This is a political convention. But needing our only son – and all of the cherished hostages – home, is not a political issue. It is a humanitarian issue.
– Jon Polin, Democratic National Convention in Chicago, Illinois, 21 August 2024
Introduction
On the morning of 7 October 2023, Palestinian militants invaded southern Israel from the Gaza Strip. In their attack coordinated across air, land, and sea, Hamas fighters streamed into Israel, killing more than 1,200 people. Hamas and its partners also kidnapped 251 people from over twenty countries – including twelve American citizens – and brought them into captivity in Gaza.Footnote 1 Among the hundreds taken were two young Israeli-American men: 23-year-old Hersh Goldberg-Polin, kidnapped from the Nova Music Festival, and 21-year-old Omer Neutra, a tank commander in the Israeli Defense Forces.
In the year after October 7, Hersh’s and Omer’s parents became highly recognisable advocates for bringing their sons and all hostages home. Both met regularly with senior US officials, including the president of the United States, to share their sons’ plights. Both were given prime-time speaking slots at presidential nominating conventions in the summer of 2024, where they were met with rapt attention and standing ovations. In the convention halls in Milwaukee and Chicago, thousands of Republican and Democratic delegates and attendees chanted ‘Bring them home! Bring them home!’ for minutes on end.
That both convention halls erupted in shared chants might suggest that hostage recovery represents a rare US policy issue with bipartisan support. Instead, the two conventions put in stark relief the variation in how the American public thinks about the purpose and form of hostage recovery. At the Democrats’ convention, ‘Bring them home!’ signalled support for hostage recovery through negotiations. Hersh’s parents, Jon and Rachel, were outspoken proponents for negotiation, pushing Israeli Prime Minister Benjamin Netanyahu to agree to life-saving concessions to end the hostage crisis.Footnote 2 The Polins thanked those people still pushing for a negotiated ceasefire and hostage deal: ‘You keep us breathing in a world without air.’ At the Republicans’ convention, conversely, ‘Bring them home!’ signalled the desire to use military force to punish hostage takers and rescue captives. Proponents of military rescue may find more in common with remarks from Ronen and Orna Neutra, who emphasised their anger, not anguish, to the crowd: ‘Where is the outrage?’ the Neutras asked. ‘This was and is an attack on Americans.’Footnote 3
The presidential conventions illustrate the political dynamics of hostage crises. While hostage taking presents a problem for foreign policy, hostage recovery presents a problem for domestic politics. As they make difficult foreign policy decisions, leaders are routinely criticised – and sometimes punished – for their action or inaction in bringing captives home. Existing research has explored how governments respond to hostage crises,Footnote 4 and how the public reacts in turn.Footnote 5 Bringing these dynamics together, recent work has explored how variation across hostage crises, including the hostages’ circumstances of capture and the cost of recovery options, affect public support for recovery.Footnote 6 However, we know less about which members of the public support hostage recovery policies – and why.
In this paper, we explore the individual-level variation in support for bringing hostages home. We argue that support for hostage recovery reflects the intersection of two different – and sometimes contradictory – attitudes: individuals’ general views about foreign policy and their partisanship. Drawing on two large, national samples of the American public, we test how Americans’ partisanship and broader foreign policy attitudes predict their willingness to support a range of hostage recovery activities, including negotiations, ransom payments, prisoner exchanges, and military rescue missions.
We find that three key factors shape Americans’ support for hostage recovery. First, compared to isolationists, foreign policy internationalists – individuals who want their government to have an active role in world affairs – are much more likely to support all forms of hostage recovery. However, variation in Americans’ type of internationalism – whether militant or cooperative – also matters in shaping support for various recovery options. Second, while broad, bipartisan majorities support recovery in general, there are crucial partisan differences regarding recovery options. Compared to Republicans, Democrats are much more likely to support life-saving concessions and less likely to support military rescue missions – a gap exacerbated when the policy is proposed by a president of the opposite party. Finally, we show that respondents with children are more likely than those without children to support some recovery options – a pattern we link to the emotional pleas often made by the parents of captives abroad.
In explaining the public’s view of hostage recovery, our work makes important contributions to scholarship and policy. First, kidnapping is a persistent and pervasive threat to civilians worldwide. Despite kidnapping’s centrality to inter- and intra-state conflict, few scholars have explored its consequences for international security.Footnote 7 Our work thus joins a small but growing body of literature that explores the dynamics of hostage taking and recovery, contributing to broader understandings of armed groups’ pattern of political violence and implications for international security.
Second, our work contributes to a robust literature on how individual-level attitudes shape public opinion on key issues in international security. This body of research has examined public preferences regarding the use of force,Footnote 8 decision making during international crises,Footnote 9 the role of moral values in foreign policy attitudes,Footnote 10 public reactions to cybersecurity threats,Footnote 11 and nuclear weapons.Footnote 12 Our research applies insights from these works to the domain of hostage recovery, a uniquely personal and emotionally charged foreign policy issue that combines questions of state responsibility, use of force, international cooperation, and humanitarian concern. In doing so, we extend the literature on public opinion and international security by highlighting how partisan, ideological, and emotional factors shape support for specific government responses to transnational hostage crises. We also contribute to the literature by increasing our understanding of public expectations about the role of government in protecting them from harm.Footnote 13
Finally, our work offers novel real-world implications for the complex politics of hostage recovery. Existing research suggests that hostage takers are more likely to target democracies because elected leaders face added pressure to respond.Footnote 14 Recent work shows that policymakers are responsive to public pressure on hostage recovery.Footnote 15 This article shows how that pressure is far from uniform and instead varies with other political concerns among the target population – crucial ramifications for democratically elected leaders facing hostage crises.
This paper proceeds as follows. First, we provide a brief background on how the US government brings hostages home, showing that different recovery options vary widely in both process and outcome. Second, we outline the individual-level factors that we expect to shape support for hostage recovery efforts. Next, we explain our research design for examining public opinion on hostage recovery and offer findings from large surveys of the American public. Drawing on these findings, we then discuss the implications for the politics and policy of hostage recovery.
Background: Hostage recovery options
Hundreds of Americans are kidnapped abroad every year. According to the US Hostage Recovery Fusion Cell, the inter-agency government body tasked with hostage recovery operations, there is a ‘persistent threat’ of Americans being kidnapped abroad.Footnote 16 The government’s ‘commitment to the safe and rapid recovery of U.S. nationals held hostage overseas’ depends on executive agencies, which ‘coordinate their efforts to leverage all instruments of national power to safely recover’ American hostages.Footnote 17 What does that mean in practice? In other words, what can a government do when one of their citizens is taken hostage?
Kidnapping – the most common form of hostage taking – functions as coercion: perpetrators take and hold a human captive in order to compel behaviour change from a third-party target.Footnote 18 Kidnappers present an explicit threat: if you do not do what we want, the hostage will suffer continued captivity, pain, and even death. Hostage takers thus offer their targets a grim choice: make concessions and win the hostage’s freedom, or do not, and suffer accordingly. Kidnapping thus clearly illustrates what scholars mean when they say that coercion requires the ‘cooperation’ of the target;Footnote 19 kidnappers give the target a painful choice for how to respond.Footnote 20
Thus, some options for hostage recovery rely centrally on cooperating with an adversary. Target governments can respond to hostage taking by negotiating with kidnappers for hostages’ release. For the US government, such negotiation is express policy. In 2015, the Obama White House released Presidential Policy Directive 30 (PPD-30) on ‘Hostage Recovery Activities’. PPD-30 emphasised the government’s willingness to communicate with anyone – including hostage takers – if it could help bring hostages home. For instance, PPD-30 allows the government to ‘assist private efforts to communicate with hostage-takers’. Moreover, the US government ‘may itself communicate with hostage-takers, their intermediaries, interested governments, and local communities to attempt to secure the safe recovery of the hostage’.Footnote 21 After the October 7 hostage takings by Hamas, for instance, the US government helped facilitate negotiations between the Israeli government and Hamas by bringing in government intermediaries from Egypt and Qatar – countries able to communicate effectively with parties on both sides.
Before or after negotiations, government targets can also accede to hostage takers’ demands. More often than not, kidnappers demand ransom payments, though they may also demand prisoner exchanges or other policy concessions;Footnote 22 such demands are often satisfied. Indeed, most hostages are freed after the target makes concessions to the hostage taker. Despite the conventional wisdom that the United States has a ‘no concessions’ policy and the re-assertion of such a policy as recently as 2015,Footnote 23 the US government regularly makes or condones concessions made by others. Technically, the prohibition on concessions applies only to government-made ransom payments to designated foreign terrorist organisations – a small fraction of kidnapping cases.Footnote 24 In other words, the oft-repeated US ‘no concessions’ policy would not and has not stopped countless American families, companies, or even the US government itself from making concessions to win hostages’ release.Footnote 25
Conversely, a target government may refuse to cooperate with hostage-taking coercion. In the wake of a kidnapping, the government may elect not to intervene. The government may worry that hostage recovery would undermine broader foreign policy aims; that extraordinary assistance to a small number of individuals would invite moral hazard; or that hostage recovery falls outside the purview of government responsibility. If the government does not act, the hostage’s family or employer may work on their own to recover a captive, sometimes relying on the extensive private market of security consultants and kidnap and ransom insurance to do so.Footnote 26
Finally, the government may decide to respond – but not cooperate. Target governments may threaten hostage takers with severe consequences, like sanctions or military force, or implore other parties to do so.Footnote 27 Target governments can engage force, using military rescue missions to recover captive citizens.Footnote 28 In the 1980s, the US military established the United States Special Operations Command (USSOCOM) headquarters and Joint Special Operations Command (JSOC) to coordinate multiple branches of the military in joint operations overseas, including the hostage recovery missions by the Army’s Delta Force and the Navy’s SEALs. Since then, US Special Forces have engaged in hostage recovery missions, alone or in tandem with partner military forces. Such missions’ purpose is twofold: first, recover captives alive, and second, punish perpetrators. Achieving the first goal is extremely difficult, requiring accurate, up-to-date intelligence; effective tactics; and risk-mitigation plans. As such, military rescues are rare and seldom succeed.Footnote 29
Theory
Hostage takings are highly salient events in international security. Transnational kidnappings in particular attract outsized media coverage compared to other forms of violence.Footnote 30 Such salience has meaningful implications for public interest in hostage recovery and for American citizens’ awareness when a fellow citizen is kidnapped abroad. For example, according to a September 2014 NBC/Wall Street Journal poll, 94 per cent of Americans were aware of the kidnapping of journalist Jim Foley and his subsequent beheading by the Islamic State – the highest proportion of Americans aware of any news event polled in the prior five years.Footnote 31
Public attention – and desire to act – is paradoxically magnified by the fact that hostage takings often feature single victims. Hostage crises are subject to a ‘collapse of compassion’, an inverse relationship that postulates that as ‘the number of people in need of help increases, the degree of compassion people feel for them ironically tends to decrease’.Footnote 32 Studies have repeatedly shown that the public pays more attention to – and is more willing to help – a single, named individual than a larger group of nameless victims. The fewer the victims and more detailed their stories, the more attention audiences pay to their suffering.Footnote 33 Hostage advocacy campaigns rely on this heightened compassion for individual kidnap victims.
Beyond their extensive media coverage, kidnappings often feature prominently in what scholars call ‘soft news’ outlets, which have a broader reach and can inform even those members of the public with limited awareness of foreign policy issues.Footnote 34 Thus, although some foreign policy decisions are insulated from public opinion because of American disinterest,Footnote 35 the public is likely to be familiar with – and animated by – hostage recovery.
We thus begin from the assumptions that hostage taking is a particularly prominent international security issue, and that hostage recovery is a salient instance of a government’s foreign policy approach. One form of hostage recovery described above – hostage rescue missions – entails the use of military force abroad. Public opinion about hostage recovery may thus reflect broader foreign policy preferences, including opinion about the use of force abroad. To understand variation in public support for government policies to bring hostages home, we start with the factors that scholars identify as shaping American public opinion about foreign policy and the use of force. In this section, we outline some of the key takeaways from this literature and draw implications for understanding public opinion about hostage recovery.
Key factors associated with public opinion about foreign policy
First, scholars point to the internationalist/isolationist spectrum as an important disposition for understanding foreign policy opinions. This spectrum captures whether an individual wants their government to have an active role in world affairs or prefers their government to focus on problems at home. This disposition acts as a heuristic for individuals with little foreign policy knowledge or interest. It is an easy mental shortcut that individuals can rely on to make sense of specific foreign policy issues. If individuals have an internationalist disposition, they may feel more favourably towards extroverted foreign policy options. Conversely, if individuals have a more isolationist disposition, they may oppose foreign policy actions that divert the government’s attentions beyond a country’s border.
Whether this disposition is fixed or movable is sometimes debated in the literature. Researchers show that Americans’ level of internationalism increases when they are co-partisans with the president.Footnote 36 Such a shift stems from individuals’ assuming that a co-partisan president would be more likely to pursue a foreign policy in line with their beliefs and values than would an out-party president. Recent work also shows that there are a number of individual-level determinants of internationalism/isolationism, ranging from domestic perceptions of the country’s economic outlook to perceptions of the international security environment.Footnote 37 Various moral foundations and ideology have been linked to internationalism.Footnote 38
In addition to sorting on an internationalist/isolationist scale, scholars further distinguish subtypes of internationalism along two main continua. First, individuals may be more inclined to support a cooperative approach to internationalism or a militant type of internationalism. In general, militant internationalism is associated with more support of the use of force, while cooperative internationalism is associated with increasing relational ties with other countries through trade or international organisations.Footnote 39 The other dimension along which scholars have sought to characterise people’s internationalist beliefs is the unilateral/multilateral dimension. Some individuals may prefer an extroverted foreign policy that is primarily pursued unilaterally, and others may prefer a foreign policy that engages other countries.Footnote 40
We expect internationalism to affect how individuals think about options for recovering hostages during a crisis. First, because internationalists tend to support a wide range of foreign policy activities, foreign policy internationalists may be more supportive of the policies and practices that lead to individuals travelling to foreign countries in the first place.Footnote 41 Given that soldiers, aid workers, or employees of international firms can work in dangerous places, they are thus regular victims of kidnapping.Footnote 42 If the public opposes sending individuals to dangerous foreign countries in the first place, they may be more likely to oppose expensive or risky options for bringing them home. Additionally, internationalists are more supportive than isolationists of the actions required for recovery: international negotiation, bargaining, and the use of force. This does not mean that some isolationists would not support hostage recovery given the risk to a co-national, but rather, that on average, internationalists should demonstrate higher and more consistent support for hostage recovery.
Beyond disagreement between internationalists and isolationists, we expect that different types of internationalists would disagree regarding hostage recovery. Hostage recovery can take the form of more militant action such as a rescue attempt as well as more cooperative action like concessions and negotiation. Such actions could be taken unilaterally or in collaboration with other states. Specifically, we expect that cooperative internationalists may be more supportive of actions to bring hostages home that include negotiated concessions to hostage takers. Militant internationalists, conversely, would be more likely to oppose such a response. The reverse could be true for rescue operations: militant internationalists would likely be more supportive of using armed force to rescue hostages, while cooperative internationalists should be more likely to oppose it. Similarly, we would expect multilateralists to be more supportive of coordinated action with allies or the countries where the hostage is held, while unilateralists would find such coordination unnecessary or constraining.Footnote 43
Negotiating with – and making concessions to – hostage takers is especially interesting to study, as it represents a hard test for cooperation. As scholars of international relations have repeatedly shown, cooperating with others is difficult in general; it is especially difficult to cooperate in the absence of trust. Targets should have little reason to trust hostage takers. Thus, we offer a hard case for the effects of cooperative internationalism on support for particular foreign policy outcomes. This discussion leads us to our first set of hypotheses.
H1: Compared to internationalists, isolationists will be more likely to oppose hostage recovery.
H2a: Higher levels of militarism will be associated with more support for military rescue to recover hostages.
H2b: Higher levels of militarism will be associated with less support for negotiations, ransom payment, and prisoner exchanges.
H3a: Higher levels of cooperative internationalism will be associated with less support for military rescue to recover hostages.
H3b: Higher levels of cooperative internationalism will be associated with more support for negotiations, ransom payment, and prisoner exchanges.
The preceding discussion about internationalism as a key predictor of foreign policy attitudes and the potential for it to shape attitudes towards hostage recovery should also take into account the role of partisanship, as partisanship and ideology may map onto different dimensions of internationalism. Democrats and liberals tend to rely on different moral foundations from Republicans and conservatives.Footnote 44 Moral foundations theory posits that individuals’ political orientations are grounded in distinct sets of moral intuitions.Footnote 45 Liberals, for example, tend to prioritise the individualising foundations of care/harm and fairness/cheating, which emphasise empathy, protection from suffering, and justice. Conservatives, by contrast, are more likely to endorse binding foundations such as loyalty/betrayal, authority/subversion, and sanctity/degradation, which emphasise group cohesion, tradition, and moral purity. These different moral emphases may lead liberals to support foreign policies consistent with cooperative internationalism, while conservatives may favour more assertive and punitive approaches aligned with militant internationalism.Footnote 46
Such moral foundations have abiding real-world implications for foreign policy preferences. For instance, liberals and Democrats tend to prioritise humanitarian outcomes, development, and diplomacy.Footnote 47 Moreover, Democrats emphasise the government’s responsibility to protect its citizens from harm. These general preferences may connect to preferences for hostage recovery: a strong desire to bring captives home, and a preference for using diplomatic means to do so.
Republicans, conversely, tend to prefer a different set of foreign policy responses, which connect to hostage recovery. For instance, Republicans tend to be sceptical of negotiations and compromise.Footnote 48 Refusing to negotiate for hostages, in particular, has been a traditional Republican talking point. Indeed, it was a Republican president, Ronald Reagan, who most famously articulated the United States’ purported ‘no concessions’ policy. In response to a 1985 hostage crisis, Reagan affirmed:
America will never make concessions to terrorists – to do so would only invite more terrorism – nor will we ask nor pressure any other government to do so. Once we head down that path there would be no end to it, no end to the suffering of innocent people, no end to the bloody ransom all civilized nations must pay.Footnote 49
Despite the fact that Reagan and every subsequent US president has engaged in hostage negotiations,Footnote 50the ‘no concessions’ rhetoric has endured.Footnote 51
Beyond the Republican association with the ‘no concessions’ rhetoric, other ideological tendencies may contribute to partisan policy preferences for hostage recovery. First, Republicans tend to be more fiscally conservative than Democrats, and thus may be more reluctant to support expensive policy options.Footnote 52 Second, Republicans are more likely to prefer ‘tough-on-crime’ stances, such as mandatory minimum sentences.Footnote 53 Such tough-on-crime stances could contribute to a refusal to reward – and preference to punish – bad behaviour.
We expect these differences to be reflected in responses to hostage crises. We hypothesise that Democrats and liberals will be more supportive of a broader spectrum of recovery policies that include negotiation and concessions, while Republicans and conservatives will be more supportive than Democrats of military operations – and less supportive of ransom payments – to bring hostages home.Footnote 54
H4a: Compared to Democrats, Republicans will be more supportive of military rescue to recover hostages.
H4b: Compared to Democrats, Republicans will be less supportive of negotiations, ransom payment, and prisoner exchanges to recover hostages.
Additionally, we expect that like with internationalism, individuals may be more supportive of hostage recovery if their co-partisan is president. The public may trust their co-partisan to pursue hostage recovery in a way that reflects their ideology. Moreover, individuals may want their co-partisan to gain political capital for bringing hostages home, while being more ambivalent about such a political victory for an out-party president.
H5: Individuals will be more supportive of recovery options when their co-partisan pursues them than when an out-party president pursues them.
Lastly, the hypotheses discussed above come from theory of public opinion about hostage recovery under the general umbrella of international security crises and public opinion about foreign policy. However, issue-specific patterns could also explain variation in support for hostage recovery. To this end, we draw on the examples laid out in the beginning of this manuscript. Often, when an American is held captive overseas, the person’s parents, children, and other family members advocate publicly for their release.Footnote 55 These public appeals frequently invoke the emotional hardship of knowing a loved one is in danger abroad. Such appeals are particularly effective due to the ‘collapse of compassion’ described above – the public is drawn to the plight of individual victims and their grieving families.
We argue that the appeal to emotions and the perspective of parents generates empathy among Americans, particularly those who themselves have children. Those who can more closely identify with the experience of having children – and perhaps imagine the pain of having one’s own child in danger – may have more willingness to support government options to bring hostages home. This theory draws on a number of perspective-taking and perspective-getting studies in political science and psychology that show the effectiveness of such techniques for generating support for myriad foreign policy issues such as accepting refugees, open migration, and support for foreign and humanitarian aid.Footnote 56
H6: Compared to individuals without children, parents will be more supportive of hostage recovery.
Research design
To test our hypotheses, we conducted two surveys with large, national samples: the first in November 2020 and the second in January 2025. The 2020 survey included 1,895 respondents recruited by Dynata, and the 2025 survey included 2,677 respondents recruited by Prolific. Dynata uses quota-based sampling to approximate national demographic distributions, and Prolific similarly allows researchers to recruit using demographic quotas from its online panel. In both waves, quotas were set on age, gender, and race to achieve broad national coverage. In our surveys, we presented respondents with a fictional, hypothetical scenario in which an American was captured abroad.Footnote 57 We randomised a range of pertinent details of the scenario, varying the captive’s gender and profession; where the captive was taken; and whether the captors were ‘rebels’ or ‘terrorists’.Footnote 58
We then offered respondents a range of possible policy options for recovering the captive. In the 2020 survey, respondents could indicate their support for negotiating in response to ransom demands; paying ransom demands (randomised at $100,000, $1 million, and $10 million); and launching a rescue mission (randomised at requiring 10, 100, or 1,000 soldiers).Footnote 59 In 2025, we added an option for exchanging one, two, or ten prisoners to win the American’s release.Footnote 60
Additionally, the timing of our 2025 survey uniquely allowed us to test H5 about the role of partisanship and the president in public opinion about hostage recovery. By fielding our survey the week of 13 January 2025, we took advantage of the transition period between US presidential administrations. In that survey, we varied whether the recovery options were proposed by ‘President Biden’ or ‘President-elect Trump’ At the time that participants took the survey – the last week of the Biden administration, and mere days before Trump took office – we considered this variation a plausible design to explore how respondents reacted to one or the other proposing hostage recovery options.
In all surveys, we asked respondents whether they supported the recovery option (approve strongly; approve somewhat; neither; disapprove somewhat; disapprove strongly). Finally, respondents were asked a battery of demographic and other questions, identifying their interest in politics, political party, internationalism, ideology, employment, income level, race, gender, age, parental status, and education level.
Findings
Our findings showcase various correlates of support for hostage recovery options. We present results from the 2020 and 2025 studies in the paper and the results from the 2016 pilot study in the Appendix. As noted above, in our 2020 survey, we included a number of experimental variations in the hostage crisis vignette in addition to measuring several demographic background characteristics of the respondents in our sample. We include control variables for treatment assignment as well as important demographic characteristics of the respondent. We report coefficients and standard errors for all demographic variables. We do not report the coefficients on the treatment variables in the results tables, because they serve as vignette-level controls rather than quantities of substantive interest; our focus in this section is on the individual respondent-level predictors of support for recovery.
The evidence presented in Table 1 shows that several demographic variables are significantly correlated with support for hostage recovery consistent with our hypotheses. Across all of the recovery options, internationalists are much more likely to support recovery compared to isolationists (H1).Footnote 61 All other demographic variables have inconsistent patterns across the recovery options. Most notable in this respect is political party. Political party is measured as a five-category variable. In the regression model, Party: Democrat is the baseline and thus does not appear in the table. The results show a relatively bipartisan consensus when it comes to supporting rescue missions and negotiating with hostage takers (H4a not supported), with Republicans and Democrats supporting both at equal rates. As we hypothesised, Paying ransoms is where Republicans and Democrats diverge, with Republicans being significantly less likely than Democrats to support paying ransoms (H4b supported). Finally, we show that individuals who have children are more supportive of all recovery options, but only the coefficient on the rescue outcome is significant at traditional levels (H6).
Table 1. 2020 correlations with demographic variables.
*** Note: Standard errors in parentheses. p < 0.01, ** p < 0.05, * p < 0.1. The OLS models include, but do not report, indicators for all experimental treatments including deservingness, size of rescue mission, size of ransom, profession of hostage, order of recovery options presented, gender of the kidnapped American, whether the captor was a rebel or terrorist group, and the location of the kidnapping. The models also include, but do not report, the respondents’ state of residence. Summary statistics tables are included in the Appendix.
Next, we turn to the 2025 results. In Table 2, we display the results from a regression model that most closely replicates the 2020 study. We see that H1 is once again partially supported with Isolationism negatively correlated with support for rescue, but not significantly related to any of the other recovery options.Footnote 62 H4a and H4b are also supported as Republicans are more supportive of military rescue but less supportive of negotiating or paying ransoms to bring the hostage home. Finally, H6 is also partially supported. While individuals with children are not different from those without children in their support of rescue, negotiations, or prisoner exchanges, they are much more supportive of paying ransoms to bring hostages home.
Table 2. 2025 correlations with demographic variables.
*** Note: Standard errors in parentheses. p < 0.01, ** p < 0.05, * p < 0.1. The OLS models include, but do not report, indicators for all experimental treatments including deservingness, size of rescue mission, size of ransom, profession of hostage, order of recovery options presented, gender of the kidnapped American, whether the captor was a rebel or terrorist group, and the location of the kidnapping. The models also include, but do not report, the respondents’ state of residence. Summary statistics tables are included in the Appendix.
To further test our more specific hypotheses about different types of internationalism and the role of co-partisanship in shaping views on hostage recovery, we report the results from two additional models in Tables 3 and 5. First, in Table 3 we disaggregate ‘internationalism’ into separate measures of the militant and cooperative versions of internationalism. We find strong support for the hypotheses associated with H2 and H3. Militarism is significantly associated with support for each recovery option in the direction predicted by H2a and H2b.Footnote 63 For a military rescue, individuals with higher levels of militant internationalism are much more supportive than those with lower levels of militant internationalism. For the outcomes more associated with concessions – negotiation, paying ransom, and exchanging prisoners – the inverse relationship is present. Individuals with high levels of militant internationalism are more opposed to these recovery options than individuals with low levels of militant internationalism.
Table 3. Results from regression models with disaggregated internationalism types, 2025 survey.
*** Note: Standard errors in parentheses. p < 0.01, ** p < 0.05, * p < 0.1. The OLS models include, but do not report, indicators for all experimental treatments including deservingness, size of rescue mission, size of ransom, profession of hostage, order of recovery options presented, gender of the kidnapped American, whether the captor was a rebel or terrorist group, and the location of the kidnapping. The models also include, but do not report, the respondents’ state of residence.
With respect to those who align more with cooperative internationalism, we show largely the reverse pattern.Footnote 64 H3a is not supported as there is no significant relationship between our measure of cooperative internationalism and support for a military rescue. However, for the other three outcomes, the pattern is in the expected direction supporting H3b: individuals that support a more cooperative style of internationalism are more supportive of actions that require cooperation with the perpetrators.
We also dig further into the mechanism explaining the differences between militant internationalists’ and cooperative internationalists’ approaches to hostage recovery. Recall that the moral foundations underpinning militant internationalists’ outlook on foreign policy could lead them to seek out more assertive and punitive actions, which could lead them to have more support than cooperative internationalists for recovery that aligns with these moral values. In Table 4 we use two additional questions from our 2025 survey as our dependent variables to understand how respondents think about punishing hostage takers as a goal of recovery. Punish 1 is a variable ranging from 1 to 4 that asks respondents how much they agree or disagree with the statement: ‘Bringing hostages home is more important than punishing hostage takers.’ The other outcome also asks about whether punishment ought to be a goal of recovery; we measure it similarly, via agreement with the following statement: ‘Hostage recovery should be used to punish hostage takers.’ As the results show, respondents with higher levels of militarism are more likely to agree with Punish 2, while respondents with higher levels of cooperative internationalism are more likely to agree with Punish 1. Together these results demonstrate the value of distinguishing between different types of internationalism. The simple isolationist/internationalist divide masks important variation in the approach to foreign policy that Americans prefer their government to take. These differences in approaches map on to the different recovery options available to the government and goals those options can achieve when an American is held captive abroad.
Table 4. Results from regression models on punish hostage takers outcomes with disaggregated internationalism types, 2025 survey.
*** Note: Standard errors in parentheses. p < 0.01, ** p < 0.05, * p < 0.1. The OLS models include, but do not report, indicators for all experimental treatments including deservingness, size of rescue mission, size of ransom, profession of hostage, order of recovery options presented, gender of the kidnapped American, whether the captor was a rebel or terrorist group, and the location of the kidnapping. The models also include, but do not report, the respondents’ state of residence.
Finally, in Table 5, we report coefficients from regression models that interact respondents’ partisanship with the partisanship of the US president. Recall that in our 2025 survey, we included a treatment that randomly assigned respondents to read that either President Biden or President-elect Trump had proposed the recovery option. We interact an indicator variable for treatment assignment with binary indicators for each response option for party identification: Democrat, Republican, or Independent.Footnote 65 If our hypothesis is supported, we should see that Democrats are more supportive of recovery options when President Biden proposes them and Republicans are more supportive of recovery options when President-elect Trump proposes them.
Table 5. Results from regression models with interaction between partisanship and Trump treatment indicator, 2025 survey.
*** Note: Standard errors in parentheses. p < 0.01, ** p < 0.05, * p < 0.1. The OLS models include, but do not report, indicators for all experimental treatments. The models also include, but do not report, the respondents’ state of residence.
Because interaction models are difficult to interpret in regression tables, we display the treatment effect of the president treatment by respondents’ party identification for each recovery option in Figure 1. We restrict Figure 1 to Democrats and Republicans because H5 explicitly hypothesises co-partisanship effects, requiring focusing on respondents who hold a partisan identification that maps onto the party of the sitting president. Independents and respondents reporting no partisan affiliation are heterogeneous in their political orientations and cannot be classified as co-partisans. For Democrats, the results support H5: for each recovery option, Democrats are less supportive of hostage recovery if the vignette they received said President-elect Trump proposed the recovery plan relative to President Biden. We also see evidence that Republicans respond to the president treatment. For all outcomes except for prisoner exchange, Republicans are more likely to support recovery if they read about President-elect Trump proposing the recovery plan, relative to President Biden.Footnote 66
Figure 1. Effect of Trump as president treatment on support for recovery options by respondent partisanship.
Discussion
The results of our surveys provide repeated, robust evidence for the correlates of support for hostage recovery. What do our findings imply for the politics of hostage recovery more broadly? In this section, we discuss the implications of our surveys for partisan responses to in- and out-party presidents and for the politics of hostage recovery in countries beyond the United States.
First, our results help explain the political dynamics at play when different American presidents have tried to bring hostages home. In particular, Figure 1 provides insight to the real-world asymmetry in public opinion for hostage deals made by Democratic and Republican presidents. In recent years, US presidents of both political parties have engaged in negotiations and made public, controversial concessions to bring hostages home.Footnote 67 Presidents Trump and Biden have engaged in negotiations and prisoner swaps with autocratic states including Russia, China, Iran, Venezuela, and North Korea to recover Americans held captive abroad. However, compared to President Trump, President Biden received disproportionate criticism for making hostage deals.
As our study has shown, Republicans are significantly less supportive of negotiating with – and making concessions to – hostage takers. Such Republican opposition is exacerbated when told that an actual Democratic president, Joe Biden, pursued those cooperative options. Thus, it is no surprise that we have seen precisely these dynamics play out in the real world. When former President Joe Biden made deals to recover American hostages, Republicans forcefully criticised him for doing so. None were more forceful in their opposition than President Trump himself, who has called Biden’s hostage deals ‘bad’, ‘unfair’, ‘extortion’ and ‘an embarrassment’.Footnote 68 Republicans dislike negotiating with or making concessions to hostage takers in general, and they like those policies even less when pursued by an out-party president.Footnote 69
When a Republican president pursues cooperative hostage recovery options, however, such criticism is muted. There are three possible reasons for this more muted response. First, Republicans’ opposition to negotiations and concessions is mitigated when told that such policies were pursued by their in-party president; they are less likely to criticise Trump for making exactly the same kinds of policy decisions as Biden. Moreover, Democrats are more supportive of cooperative hostage recovery policies in general. While Democrats’ support for such policies is reduced significantly when proposed by Trump, Democrats remain more favourable than not to hostage deals. Finally, Democrats could resist criticising hostage recovery deals for fear of being called unpatriotic.
A final point worth raising is the persistent and bipartisan support for hostage rescue missions – the public’s clearly preferred option for hostage recovery. Like the existence of a ‘no concessions’ policy, we consider the support for military rescue to stem from popular misconceptions about US hostage recovery – perhaps one driven more by Hollywood than Washington. Despite rescue missions’ public popularity, they are rare events with very high failure rates.Footnote 70 Given the difficult-to-achieve conditions that are required for success, rescue missions cannot be considered the optimal option to bring most hostages safely home. Nevertheless, leaders have – in spite of their better judgement – succumbed to public pressure and pursued rash rescue missions. None is more notorious than ‘Operation Eagle Claw’, President Jimmy Carter’s ill-advised attempt to recover fifty-two Americans held in the US Embassy in Tehran. Carter’s advisors knew that the mission was unlikely to succeed, calling it ‘self-defeating and probably suicidal’.Footnote 71 The allure of a possible victory convinced the president to ignore the factors that security personnel consider to be crucial prerequisites to launch a rescue mission: the ability to minimise ‘risk to force (the guys going in) and risk to mission (the hostage)’.Footnote 72
Our study explores support for hostage recovery among a population of American respondents. Extrapolating beyond the United States, to what extent might these same dynamics travel to other countries, and in particular, to other Western, democratic targets of hostage-taking violence?
On the one hand, we would expect some of the factors shaping individual-level preferences in the United States to similarly explain public opinion in other Western democratic countries. Individuals’ broad preferences regarding internationalism and isolationism, as well as militant and cooperative internationalism, should be expected to correlate with hostage recovery preferences beyond the United States. Individuals’ political ideology on a liberal/conservative spectrum should also travel. After October 7, for example, liberal Israelis were much more supportive of recovering hostages through prisoner exchanges with Hamas, while conservative populations and politicians opposed making concessions for hostage recovery.
On the other hand, however, there is reason to believe that the baseline of public support for hostage recovery would be higher in other targeted states. In countries including France, Israel, and Canada, hostage crises spur major public protests,Footnote 73 maintaining public pressure until hostages are brought safely home. Public opinion regarding hostage recovery might correlate with variation in a sense of ‘national solidarity’; where such solidarity is higher, the public will be more likely to demand hostages’ safe return.Footnote 74 Moreover, we expect that the public in countries with more robust social welfare systems would have a greater sense of collective responsibility for citizens’ safety, whereas countries with a deeper emphasis on individual responsibility would have less support for government assistance.Footnote 75 Finally, the public may be more supportive of military responses for a more physically proximate hostage-taking threat, or regarding an adversary with a history of kidnapping co-nationals. All of these factors could raise the floor for support and decrease the degree to which partisan and ideological tendencies influence opposition to recovery. Future work should investigate these dynamics, exploring hostage policy in comparative context.
Conclusion
In this article, we explore variation in public opinion about hostage recovery. Hostage crises are salient foreign policy events, and public reactions play a central role in how elected governments approach the resolution of said crises. We theorise that the public views these crises through a similar lens as other foreign policy issues, particularly international security and issues of the use of force. We test several hypotheses drawn from the literature on foreign policy public opinion as well as examining the role of emotions and empathy in shaping parents’ support for various recovery options in light of the emotional pleas victims’ family members often make.
The evidence supports our hypotheses to a strong degree. We use surveys fielded to national samples of Americans over multiple presidential administrations to test the hypotheses. The data reveal that internationalism – including subtypes of militarism and cooperative internationalism – as well as partisanship, the partisan identity of the president, and having children shapes the way respondents view hostage crises. Like other foreign policy issue areas, internationalists, and especially militant internationalists, are more likely to support hostage recovery via military rescue, while internationalists, and especially cooperative internationalists, are more likely to support bringing hostages home through the use of concessions like ransom payments or prisoner exchanges. Partisanship also shapes views on hostage recovery; though Democrats and Republicans support rescue missions and negotiations at equal rates, Republicans are much less likely to support paying ransoms to bring hostages home. Moreover, whether the president is a co-partisan also fundamentally alters the respondents’ support for various hostage recovery options: both Democrats and Republicans are more likely to support recovery options if they were proposed by a president of their own party, and they are more likely to oppose recovery options proposed by an out-party president. Together, these results provide insight into the politics of hostage recovery and the factors that inform Americans’ views on the resolution of hostage crises.
Our findings resonate with the reception that the Israeli-American hostage families received during the 2024 presidential nominating conventions. Americans are indeed broadly supportive of bringing hostages home, but the method of recovery, purpose of recovery, and leader managing the recovery matter a great deal. With hostages’ parents at the podium, Democrats seemed more likely to support a solution achieved through negotiated concessions, whereas Republicans cheered for a forceful response.
These findings have vital implications for the politics of hostage recovery – an empirical phenomenon in which public opinion matters a great deal to real-world outcomes. As past work has demonstrated, hostage takers are more likely to target democracies than non-democracies, precisely because of the relationship between policymakers and the public. Citizens in democracies, with access to free press and intolerant of limits to civil liberties, are more likely to pressure their governments to respond when a fellow citizen has been taken hostage.Footnote 76 Recent work has demonstrated the existence of such policymaker responsiveness in both Israel and the United States; democratically elected leaders care about public opinion on foreign policy, and they adapt their behaviour accordingly, including on hostage recovery.Footnote 77 Our findings suggest that target governments will have a harder time garnering public support for some hostage recovery activities, like ransom payments, than for others, like military rescue missions. They also suggest that enterprising perpetrators might compound their hostage-taking damage through provoking controversies in partisan politics, exacerbating the domestic political dilemma of hostage recovery.
This article demonstrates how some individual characteristics among the public shape support for hostage recovery. Beyond our findings, we expect that several other factors at the heart of hostage crises would shape the politics of hostage recovery – each of which provides a promising avenue for future research. First, our work focuses on foreign policy attitudes, partisanship, and parenthood as relevant factors shaping public opinion about hostage recovery. But we anticipate that other individual traits, like gender or national attachment, would correlate with hostage recovery preferences. Second, future work should explore how the public responds to different forms of hostage-taking violence. This paper explores public opinion about one type of hostage taking: kidnapping by non-state armed groups. However, in recent years, a new form of hostage taking has risen in prominence, as autocratic states like Russia, China, and Iran increasingly use their criminal justice systems to hold Americans hostage.Footnote 78 As successive US administrations have negotiated with adversarial states to win prisoners’ release, future work should explore public opinion about this new and shifting form of hostage taking.
Third, future work should investigate how perpetrators’ identities affect the politics of hostage recovery. Prior work has shown that – despite the US government’s well-known mantra that they ‘won’t negotiate with terrorists’ – the American public is not any less supportive of making concessions to ‘terrorists’ than other non-state kidnappers.Footnote 79 But future work could explore how specific, named perpetrators might illicit a different response. Existing work suggests that experimental designs relying on abstract or hypothetical actors yield different results than those based on real people.Footnote 80 We don’t know, for instance, whether respondents had any real-world hostage crisis or recovery dynamics in mind when answering survey questions, which could shape their beliefs about the risks of recovery options or chances of success. Scholars could thus test the dynamics of public opinion during a real, ongoing hostage crisis, to explore how findings change between hypothetical and real-world scenarios.
Finally, future work might explore how variation across particular hostages and hostage crises shapes public opinion about bringing captives home. Prior work has shown that American support for hostage recovery is shaped by the circumstances of capture: the American public is less supportive of hostage recovery when the public believes that hostages were personally responsible for putting themselves in danger.Footnote 81 But future work should continue to explore other aspects of hostage ‘deservingness’ – including how characteristics like race, gender, and profession shape public support across cases. A large body of work identifies a ‘missing white woman syndrome’ in domestic abduction cases: young, white, female victims of abductions receive far more media coverage than their male or non-white peers.Footnote 82 Scholars of public opinion and foreign policy should explore how such traits translate to international violence. Such future work would provide insights on the domestic politics of international security, showing how hostage-taking violence reverberates through target states.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/eis.2026.10051.
Acknowledgements
The authors would like to thank their survey participants for contributing to this research. For their insightful comments on prior drafts, the authors would like to thank Elizabeth Acorn, Karl Gustafsson, Martin Naunov, Dick Samuels, Jessica Stanton, Rachel Whitlark; workshop participants in Berlin and Evanston; and two anonymous reviewers and the editors of the European Journal of International Security.
Funding statement
Funding for this project was generously provided by the Humanities, Social Sciences, and Arts Grant from the Office of the Provost at Northwestern University.
Competing interests
The authors have no conflicts of interest to declare.
Danielle Gilbert (https://www.danigilbert.com/bio–cv.html) is an assistant professor in the Department of Political Science and a faculty associate at the Institute for Policy Research at Northwestern University.
Lauren Prather (http://www.laurenprather.org) is an associate professor in the School of Global Policy and Strategy at the University of California, San Diego.
Open access
