The Things They Carried: Generational Effects of the Vietnam War on Elite Opinion Abstract Do foreign policy elites who shared in formative political experiences also share similar views

Do foreign policy elites who shared in formative political experiences also share similar views on subsequent policy issues? Proponents of a generation effect suggest they do, but this argument overlooks two facts: that not everyone experiences major historical events in the same way, and that different experiences might give rise to quite different policy views. Here we investigate the impact of the Vietnam War on elite opinion about foreign policy during the following two decades using elite surveys conducted by the Foreign Policy Leadership Project (FPLP) from 1976 through 1996, assessing their susceptibility to what has been called the Vietnam Syndrome. Not surprisingly, we find that age and military service influenced elite opinion about the Vietnam War. More importantly, we find that different trajectories of opinion about the Vietnam War influenced later views about a wide range of foreign policy issues during the Cold War, even after controlling for party identification and ideology. However, there is little evidence that these effects persisted after the end of the Cold War, even on matters like civil war intervention where the experience of the Vietnam War was arguably still relevant. The "Vietnam Syndrome" that we find is restricted to the Cold War rivalry. Jonathan M. DiCicco, Canisius College Benjamin O. Fordham, Binghamton University Authors’ note: A previous version of this article was presented at the 2016 annual meeting of the International Studies Association. The authors thank Jessie Rumsey, Emily Meierding, participants in the World Politics Workshop at Binghamton University, and the editors and anonymous referees at ISQ. We are responsible for any remaining errors.


Introduction
In a Wall Street Journal op-ed published a half-century after the United States' escalation of the Vietnam War, former Secretary of Defense Chuck Hagel and retired US Senator Bob Kerrey asserted, "The years of the Vietnam War defined a generation and influenced and changed America. It profoundly affected the nation's culture and politics. No corner of society was untouched. The war framed and informed a generation and forced the questioning of institutional thinking, particularly in foreign policy." 1 Hagel and Kerrey's assessment of the Vietnam War's impact is not unusual, but it is telling.
Like others before them, the former Senators from Nebraska characterize the Vietnam War as a generation-defining event. And, like others before them, they suggest that the war affected Americans' foreign policy attitudes, shaking their confidence in the establishment and its faith in US military might. Though Hagel and Kerrey do not use the expression, others have put a label on this lingering shadow of doubt about the efficacy of US foreign interventionism: the "Vietnam Syndrome." Hagel and Kerrey share more in common than their Cornhusker roots and views on the Vietnam War's impact. Both are decorated veterans of the war, and both are foreign policy elites of the first order. 2 That these two men would share strong views on the war's effects is unsurprising, despite their partisan differences. But did the Vietnam War inform an entire generation, as they suggest? Previous studies of elite opinion found little evidence of generation effects on foreign policy attitudes (Holsti and Rosenau 1990;Holsti 1996). Our analysis of survey data from the Foreign Policy Leadership Project (FPLP), conducted between 1976 and 1996, helps to account for the disjuncture between Hagel and Kerrey's conventional wisdom and the underwhelming evidence for generational effects in Holsti and Rosenau's analyses. Parsing the effects of distinct Vietnam War experiences allows us to identify patterns of elite attitudes toward the war and toward subsequent military engagements-patterns that reflect differences between generations, but also within the "Vietnam generation" itself. Different wartime experiences and contrasting conclusions about whether the war was a mistake gave rise to divergent views about American foreign policy in the years that followed. In this sense, our findings contribute to recent scholarship on generations, political socialization, and politics (e.g., Bartels and Jackman 2014;Jennings, Stoker, and Bowers 2009;Johnson and Dawes 2016;Neundorf and Niemi 2014;Schuman and Corning 2012).
That our study captures distinct foreign policy attitudes among people who experienced the Vietnam War in different ways comports with recent studies of wars' effects on the political attitudes of those who lived through those wars (e.g., Bellows and Miguel 2009;Blattman 2009;Grossman, Manekin, and Miodownik 2015;Hong and Kang 2017). Consistent with such studies, we hypothesize and find that wartime military service is associated with distinctive attitudes toward war-the Vietnam War, and especially in our study, prospective wars-and that these attitudes tend to persist. To gauge the persistence of elite attitudes shaped by war experiences, we use trends in attitudes toward the Vietnam War to predict elite opinion concerning prospective uses of US power abroad, both during and after the Cold War. Doing so allows us to evaluate the claim of a "Vietnam Syndrome" among the people most likely to affect foreign policy.
Not surprisingly, the elite survey data reveal that many respondents changed from support to opposition as the Vietnam War dragged on. For many in this group, war-related trauma came not from combat but from the shattering of prior assumptions about the war and about the United States' role and efficacy in world affairs. Such a change of heart is arguably more consequential for elites closely involved in foreign policy than it would be in the general public. Survey respondents who shared this experience of changing their minds perhaps are the core group of skeptics who suffered most acutely from a "Vietnam Syndrome." Members of this group proved reluctant to embrace interventionism and tended toward "dovish" opinions in post-Vietnam surveys.
However, survey data from the 1990s show that this dovish tendency among those who came to oppose the Vietnam War evaporated after the Soviet Union's collapse. Similarly, the "hawkish" attitudes of Vietnam War supporters that prevailed during the Cold War were attenuated in the 1990s. As we argue below, these changes cannot be attributed solely to political ideology and partisanship-rather, they reflect the importance of international context, and in particular, overriding concerns about the Cold War. The disruption of earlier, significant patterns suggests that the hawk-dove dichotomy depended on the Cold War frame-and that among US elites, war-influenced attitudes were impermanent.
In sum, it may be tempting to speak of a "Vietnam generation" among US foreign policy elites, but it is more accurate to state that there are multiple generation effects shaped by individual experience and the international context. Within that age cohort, patterns emerge: those who served in the military and those who changed their mind about the war exhibit distinct attitudes concerning the war itself and subsequent projections of US military power. And to the extent that their divergent attitudes can be characterized as hawkish or dovish, these patterns break down after the Cold War, indicating that fears of a permanent Vietnam Syndrome were misplaced. To say, then, that the war "framed and informed a generation" may afford too much influence to the war itself and not enough to the constraining context of the Cold War. The Vietnam War's effects on the foreign policy attitudes of elites-members of a "Vietnam generation" and otherwise-were patterned and persistent, but did not transcend the effects of the Cold War rivalry and its demise.
To arrive at these conclusions, we begin with a brief overview of the notions of a Vietnam generation and Vietnam Syndrome. Building on extant international relations scholarship, we develop our expectations and two-stage research design. We then present the empirical results, which obtain even when we control for the background variables most consistently associated with elites' foreign policy attitudes in previous analyses: partisanship, ideology, and occupation (Holsti and Rosenau 1990, 116). We conclude by discussing some implications of our findings, which might be of particular interest in an era characterized by disillusionment with US military adventurism, pressures for restraint and retrenchment, 3 and the burgeoning maturation of a generation that experienced its coming-of-age during the wars of 9/11.

The Vietnam War's Legacy in the U.S.: A Generation Averse to Using Force Abroad?
Generational analysis (e.g., Mannheim [1928Mannheim [ ] 1952 is built on observations about the human life cycle as it transects the linear passage of time. The stream of political events may be depicted 4 along a linear timeline, but not all living persons experience each event at the same life-cycle stage; some will be children, others teenagers or young adults, and some mature adults. Political events may be salient to individuals' lives during any life-cycle phase, but dramatic events are thought to have particularly lasting effects when subjects are old enough to perceive political events but young enough that their attitudes and belief systems are still developing. 4 During these "formative years" a generational identity may be forged through a cohort's shared experience of salient events. Attitudinal persistence models build on this notion with the "impressionable years" hypothesis (Alwin and Krosnick 1991;Sears 1990;Sears and Brown 2013). In part because they have not yet engaged in adult interaction with the political world, adolescents have a weak "experiential base" for political attitudes to take hold. Young adults are especially susceptible to the force of external events, and by the end of that formative life stage-roughly ages 16-30political identity tends to be more fully formed and less pliable, a tendency supported by empirical studies of political beliefs' consistency in panel survey data (Jennings and Zhang 2005; also Alwin and Krosnick 1991;Jennings 1996;Sears and Funk 1999). As Bartels and Jackman (2014, 10) note, "the notion that early formative experiences have some special power to stamp different generations with distinctive political attitudes and beliefs has been one of the most familiar and influential ideas in the literature on political socialization." As political events go, the Vietnam War-given its prominence in media reporting and the potency and persistence of debate about the war-clearly had extraordinary potential for shaping a generation who experienced it as young adults. Even before the war was over, at least one scholar speculated that it would push a generation of future foreign policy elites-those who were in their 20s during the war-toward a "Vietnam paradigm" of non-interventionism in world affairs (Roskin 1974). Scholars continue to apply the impressionable-years logic to capture the shared experience of the "Vietnam generation." Busby and Monten (2008, 457-60) for example identify the Vietnam War as a formative event, and code for a "Vietnam generation" in their study of establishment internationalism among foreign policy elites. Using open-ended survey questions about "especially important events," Schuman and Corning (2012, 7) find that among respondents' judgments of pre-1985 events the Vietnam War ranks just behind World War II, and plays a special role for those who experienced the war during their "critical years" (2012,. Leery of military activism in the Vietnam War's wake, Americans-and perhaps the Vietnam generation in particular-might appear more hesitant to deploy US armed forces overseas, especially in the developing world. Conventional wisdom and the media enshrined the intuition that the War had a durable impact on Americans' attitudes toward related issues, making them more pessimistic about interventionism and the United States' role in the world. This "Vietnam Syndrome" was characterized by The Economist as "scepticism… about both the legitimacy and the efficacy of America using military power overseas." Scholarship is more equivocal on the matter. Though a variety of studies have analyzed the Vietnam War's influence on attitudes toward subsequent military activism, "three decades of scholarship on war and public opinion have not produced a scholarly consensus on a crucial question: whether the Vietnam War fundamentally transformed U.S. public opinion on the use of force" (Eichenberg 2005, 140). 5 Many studies in this vein have focused on mass public attitudes, exploring phenomena like "casualty sensitivity" and declining public support for war (Mueller 1973; see also Gartner and Segura 1998;Gelpi, Feaver and Reifler 2005-6;Kriner and Shen 2014). But the Vietnam Syndrome cannot be reduced to mass aversion to mounting casualties. Its symptoms are broadly consistent with findings that the American public is "pretty prudent" with regard to military intervention (Jentleson 1992;Jentleson and Britton 1998;Oneal et al., 1996), inasmuch as the public seems to prefer military actions focused on "foreign policy restraint" rather than "internal political change" (Jentleson 1992), and conditions its support for intervention on the probability of success (e.g., Eichenberg 2005;Gelpi, Feaver and Reifler 2005-6). Taken together, these studies depict an American public reluctant to support military interventions that might resemble the Vietnam War, even if the studies do not directly link this reluctance with the War's impact.
Given the connections' ambiguity, it would seem the diagnosis of a "syndrome" is inconclusive.
However much the American public as a whole has (or has not) suffered from a Vietnam Syndrome, the mass public's influence on policy and on decisions to deploy forces abroad is limited; the prime mover, more likely, is the American foreign policy elite. Elites are sensitive to the electorate, but also inform and manipulate the public, giving cues and divulging information through the media (e.g., Berinsky 2009;Powlick and Katz 1998). They are more likely than less interested and educated citizens to have opinions that are consistent across issues and over time (e.g., Converse 1964). Their views may also differ from the general public in some respects, as in their near-unanimous support for US internationalism while isolationism remains a minor but persistent tendency in the mass public. Elites, then, may be more inclined than the "pretty prudent" public to embrace the use of force abroad to achieve national objectives 6 , which possibly makes elite opinion a tougher test of the Vietnam Syndrome than mass public opinion.
A key question, then, is did the shared experience of the Vietnam War shape the attitudes of those with greater access to the levers of power (or significant potential to nudge those levers)? It is these foreign policy elites to which we turn our attention.

Not All Baggage Looks the Same: Unpacking the Vietnam Generation of US Foreign Policy Elites
A common assertion is that the Vietnam War shaped a generation of young Americans-and (by implication) as they aged and gained status, a generation of American foreign policy elites.
Because elites-especially those who eventually attain influential positions in government, media, and other institutions-can have a significant impact on policy decisions, the question of whether the Vietnam War influenced that birth cohort's worldviews is of interest to scholars and analysts of US foreign policy. In 1980, Ole Holsti and James Rosenau published an empirical study investigating this possibility. In their words, The generational argument, if valid, has some important implications for the future conduct of this nation's external relations because it suggests that members of the "Vietnam generation," as they achieve positions of leadership and influence during the 6 By Wittkopf and Maggioto's (1983, 306) reckoning, "Elements of the internationalist world view widely embraced by elites during the post-World War II period include a sense of active, global responsibility, a fear of communism, a conviction that the Soviet Union's expansive tendencies must be contained, a preference for military preparedness, and a willingness to use interventionist means to realize American objectives." next several decades, will bring to their roles an intellectual baggage radically different from that of the leaders they are replacing. (Holsti and Rosenau 1980, 3) Such tests of hypotheses concerning a "Vietnam generation" were arguably premature. Holsti and Rosenau (1980, 1986, 1990 themselves admitted that the tests were preliminary. Others (e.g., Yoon 1997) assessed whether a generational shift in attitudes had occurred by analyzing US foreign policy behavior, but such studies are at best indirect assessments of attitudes-and are even more premature, since generational replacement at higher levels of government had not yet occurred in the period under examination (Roskin 1974, 587-8).
Though preliminary, Holsti and Rosenau's contributions on elite opinion and foreign policy issues must not be discounted. Their work, along with that of Kegley (1986), Wittkopf (1986), and others revealed much about the attitudes and belief systems of American foreign policy elites. But even Holsti's 1996 book-which considers the generational hypothesis across five iterations of the FPLP survey and, unlike the earlier collaborative works, includes post-Cold War survey results-finds only weak and inconsistent support for generation-based patterns in respondents' attitudes. As Holsti acknowledges, "there is little evidence presented here to sustain the hopes-or fears-that the inevitable replacement of one generation by its successor will give rise to significant changes in public or elite attitudes toward foreign affairs" (1996, 166).
Holsti's conclusion follows from weak empirical associations between birth cohort and attitudes toward controversial foreign policy issues (e.g., Holsti and Rosenau 1990, 116-7;Holsti 1996). One problem with this approach to the Vietnam generation is that it implicitly assumes that the war had a uniform effect on those who experienced it. To be sure, the war may have had a greater impact on some individuals than it did on others, but Holsti's analysis still assumes that it pushed everyone it affected toward the same policy conclusions. Such an assumption makes sense for historical events about which there is an elite consensus; for instance, the experience of World War II is widely supposed to have pushed all Americans away from isolationist attitudes about the country's role in the world (e.g., Roskin 1974, 567 ff.). Given the elite consensus on this point, this is a plausible claim. However, no such consensus coalesced about the meaning of the Vietnam War.
For some, the Vietnam War revealed grave problems with the assumptions behind American foreign policy (e.g., Caputo 1977, xiv;Ellsberg 2002, vii-x). For others, it revealed limits on public willingness to make sacrifices necessary to ensure the country's security (e.g., Mann 2005, xvi-xvii, 52-4). This divergence is likely to be especially pronounced among elites who are highly engaged with American foreign policy. For an event like the Vietnam War, the weak results turned up in previous research about generation effects might stem from the fact that they effectively average across these two very different groups. And the differences do not end there-as we show below, not all members of the cohort experienced the Vietnam War in the same way, reacted to the war in the same way, or evaluated subsequent intervention opportunities in the same way.

Effects of the Uniform, but not Uniform Effects: Expectations
We do not expect to find uniform effects among members of the "Vietnam generation." Our expectation of intra-generational variation comports with recent scholarship on how specific war experiences affect individuals and their political attitudes and activities. For example, findings reported by Bellows and Miguel (2009) and Blattman (2009) from Sierra Leone and Uganda, respectively, suggest that individuals with firsthand experience of war trauma are especially likely to be engaged and mobilized in postwar political life. Evidence from South Korea suggests that wartime trauma impacts individuals' political attitudes, such as trust in the government, and that such attitudes can persist even 60 years after war's end (Hong and Kang 2017). These micro-level studies from around the world suggest, perhaps unsurprisingly, that distinct wartime experiences can shape individuals and their political attitudes in different ways. If, as we suspect, such variation gives rise to significant intra-generational heterogeneity, then broad generalizations about the Vietnam generation's tendencies will not hold (Mayer 1992). Even if one set of experiences and policy conclusions predominated and characterized the "generation" in the future, as it appears to have done during World War II, there might still be minority views-and such differences are particularly important for controversial events like the Vietnam War. Rather than a single "generation effect," it is better to think of several possibly divergent "generation effects." For instance, those who served in the military might well view a war differently than those who did not (e.g., Schreiber 1979). Horowitz and Stam (2014, 528) frame the issue in these terms: "Does military service increase familiarity and knowledge about the use of force, making those who serve more likely to support military action, or does the exposure to danger in the military make those who serve more hesitant to use force in the future?" Grossman, Manekin, and Miodownik (2015) note that international relations scholarship tends to be polarized on this question. Consistent with scholarship associated with Huntington (1957) and Janowitz (1960), some predict that veterans will exhibit hesitancy to embrace military solutions, and instead will see war as a last resort. By contrast, a school of thought associated with Posen (1984) andSnyder (1988) "argues that military experience leads to militaristic attitudes and behavior." The logic is that soldiers' intense socialization is "intended to increase their aggression toward rivals, inculcate a sense of identification with the military, and make them more comfortable with the use of force… resulting in greater acceptance of military solutions to conflict" (Grossman, Manekin, and Miodownik 2015, 984-5). Those authors present evidence for the latter, and their findings guide our predictions.
Military service is not the only way experiences might differ-for those in potentially influential positions, simply changing their mind about a policy can itself be important. 7 Doing so is costly because it alienates political allies who remain committed, especially to a highly salient policy like the prosecution of a war. If expressed openly, the new perspective could change or derail an individual's career. In addition to such risks, changing one's mind requires overcoming tendencies toward psychological inertia. For these reasons, the process of changing one's mind can be a significant, potentially formative event in its own right (e.g., Janoff-Bulman 1992). In an influential article, Tedeschi and Calhoun (2004) assert that cognitive processing of a traumatic experience or crisis can generate post-traumatic growth, a developmental process in which the individual has survived the trauma but been transformed in "deeply profound" ways (2004,4). Post-traumatic growth is not the direct result of the trauma itself, but rather of the individual's attempts to process it: It is the individual's struggle with the new reality in the aftermath of trauma that is crucial in determining the extent to which posttraumatic growth occurs… Psychological crisis can be defined in relation to the extent to which the fundamental components of the assumptive world are challenged… The "seismic" set of circumstances severely challenges, contradicts, or may even nullify the way the individual understands why things happen, in terms of proximate causes and reasons…Cognitive rebuilding that takes into account the changed reality of one's life after trauma produces schemas that incorporate the trauma and possible events in the future (Tedeschi and Calhoun 2004, 5).
Thus, the struggle to process the meaning and import of a trauma (such as an ineffectual or morally objectionable war) can reshape one's thinking about similar phenomena in the future (such as prospective military interventions resembling the problematic war).
The process of changing one's mind about the war suggests a learning process. We salute the brave souls who venture into what Levy (1994) characterized as a "minefield"-for example, Bartels and Jackman's (2014) innovative Bayesian updating model that accounts for the extraordinary impact of formative events by assigning them heavier weights than subsequent events later in life-but we do not follow their lead here. We neither propose nor test a general model of political learning. Instead, we isolate and characterize the effect of the Vietnam War experience (or more accurately, different sorts of Vietnam War experiences) as foreign policy elites age and mature, and assess whether such effects correspond to what is commonly understood as "hawkishness" or "dovishness" of attitudes about the use of American military power abroad, with an eye toward the persistence of an alleged Vietnam Syndrome.
For an event like the Vietnam War to have a lasting influence is more plausible for elites who are preoccupied with foreign policy than it would be for the general public. If being part of the "Vietnam generation" had an important influence on elite policy attitudes, then differences in attitudes toward the war associated with age and military service should persist over time, remaining roughly the same in surveys taken long after the Vietnam War ended. As we discuss in the conclusion, however, our results have more to say about intra-generational differences and the overarching Cold War context than they do about a chronic Vietnam Syndrome.

Hypotheses and Research Design
We test the impact of the Vietnam War on elite opinion in two stages, one to establish differences between generations and within the "Vietnam generation," and a second to discern whether respondents' views on the Vietnam War predict attitudes toward subsequent interventionism, either exclusively within the Cold War context or in general. We first test whether those who had specific types of exposure to the Vietnam War-those of draft age and those who actually served in the military-took different positions on the event than others.

Hypotheses
Our first hypothesis thus seeks to ascertain whether we can identify a "Vietnam generation." We expect the members of the Vietnam-era cohort to register attitudes distinct from those respondents who had already aged past their formative years. In keeping with the logic of posttraumatic growth viewed through a generational lens, we propose:

H1.
Elites from the "Vietnam generation" should be more likely to have turned against the Vietnam War than those who were too old to be drafted. On the other hand, some might have turned against the Vietnam War not because they were skeptical of the Cold War but rather because they concluded it was a civil war in which US intrusion was inappropriate, unwarranted, and doomed to fail. These observers are motivated to avoid the dangers of intervening in civil wars, a conclusion that has no obvious "expiration date." If the longevity of political attitudes shaped by war trauma is as durable as other studies suggest (e.g., Hong and Kang 2017), post-Vietnam reluctance to intervene in civil or regional conflicts, especially in less-developed areas not considered paramount in the Cold War, should persist even after the Cold War's end: H4. Those who opposed the Vietnam War should oppose subsequent interventions in civil or regional conflicts, especially in less developed areas of the world. This pattern should persist after the Cold War.
If differences over contemporary policy issues based on positions taken during the Vietnam War outlast the Cold War, we can conclude that the second effect was more important than the first.
Below we explain how we use the survey data to parse these effects and to distinguish the Cold War's contextual influence from a more general reluctance to engage in interventionism.

Data, Methods, and Controls 8
We test our hypotheses using the elite surveys conducted by the Foreign Policy Leadership Project (FPLP) between 1976 and 1996 (Holsti and Rosenau 1984;. Holsti and Rosenau surveyed a relatively large sample of potentially influential Americans every four years during this period. 9 The sample drew on several relevant professions including military officers, public officials, university professors, doctors, lawyers, and media figures. With more than half holding graduate degrees, this group was much more highly educated than the general population. As Table 1 indicates, the proportion of each profession within the FPLP sample as a whole was relatively consistent over time. The same individuals were not sampled, but the same groups were. One obvious problem with an elite sample of this sort is that the "policymaking elite" is not a clearly defined population. The structure and composition of this group is not obvious; for instance, high-ranking military officers clearly belong in the sample, but what share of the elite 9 The FPLP data have several advantages over the similar elite samples gathered by the Chicago Council on Global Affairs (CCGA, formerly Chicago Council on Foreign Relations). First, the FPLP sample is more than four times larger than that typically gathered by the CCGA. Second, the FPLP consistently included items about respondents' retrospective views of the Vietnam War and about their military service. Third, the CCGA surveys include little demographic data, including age, so we cannot test hypotheses about generational effects. "population" do they comprise relative to other groups such as academics and Foreign Service officers? Without a convincing answer to this question and corresponding set of weights, it would be a mistake to draw strong conclusions about the prevalence of particular views among policymakers. We will refrain from such conclusions, focusing instead on sources of differences within the sample.
While defining the correct composition of the policymaking elite and weighting the FPLP sample accordingly is beyond this article's scope, in our analysis we do consider the way this sample was gathered. Holsti and Rosenau sampled several professions relevant to foreign policy, including educators and military officers. One potential problem this sampling strategy poses for our analysis is that we might confound the effects of individuals' views during the Vietnam War with tendencies associated with their post-war professions. This is especially problematic because their view of the war might have influenced their choice of profession, which then served to preserve and reinforce their perspective. To handle this issue, we include a fixed-effect dummy variable for each occupational category in models of post-Vietnam foreign policy opinions. This is a conservative estimation strategy. To the extent that respondents' views on the Vietnam War influenced their choice of profession, we might underestimate the impact of the War on their subsequent opinions.
The first part of our analysis estimates the effects of belonging to the Vietnam generationbeing among those eligible to serve in the War-and of actual service in Vietnam on retrospective attitudes toward the War. Here, we are interested not only in respondents' ultimate conclusions about the war but also into their pathway to this view. Turning against the war after initially supporting it is potentially different than the experience of opposing the war from the beginning. Moreover, such conversions are a staple of memoirs written by the elites the FPLP surveyed, and should be more likely among those who were relatively young at the time.
The second and more important part of our analysis evaluates the association between respondents' positions on the Vietnam War and their subsequent policy views. We selected policy questions from the FPLP survey with two criteria in mind. First, we chose items that asked respondents' views of contemporary policy questions rather than general philosophical attitudes, also avoiding those that were explicitly linked to the Vietnam War. It would be less surprising-and less interesting-to find that broader policy outlooks, such as opinions about containment of the Soviet Union, were linked to Vietnam. Later issues with no immediate substantive linkage to the Vietnam War offer a better test of the "Vietnam Syndrome." Second, we use survey questions both about policy issues arising from the U.S.-Soviet rivalry and about interventions in less developed areas that were not clearly related to the Cold War. This allows us to discern whether differences over Vietnam exerted their effects mainly through Cold War attitudes, through attitudes toward intervention in less-developed countries, or both.
Assessing the impact of respondents' opinions about the Vietnam War and their subsequent policy views raises an important issue of causal inference. Such an association could arise because respondents' ideological commitments or party loyalties shaped both their views about the War and their later opinions. If this is the case, then there might be no causal relationship between these opinions. To avoid this spurious inference, we control for party identification and ideology when estimating the effect of Vietnam on later policy views. This is a conservative estimation strategy. The FPLP survey questions asked about respondents' current party identification and ideology. If the Vietnam War shaped these basic attitudes-as it might have among foreign policy elites-then models that include these controls will underestimate the impact of the Vietnam War experience on subsequent policy views. For present purposes, avoiding a spurious causal inference is more important than underestimating the War's impact.

Age Cohorts, Military Service, and Attitudes toward the Vietnam War
The The survey allowed respondents to choose "I tended to favor a complete military victory," "I tended to favor a complete withdrawal," "I tended to feel in between these two," and "not sure." As the question indicates, the survey recorded respondents' opinions about Vietnam both "when the war first became an issue," and "toward the end of U.S. involvement." Holsti and Rosenau (1984, 33) used responses to this question to sort positions on Vietnam based on the trajectory of the respondent's views on the war. 10 Our categorization scheme is similar to theirs, and includes five groups: Consistent Supporters: Those who favored a complete military victory at both points Converted supporters: Those who favored a complete victory at the end of the war, but not at the beginning Ambivalents: Those who held an "in between" or "not sure" position at the end of the war Converted critics: Those who favored complete withdrawal at the end, but not at the beginning of the war Consistent Critics: Those who favored complete withdrawal at both points Table 2 presents the percentage of the sample that fell into each category in the six FPLP surveys. The distribution of the sample over these categories remains roughly the same over time. As conventional wisdom about the Vietnam War would have it, "converted critic" is the modal category, comprising more than a third of the sample in every year. Much of the Vietnam War's impact hinges on whether converted critics continued to have distinctively dovish policy views over time. This is the essence of the "Vietnam Syndrome."  Holsti and Rosenau (1984, 33). Consistent supporters (critics) supported (opposed) the war in both its early and later stages. Converted supporters (critics) supported (opposed) the war only in its later stages. Ambivalent respondents did not take a clear position in the final stages of the war. Table 2 exclude the small fraction of the sample born after 1960. As children, they were unlikely to have had opinions about the war during its early stages. As Table   3 suggests, the FPLP sample was a relatively old one. Few people achieved the professional status necessary for inclusion before the age of 30, so even in 1996 only 3.5 percent were born after 1960. The comparisons that follow are between the Vietnam generation and their elders. about the Vietnam War. To do this, we estimated multinomial logit models of the five trajectories noted above for each survey. These models included a variable indicating whether respondents were young enough to be eligible for the draft during the War-our definition of the Vietnam Generation-and another indicating whether they served in the war. We also included a control for gender to ensure that military service does not proxy a gender effect. Second, military service in Vietnam tended to bolster support for the war. Most Vietnam veterans were also part of the "Vietnam generation" as defined here, so these constitute an important exception to the generation's broader trajectory toward anti-war views. Some of this pattern is certainly the result of self-selection. Members of the privileged group in the elite sample who had doubts about the war could probably have avoided military service. Our point here is not that serving in Vietnam caused the differences that emerged from the war, but only that it reinforced them and contributed to their durability.

The numbers in
The patterns in Figure 1 generally hold up in later surveys. In nearly every survey, nonveterans from the Vietnam Generation were significantly more likely to be critics or converted critics and less likely to be supporters. Similarly, Vietnam veterans were more likely to be supporters and less likely to be critics or converted critics. Figure

The Vietnam War and Later Foreign Policy Positions
The most important claim of the Vietnam Syndrome argument is that positions adopted during the war continued to influence policy attitudes for years afterward (hypotheses 3 and 4). We test this claim by using the assessments of the Vietnam War from the preceding section to predict subsequent policy views. Here we report marginal effects in graphical form. Marginal effects that are statistically significantly different from those reported for respondents who were ambivalent about the Vietnam War are marked with asterisks. The appendix contains the full numerical results of each model. In every case, we control for party and ideology, so the marginal effects we report are for ideologically moderate respondents who did not identify with either major party.

Vietnam and Later Issues: The Cold War Years.
We turn first to the surveys administered during the Cold War. The results we will report come from logit models in which the dependent   1980, 1984, and 1988, the relationship remains at least as strong for the remainder of the Cold War-and without explicit Vietnam War framing.
As one might expect, ideology also influenced attitudes toward intervention in Angola.
Among respondents ambivalent about the Vietnam War, those who were "very liberal" had just a 0.05 probability of strongly disagreeing with the anti-interventionist statement in the survey.
"Very conservative" respondents had a 0.16 probability of doing so, with other ideological selfplacements falling between these two extremes. Neither Republican nor Democratic Party identifiers were more likely to strongly disagree with the statement than those who reported no party identification. These results are not surprising. For our purposes, the most important thing about them is that positions on the Vietnam War made a difference even when we control for these well-established influences on opinion. The impact of positions on the Vietnam War was not simply a reflection of partisan or ideological differences. 16 Table A-4 in the appendix presents the numerical results of the models, including the significance tests.
We also tested whether restricted models that excluded the Vietnam opinion variables fit the data equally well or better. Model fit statistics reported in Table A-4 strongly support the full model in all three cases.
As was the case in 1976, the legacy of the Vietnam War is apparent even though we control for ideology and party identification. Our results show that conservatives were substantially more hawkish than liberals in response to all three of the 1980 questions. Party also had some effect, though it was not consistent. Democrats were more hawkish than those with no party identification in response to the question about Iran. Republicans were more hawkish in response to the question about the SALT process. These results make sense in terms of the politics of the time, yet views about the Vietnam War still mattered even after we consider these partisan and ideological patterns. The interventions in Nicaragua and El Salvador were closely tied to the Cold War. By contrast, the principal targets in the Persian Gulf and Libya were not communists. Figure 6 presents the results of the analysis, and the now-familiar pattern found in 1976,1980, and 1984 appears once again. 20 The step-like progression obtains in almost every case, with supporters of the Vietnam War adopting the most hawkish positions and critics the most dovish-a result that emerges with equal force, even for the interventions that had the least to do with the Cold War. This result obtains in spite of the controls for ideology, which also had strong effects, and party identification, which rarely did.
20 Table A-6 in the appendix presents the numerical results of the logit models, including the significance tests noted here. We once again tested whether restricted models that excluded the Vietnam opinion variables fit the data equally well or better. Model fit statistics reported in Table A 23 Table A-7 in the appendix presents the numerical results of the logit models, including the significance tests noted here. As with the other surveys, we tested whether restricted models that excluded the Vietnam opinion variables fit the data equally well or better. We found support for doing so in two of the three cases, but even in these instances it was not as strong as it had been in earlier surveys. These results once and for all" (Herring 1991, 104 These results accord with the near-absence of the step-like Cold War pattern in Figure 8.
administration's policies. Second, the results presented here control for both party and liberalconservative ideology. These considerations had substantial effects on most of the questions asked in the 1992 and 1996 surveys, just as they did during the Cold War. Conservative respondents were no longer relatively hawkish in the 1996 survey, which probably does reflect their distrust of operations launched during the Clinton administration. Indeed, they tended to oppose the Bosnia deployment. However, the Cold War pattern breaks down even on questions where conservatives adopted hawkish positions like those they had espoused during the Cold War, as they did with respect to the defense of Saudi Arabia and the use of troops to support an uprising in Cuba.
One might argue that the interventions considered in the 1996 survey are simply different than those considered in earlier surveys and should not be expected to produce the same pattern of elite disagreement seen during the Cold War. For example, humanitarian rather than security concerns arguably motivated the troop deployments in Bosnia and Somalia. The fact that many elite observers did not view post-Cold War interventions in the same way is precisely the point.
In the absence of the Cold War as an organizing device, the meaning of being a "hawk" or a "dove" changed-and patterns symptomatic of a Vietnam Syndrome appeared to go into remission. This is not to say that the Vietnam Syndrome was psychosomatic. Our results suggest that the Vietnam War crystallized elite views on a variety of confrontational policy issues, and that it had a particular impact on the Vietnam generation, leading them to more hawkish positions if they served in the war and more dovish positions if they did not. But the Cold War's end disrupted the context in which "hawk" and "dove" denoted clearly distinguishable positions, and marked an end to patterns of elite opinion that had held since the mid-1970s.

Discussion and Conclusions
Our findings on the Vietnam War's effect on elite opinion have implications for three phenomena associated with cynical attitudes toward using US power abroad: the "Vietnam Syndrome," a "Vietnam generation," and whether one or both of these prompted a persistent, dovish turn in the attitudes of foreign policy elites. Though the results allow us to endorse neither an uncomplicated version of the "Vietnam generation" claim nor an unrefined diagnosis of a "Vietnam Syndrome," we are able to present a more nuanced picture of elite opinion in the shadow of the Vietnam War. This picture is relevant today because the Vietnam-era cohort now occupies high-level government positions, because scholars are investigating the effects of war on political attitudes and generations, and because a new generation whose impressionable years coincided with U.S. wars in Iraq and Afghanistan is maturing. Amid pressures for US retrenchment and reduced interventionism, a reconsideration of the Vietnam Syndrome is timely and has renewed salience.
A more refined diagnosis must acknowledge that the incidence and effects of a Vietnam Syndrome are shaped and constrained by individual experience and international context. First, the Vietnam War's effects on attitudes toward interventionism depended on how individual views had evolved during the war. The vast majority of foreign policy elites surveyed by Holsti and Rosenau reported that they had initially supported the war, just as most of the American public did. Those who later turned against it-the modal group of "converted critics"-generally favored relatively dovish positions on many subsequent foreign policy issues, including a range of military interventions (some Cold War-related, some not). By contrast, those who had continued to support the Vietnam War through its final years took a significantly harder line on these issues through the 1970s and 1980s. Notably, this pattern is not simply a subsidiary effect of partisanship and ideology. It holds up even when we control for those well-known considerations.
Second, these attitude trajectories resulted in part not just from being part of the "Vietnam generation," but from specific experiences within it. Consistent with findings of recent microlevel analyses of war's effects on individuals, our examination of the FPLP data identifies significant, patterned differences in elite attitudes toward the Vietnam War itself. Those who served during the war were more likely to have continued to support it during its final years.
Respondents from the Vietnam generation who did not serve were more likely than their older compatriots to have turned against the war by the time it ended. Indeed, it seems that having turned against the war was a formative experience in and of itself, though notably a different one from combat or military service. Differing attitudes toward the war arising from these distinct experiences predicted divergent attitudes toward future uses of US power abroad, even within the same cohort. Overall, it is clear that the formative experience of the Vietnam War yielded heterogeneous generational effects, rather than a uniform generational effect.
Third, our results suggest that the Cold War was a critical contextual variable that sustained elite divisions that emerged over the Vietnam War. Individuals' trajectories of opinion about the Vietnam War made a substantial difference for their later policy positions, but only while the Cold War lasted. In principle, the American experience in Vietnam might have led elites to conclude that intervention in civil and regional conflicts was unwise, regardless of whether these interventions related to the Cold War. In fact, Vietnam-related patterns of elite opinion ended in the wake of the Soviet Union's collapse, suggesting that the "Vietnam Syndrome" among foreign policy elites pertained to the Cold War rather than to military intervention in general. The evidence is unequivocal: in response to the eight survey items we examined from the 1992 and 1996 FPLP studies, elites who had turned against the war in Vietnam exhibited attitudes no different from those who had been ambivalent about the war. This result contrasts sharply with the Cold War era, during which these "converted critics" held distinctively dovish views on 11 of 14 survey items we examined. Indeed, in many post-Cold War cases, even consistent critics and consistent supporters of the Vietnam War no longer held distinctive attitudes about major foreign policy issues, in stark contrast to their reliably polarized attitudes during the Cold War.
The fact that positions on the Vietnam War cease to reliably predict patterns of elite opinion after the Cold War suggests an answer to one potential criticism of our claim about the impact of the Vietnam War: that opinions about the War proxy broader "hawkish" or "dovish" attitudes. If this were the case, these differences should have been apparent on later military interventions.
The rejoinder that interventions like those in Bosnia or Somalia differ from earlier uses of force underscores this point. The sense that these interventions did not engage real security interests is precisely why the Cold War was so important. Presidents frequently cited concerns about the promotion of human rights and democracy during Cold War interventions, but the security frame defined by the U.S.-Soviet rivalry appears to have predominated. After the Cold War, the labels "hawk" and "dove"-already artifacts of the Cold War 26 -had little residual meaning. That our results highlight the context-dependence of these labels suggests that perhaps they should be treated as relics. Labels that oversimplify can mislead.
Our introduction echoed the words of decorated Vietnam veterans and foreign policy elites Chuck Hagel and Bob Kerrey, who also asserted, "Nations, like individuals, are products of their