentiment, should be viewed as conversational settings in which elite policy preferences can become embedded. A qualitative exploration of the survey questions wordings reveals that human rights legal frames contrary to the US government's position were largely excised from the debate in both cases.
Some might counter that there is value in measuring attitudes towards government policy frames, and perhaps even using the government's language in order to assess how individuals respond to mainstream discourse on foreign policy issues. However, such discourse does not float freely and is usually contested, particularly by human rights INGOs. Thus, a measurement that does not also incorporate the elite counter-messaging from actors such as large INGOs can implicitly reify one side of a normative debate. For that reason, it is problematic to assume that the findings from such studies are internally or externally valid as measures of some objective public opinion ‘out there’. We believe that polls should be understood by IR scholars as forms of communicative action, themselves shaping both concrete knowledge and political preferences, including about legally salient foreign policy issues – and the nature of these effects should be studied, and carefully controlled for in IR survey experiments.
We do not of course argue that these polls are a sole or even primary driver of public socialisation into international law. There are a multiplicity of ways that both elite and public understandings of human rights norms can change – whether it is through the INGO campaigns aimed at altering the perceived ‘appropriateness’ of a given government policy – like the use of landmines – or through a more historic evolution of the perceived strategic consequences of a given government policy – like leadership assassination.Footnote 113 Taking a social constructivist approach, we have hypothesised that polls could be another salient site of everyday socialisation, given their accessibility and prevalence in public discourses due to publication in the mass media, and we find support for this hypothesis in this systematic plausibility probe of the discursive properties of polls. For this reason, we suggest further empirical research is needed to measure their impacts on attitudes, something we think would be a useful contribution to debates over sources of normative change in the international arena.Footnote 114
Such a research agenda could have important implications for stakeholders and scholars. For example, if the hypotheses developed in this article about how much polls can affect norm socialisation are borne out after further empirical testing, it would have implications for the efforts of norm advocates, such as human rights NGOs. Rather than merely attempt to diffuse norms through mailings, reports, or press releases, a policy implication would be that human rights NGOs are missing an opportunity to educate respondents and frame debate by embedding their own narratives in public opinion polls.
Second, such a research agenda would yield both new avenues of inquiry and potential methodological insights for scholars of public and opinion and international law. We have established that the discursive structure of poll questions may predispose respondents towards certain understandings, but empirical research is needed to test the magnitude of this hypothesised effect. Certainly other factors besides surveys themselves may influence the waxing and waning of public opinion on national security issues, and further research is needed to unpack whether, how much and how lasting an effect survey participation has on citizens’ understandings of complex social issues. Carefully constructed survey experiments – constructed to avoid any ethical pitfalls resulting from discursive priming described here – could explore the impact of such frames not only on survey measurements but also on respondent learning. Scholars of international norm diffusion could also more closely test the extent to which understandings taken from the relatively small number of survey respondents nationwide travel to wider audiences through media effects or word of mouth, as some evidence suggests.Footnote 115
The insights here may offer not just fresh research questions but also methodological cautions to the burgeoning experimental survey literature on international law. For example we noted that Kreps's survey experiment, which critiques a polling establishment too quick to adopt elite US government narratives, also replicates some of the problems of the very polls she criticises.Footnote 116 Survey experimenters could pay more attention to measuring norm impacts while minimising the impact of their surveys on understandings of norms.
Finally, scholars of international norm and law diffusion may benefit from expanding their focus beyond the efforts of NGOs, IOs, and formal policy arenas, and more on the mundane, everyday sites where intersubjective understanding of IR get constructed – classrooms, survey experiments, simulations, media representations, and varieties of popular culture. Opinion surveys lie at the intersection of these types of political communication, constituting both a learning environment for participants and a set of persuasive tools for audiences. IR scholars should pay closer attention to ways in which they are used not just to measure, but to construct, international normative understandings.
Acknowledgements
We are grateful to the Department of Political Science at University of Massachusetts for a Summer Research Grant that seeded this project. Helpful feedback was provided by participants in the Conflict, Violence, and Security workshop and Flashapalooza at University of Massachusetts, at the International Studies Association Annual Conference in 2017, and by several anonymous reviewers. We are particularly grateful for comments by Ardeshir Pezeshk, Usmaan Farooqui, Jamie Rowen, and Timothy Pachirat.
Alexandria J. Nylen is a political science PhD Candidate at the University of Massachusetts, Amherst. Her research interests include international law, critical security studies, and transnational advocacy networks. She is currently writing a dissertation on the relationship between issue framing and transnational network connectivity, using advocacy against armed drones as a case study.
Charli Carpenter is a Professor in the Department of Political Science at University of Massachusetts, Amherst. She has published three books and numerous journal articles on international law and politics, raised two members of the American electorate, and consulted for the government, the United Nations, and multiple humanitarian and human rights NGOs. She is currently using experimental survey methods to study the politics of law, and sees this as ironic.
