Who Supports a Feminist Foreign Policy for the United States?

July 21, 2017

(Cross-posted with the Chicago Council on Global Affairs’ Running Numbers blog)

In August 1995, President Bill Clinton established the President’s Interagency Council on Women and declared: “We are putting our efforts to protect and advance women’s rights where they belong–in the mainstream of American foreign policy.” Elsewhere, I have described the subsequent growth of policy activity on behalf of global women’s rights and traced which political actors have been most active in its implementation.

In this post I ask a different set of questions: how strong is popular support for a “feminist foreign policy” that makes women’s rights a central priority? What segments of the population are most supportive?  Is support for global women’s rights correlated with other policy attitudes?

Some answers to these questions can be found in the annual Chicago Council Survey of Americans on foreign policy issues conducted by the Chicago Council on Global Affairs. In the 2016 survey, the Council asked respondents to rate the importance of a number possible foreign policy goals for the United States, including (among others), combating terrorism, maintaining military superiority, limiting global warming, and combating world hunger. Included among these foreign policy goals was promoting the rights of women and girls around the world. (In the analysis to follow I combine this question with a second that has very similar wording and yields almost identical results).

One feature of the results of the survey is that Americans consider most of these goals either very important or somewhat important. For example, 97 percent of respondents consider combating terrorism a very or somewhat important goal, versus 91 percent for maintaining military superiority, 90 percent for combating global hunger, and 83 percent for limiting global warming. A substantial majority –87 percent—considers promoting the rights of women and girls an important or somewhat important foreign goal.

Clearly, most foreign policy goals are considered important to some extent. The more interesting question is which segments of the population say the policy is very important. This seems all the more relevant given news reports that President Trump’s budget proposes to eliminate the office that administers global women’s issues. As Congress deliberates on that budget, it is important to know both the overall level of popular support for advancing global gender rights and who is more skeptical.

The graphic below shows the breakout of those who believe that promoting the rights of women and girls is a very important policy goal. There are few surprises. Women, Democrats, and those who intended to vote for Hillary Clinton in 2016 are far more likely to identify global gender issues as “very important.” Black Americans are also very supportive. [Click on image to enlarge]

I also examined what other foreign policy attitudes were related to support for global gender rights. Those who believe that strengthening the United Nations would be an effective way for the US to achieve its goals also support the pursuit of global gender rights. The finding is not surprising because the United Nations has been a crucial forum for advocacy on global gender issues. In addition, scholars have found that opinions of the UN and opinions favoring global justice tend to cluster among citizens who favor a cooperative approach to global affairs or a redistribution of global resources (Gravelle, Reifler, and Scotto 2017).

A second correlation may be a bit more surprising to some readers: support for the pursuit of global women’s rights is also strongest among respondents who are very worried about being the target of gun violence. At first glance, a relationship between worries about gun violence and global gender rights may not seem obvious. Several factors likely explain it. First, fear of gun violence is highly gendered –women are more likely to worry about being a victim of gun violence. Second, women are no doubt aware that women are often the victim of violent crime and violent global conflicts. Securing the rights of women is not just a question of justice –it is a question of personal safety. As Valerie Hudson and her colleagues have put it, the security of women and the reduction of global violence are integrally related (Hudson et al 2012). What may be surprising is that a simple question about gun violence elicits responses that are cosmopolitan in their overtones.


Gravelle, Timothy B., Jason Reifler, and Thomas J. Scotto. 2017. “The Structure of Foreign Policy Attitudes in Transatlantic Perspective: Comparing the   United States, United Kingdom, France and Germany.” European Journal of Political Research, March. doi:10.1111/1475-6765.12197.

Hudson, Valerie M., et. al. (2012). Sex and world peace. New York: Columbia University Press.


Principal Actors in the Pursuit of Women’s Rights in US Foreign Policy, 1995-2015

July 25, 2016

Richard C. Eichenberg (Tufts University)

With the assistance of

Elizabeth Robinson, Tufts 2015

Lily Hartzell, Tufts 2018

Lara LoBrutto, Tufts 2017


Second in a series; the first post in the series is here



 In the first post in this series, I described a new dataset that measures the volume of policy activity announced by the Office of Global Women’s Issues in the US Department of State, the office that is a principal location for the formulation and implementation of US government actions in pursuit of global gender equality.

To review, my data on policy “actions” include: a government official speaking in public; all remarks, testimony, speeches, roundtable discussions, interviews, and public appearance; and all press releases, reports, fact sheets, newsletters, op-eds, legislation, and blog posts.

The first post in the series described the yearly evolution of the volume of policy activity.  In this post, I describe which policy actors are most often associated with these policy actions.


Review: The Evolution of Policy Activity across Presidential Terms

 The following graphic displays the total number of policy actions announced during each presidential term (or partial term).  With one exception, policy activity shows incremental growth since the establishment of the President’s Interagency Council on Women in 1995. The exception is the first term of President Obama from 2009 through 2012: the volume of policy activity was more than twice that during any other presidential term (note that activity in the second Obama term will increase when the data are updated through January 2017).

(click image to enlarge)



The “Wellesley Effect” : Principal Actors Associated with Policy Actions

 As I noted in my first post,  the most obvious hypothesis to explain the large increase in activity in the first Obama term is the personal engagement of his Secretary of State, Hillary Rodham Clinton, who had also been the first honorary Chairwoman of the President’s Council on Women  and whose activism on issues of global women’s rights is longstanding.  Of course, variations in the engagement of Secretaries of State will also depend on other factors, especially the urgency of competing policy issues that compete for the Secretary’s attention.

My data collection allows me to measure the personal involvement of major political actors in the actions taken in pursuit of women’s rights.  Specifically, for each action that is tallied in the database, my research team also tallied each political actor who was associated with the action.  Thus, if Secretary of State Clinton or Secretary Colin Powell made a speech announcing a gender equality initiative, they are counted as being associated with that action (two or more actors may be associated with the same action, but the action itself is counted only once).  The research team tallied the involvement of the following actors: the President, First Lady, Secretary of State, and Ambassador for Global Women’s Issues (the title of the latter position varies over the years).

On average, 60 percent of actions are associated with at least one specific actor (others are simply announced as departmental or USGOV initiatives), but the percentage varies widely (the standard deviation since 1995 is 17 percentage points).  Perhaps unsurprisingly, Presidents and First Ladies are associated with fewer actions (together about 15 percent), while the Secretary of State and Ambassador for Global Women’s Issues are each associated with 20 to 25 percent of actions.

The relative involvement of each Secretary of State and Ambassador for Global Women’s Issues reveals several interesting patterns (see graphic below).  First is a finding that I call the “Wellesley Effect”:  Secretaries Albright and Clinton were associated with the largest percentage of actions by far – almost twice the percentage of any other Secretary of State.  Second, the increasing importance of the Ambassador for Global Women’s Issues is evident in the very high percentage of actions in which they were involved during both Obama terms under Secretaries Clinton and Kerry (Ambassador Melanne Verweer under Secretary Clinton and the first eight months under Secretary Kerry, and Ambassador Catherine Russell from August 2013). Third, the three Secretaries whose involvement exceeded that of their Ambassadors were women: Albright, Clinton, and Rice.  Finally, the data reveal the effect of competing priorities on the Secretary of State.  Secretaries Powell and Kerry, under whom the Ambassador is associated with a larger share of actions, were engaged in intense diplomatic activity on security issues: Powell on the wars in Iraq and Afghanistan, and Kerry in mediation of the Israeli-Palestinian conflict, negotiations on the Iranian nuclear agreement, and multiple negotiations on the Syrian civil war and the war against ISIS.









One final way to characterize the involvement of the principal actors is to ask which actor was associated with the largest percentage of policy actions during each presidential term.  The table below provides this information.  Several things stand out. First, the most active actors are all women, which suggests that it does indeed make a difference to have women represented in the executive policy process.  Second, Hillary Clinton has twice been the most active actor in this policy field, as First Lady in 1995-1997 and again as Secretary of State in the first Obama term.

Presidential term               Most active (% of actions associated with actor)

Clinton  1995-1997    First Lady Clinton (31%)

Clinton  1997-2000    Secretary Albright (39%)

Bush 2001-2004      Senior coordinators Palmerlee and Ponticelli (28%)**

Bush 2005-2009         Secretary Rice (17%)

Obama 2009-12  Secretary Clinton/Ambassador Verweer (tie 37%/ 36%)

Obama 2013-2015     Ambassadors Verweer and Russell (38%)


** The Senior Coordinator for Global Women’s Issues was the predecessor office of the Ambassador for Global Women’s Issues, which was established in 2009.


Technical note:

The Codebook for the data collection is now available.

Policy Activity in the Office of Global Women’s Issues, US Department of State 1995-2015

June 27, 2016

Richard C. Eichenberg, Tufts University

Elizabeth Robinson (Tufts 2015)

with the assistance of

Lily Hartzell (Tufts 2018)

Lara LoBrutto (Tufts 2017)

First in a series of posts

Update 7/12/2016: codebook now available


In August 1995, President William Clinton established the President’s Interagency Council on Women and declared:  “We are putting our efforts to protect and advance women’s rights where they belong–in the mainstream of American foreign policy.”  In January 2013, President Barack Obama issued a Presidential memorandum on the “Coordination of Policies and Programs to Promote Gender Equality and Empower Women and Girls Globally.”  The memorandum directed the Assistant to the President for National Security Affairs to establish an inter-agency working group to coordinate government-wide implementation of policies to promote gender equality and advance the status of women and girls internationally.

There has been some scholarly attention to the subject. Karen Garner chronicled the origins and development of the policy in the Clinton administration and analyzed the role of important political figures in the pursuit of gender equality.  More recently, Valerie Hudson and Patricia Leidl traced the evolution of the policy through the Obama years and offered critical analysis of its implementation.  Helen Laville reviewed critical perspectives on the role of women’s rights in American policy since 1945.

Nonetheless, questions remain about the priority that is placed on the pursuit of women’s rights in foreign policy.  Has declaratory policy actually led to increased policy activity?  In what policy areas?  Is the policy focused on specific countries or regions? Which presidents and secretaries of state have been most active in the pursuit of global gender equality, and on which policy problems do they focus?

In this and subsequent posts, we describe a new data set that documents the volume of policy activity announced by the bureaucratic organization that is most responsible for the pursuit of gender equality in US foreign policy: the Office of Global Women’s Issues in the US Department of State (OGWI –the name has varied over the years). In this first post, we provide a brief description of the procedures that we employed to collect the data and an overview of the number of policy actions in the years 1995-2015.  Subsequent posts will focus on the apparent priority placed on gender equality by different secretaries of state; the programmatic focus of policy; and the regional focus of policy actions. (A separate data collection on gender equality actions in USAID is underway)


Measuring Policy Activity

We define policy activity as any policy action that is publicly announced or otherwise reproduced on the OGWI website. These actions appear on the main webpage of the Office of Global Women’s Issues (on the “Releases” link on the left sidebar), which further links to the following two categories: “Remarks, Testimony” and “Other Releases.”  (A list of links for all actions since 1995 are reproduced in a Codebook that we will eventually distribute with project materials).

Within the two categories above, we counted any occurrence of the following types of actions: a government official speaking in public; all remarks, testimony, speeches, round table discussions, interviews, and public appearance; and all press releases, reports, fact sheets, newsletters, op-eds, legislation, and blog posts.

Two examples illustrate the sorts of “action” that we count.   In June of 2008, First Lady Laura Bush traveled to Afghanistan to meet with a variety of women’s groups and to announce several grants to support universities and literacy programs.  More recently, in March 2016 Secretary of State Kerry announced a $47 million initiative for a Global Strategy to Empower Adolescent Girls.

We do not claim that this measure of the policy output of one office provides a comprehensive picture of the government’s commitment to women’s rights in foreign policy. That will require a fuller study of the budgetary resources and personnel that are committed to the policy, together with a qualitative assessment of the depth and persistence with which the policy is pursued. A full picture would also include a study of legislative actions, such as the Women’s Peace and Security Act, which is now making its way through both houses of the US Congress.

Nonetheless, the OGWI is the US government’s primary implementing office in the pursuit of global women’s rights, and we believe that a description of the volume of its policy activity and its programmatic and geographic priorities represents an important tool for tracing the evolution of an important foreign policy priority.


Overview: Total Policy Actions, 1995-2015

The graphic below displays the number of policy actions announced by the OGWI in each year since 1995.  Three features of the data stand out. First,  the priority of women’s rights in American foreign policy has indeed grown.  In the last two years of President Clinton’s first term, a total of 12 policy actions were announced.  In the last two years of President Obama’s first term, at least one policy action was announced every two to three days (151 actions in 2011).  Although there was a decline after 2012 in the volume of policy activity, in 2015 the number of actions was nonetheless higher than in any year prior to the Obama administration.










Second, there are two identifiable spurts in policy activity:  during the middle years of the Bush presidency (2002-2006) and most prominently during the first term of the Obama presidency (especially 2010-2012).  In the case of the former, the spike in activity occurs during the most intense years of the wars in Afghanistan and Iraq.  In a future post, we will explore the extent to which policy activity was in fact focused on those two countries.  In the case of the increase during Obama’s first term, it is of course possible that policy activity was simply following the President’s own priorities, but that explanation falters in light of the substantial decline in activity that occurred after 2012. It therefore seems equally plausible that the peak of policy activity during the first Obama term was due to the personal interest and priorities of Secretary of State Hillary Clinton, who had been the first honorary chair of the President’s Interagency Council on Women before giving her famous 1995 speech to the United Nations Fourth World Conference on Women in Beijing.  In a future post, we will explore the extent to which each Secretary of State was personally associated with the policy actions announced during their tenure.

Third, although policy activity does grow incrementally, there are also periods of decline, specifically at the end of the second Bush term and the end of the first Obama term.  In each case, the likely explanation is a pressing competing priority on the agenda of the Secretary of State (the Iraq surge under Secretary Condoleeza Rice and the negotiations with Israel, Palestinians, and Iran undertaken by Secretary of State Kerry).  Nonetheless, each decline is followed by a period of increase, which suggests that women’s rights have become an institutionalized priority in American foreign policy.


Technical Note

Because of the large spike in policy activity during the tenure of Secretary of State Hillary Clinton and the political interest that surrounds her candidacy for president, we invested additional effort to double-check the accuracy of our data for the period of her tenure as Secretary and the period following her tenure.  The data through 2012 were originally collected by Coder 1.  To double-check her work, two additional coders replicated the data collection for the period 2009-2015.  For the period of overlap (2009-2012), the number of policy actions recorded by Coder 1 and Coder 3 were nearly identical, but Coder 2 counted fewer actions in two years (22 fewer in 2011 and 17 fewer in 2012).  For the period 2013-2015, Coders 2 and 3 were nearly identical in their counts.  We chose a conservative method for reconciling the two years of difference during 2011 and 2012 and all other years of overlap: we averaged the number of policy actions recorded by the three coders.  This method has the effect of slightly lowering the totals for these two years in particular.  For other years, as noted, the counts for all three coders are nearly identical so that averaging has little impact.



Europeans’ Support for Defense Spending is Higher Than Any Time in the Last Fifteen Years

June 15, 2016

A recent report from the Pew Research Center’s Global Attitudes Project observes that a lack of support for increased defense spending suggests that Europe is “reluctant to support the use of hard power.”

“Only in Poland (52%), which spends 2.2% of its GDP on the military, and the Netherlands (49%), which spends 1.2%, does roughly half the public support increasing outlays on national defense. Despite commitments by their governments to boost military spending, around half the public in France (52%) and Spain (52%) want to keep defense spending the same as it is today, as does a plurality in Germany and Greece (both 47%). A third of the public in Spain (33%) and about a quarter in Italy (23%) favors cutting military outlays.”

I believe that these percentages substantially underestimate Europeans’ support for defense spending.  Most scholars of defense spending focus not on a single percentage favoring increases or decreases, but rather on the balance of opinion between these two categories (the signal work is by Christopher Wlezien).  After all, for elected leaders, the important political question is whether constituents want more or less defense. Further, most scholars tend to ignore the response category that says spending should “remain the same” or that it is “about right.”  Some citizens may feel this way, but many choose this neutral option because their opinion is not intensely felt one way or the other.  In other words, “keep the same” is not as politically relevant as the percentage of citizens who actually want an increase or a decrease.  Further, most politicians will want to know if and how opinion is changing.

I address both of these issues in this post.  First, I recalculated support for defense spending in the Pew survey using a simple formula for the balance of opinion that wants an increase or a decrease in the defense.  This measure –net support–  divides the percentage who want an increase by the total percentage who favor an increase or a decrease (% favor increase / %favor increase+% favor decrease).  Put simply, this measure summarizes support for increasing defense spending as a percentage of the total who want a change in spending.  Second, using surveys conducted by the German Marshall Fund’s Transatlantic Trends project, I compare this year’s findings from the Pew Survey to earlier years.

The results for four countries are shown in the graphic below.  Clearly, this measure of support for defense spending is higher in the three European countries than it has been since 2002 (in France and Germany considerably so).  Were I simply to show the percentage who favor an increase, the trend would be exactly the same. Paradoxically, support for defense spending is lowest in the US in 2016.















The same increase in support for defense since 2002 is evident in the other countries sampled by Pew, but showing them all would take considerable space.  Instead, in the graph below, I summarize net support for increased defense spending in 2016 compared to the average net support from 2002 through 2012.  The increase in support over the last fourteen years is considerable in all but Spain. Before 2016, support for increased spending was a majority in only two countries, the UK and Poland.  In 2016, it is a majority in all but Spain. In France, Germany, the Netherlands and Sweden, the increase over the recent average is particularly large.












To summarize: I have argued that net support for defense spending is a preferable measure of citizen preference because it summarizes the balance of opinion between those who want increases and those who want cuts, which I argue is a more politically relevant measure.  Second, by comparing the recent Pew question on support for defense spending to past surveys, we see that citizens are more receptive to increases in defense than at any time in the last fourteen years. Whether this means that Europeans are reluctant to embrace “hard power” is a different question, but for the moment lack of support for defense spending is not a source of that reluctance.

Full disclosure: I served as an academic adviser to the GMFUS Transatlantic Trends project from 2004-2014; the views expressed here are my own.


Does the American Public Agree with Donald Trump’s Views of NATO?

March 24, 2016

Richard C. Eichenberg, Tufts University

Twitter: @IkeEichenberg

America’s most notorious presidential candidate recently expressed critical views of the US commitment to NATO in a meeting with the editorial board of the Washington Post.  These views have no doubt increased European uncertainty about the steadiness of the US commitment, especially as they come on the heels of several years of European disquiet with the Obama administration’s “pivot to Asia,” its approach to the conflict in Syria, and President Obama’s critical comments about “free riders” in his interview with Jeffrey Goldberg (nicely reviewed and analyzed by Simone de Galbert  here).

Are Americans still committed to Europe’s security?  Based on many years of polling, there are two answers: the first is “Yes,”  but the second is “it depends on whether you ask Democrats or Republicans.”

The first chart below shows answers to a variety of questions that elicited positive and negative assessments of the NATO Alliance in national polls since 1974.  Since 1974, an average of 62% of Americans have expressed a positive view of NATO (they want to keep or increase the commitment to NATO or believe that NATO is essential to American security).  The positive assessments do fluctuate, but not by much, which suggests that the commitment is rooted in a solid affinity of interests and common identity with Europe. The three most recent polls are slightly lower, but not lower than some readings in the past, and in any case we might easily surmise that they have risen once again after the terrorist attacks in Paris and Brussels (the survey on NATO’s essentiality was not conducted in 2015 or 2016).












A second question shown in the graphic below reveals a sharp divide between Democrats and Republicans on the security partnership. From 2004 to 2014, the German Marshall Fund of the US asked a question about the US-European security partnership.  The question reads: “Do you think that the partnership in security and diplomatic affairs between the United States and the European Union should become closer, should remain about the  same, or should the United States take a more independent approach from European Union?”  Below, I show the percentage of Democrats who prefer a more independent approach from Europe: it averages 21% since 2004 (obviously, remaining Democrat respondents prefer a close or closer partnership).  Among Republicans, however, an average of 44% prefers a more independent approach, and the average since 2009 has been 47%.  Put bluntly, the solid consensus of Democrats is for a close or closer partnership with Europe, while Republicans are divided. (The same divided pattern is evident in party views of NATO’s essentiality –Republicans are less likely to find it essential—but a modest majority of Republicans and a large majority of Democrats find  NATO essential).







None of these figures explain why Republicans are less committed to Europe, and without a historical party breakdown, it is difficult to tell if the trend toward independence is recent or longstanding.  What the data do tell us is that Trump’s criticisms will find some resonance among Republicans.  Put differently, his views are unlikely to damage his electoral prospects. On the Democratic side, the figures suggest that support for continuing partnership with Europe is a view that is likely to find more resonance than a critical one.

For reference:

Most of the surveys above are taken from the German Marshall Fund’s Transatlantic Trends series. Party breakdown does not include Independents who may lean toward Democrats or Republicans.


Would American and Allied Public Opinion Support Military Action if the Iran Negotiations Fail?

July 2, 2015

Richard C. Eichenberg

Twitter: @IkeEichenberg

 Negotiations continue in Vienna on an agreement that would limit the Iranian nuclear program.  As the talks continue, it is useful to ask: how will public opinion react if the negotiations fail?

Although the wisdom and likely effectiveness of a military strike against Iran have been much debated, the position of the American government has long been that military action remains “on the table.”    Whether President Obama or any future President would choose the military option is obviously uncertain, but one factor that will likely condition the choice is the anticipated reaction of public opinion, both in the United States and in allied countries.

In the US, public support for military action against Iran has been lukewarm over the past ten years –an average of 45 percent from 2002-2013 –although it climbed to over 50 percent after President Obama initiated negotiations with Iran. This is lower than the 60 percent of the public that supported war with Iraq prior to 2003, and it is probably the frustrations of the Iraq and Afghanistan wars that have left Americans less than enthusiastic about further military endeavors.

Another reason may be the simple fact that negotiations are underway and that some progress has already been made in the form of an interim agreement (which the public supports). Public opinion generally favors negotiation and nonmilitary solutions so long as these are realistic alternatives, but pollsters rarely test respondents by asking them what action they would favor should negotiation prove fruitless.

An important exception is a series of questions on Iran in the German Marshall Fund’s Transatlantic Trends surveys. A first question in the battery asks: “As you may know, efforts to prevent Iran from acquiring nuclear weapons are under way. Which of the following do you think is the best option?”  Respondents are then offered alternatives that include “military action,”  “economic incentives,”  “economic sanctions,” “support to opponents of the government,” and “use of computer technology to sabotage nuclear installations.”  A final response option is to “accept that Iran could acquire nuclear weapons.”

This is a complicated question, but the public’s preferences are clear:  an average of only 7 percent of Europeans in 2014 chose “military action” over the alternatives (in the US it was 12 percent).  Among the nonmilitary options, economic incentives and economic sanctions were far and away the most popular options.  The breakdown of responses to this question have been very stable since 2010.

A second question was directed to those respondents who chose nonmilitary options in the first question: “And now imagine that all of these non-military options have been tried and the only option left to prevent Iran from acquiring nuclear weapons is the use of military force. In that case, should the [European Union/United States/Turkey/Russia] take military action against Iran, or should [it/they] simply accept that Iran could acquire nuclear weapons?”

The responses to this second question are shown below: (click to enlarge image)

iranmil14An important feature of the responses to this question is the unusually high number of “don’t know” and refusals—an average of 22 percent and as high as 30 or 40 percent in some countries. Many respondents simply do not want to answer this question or express an opinion, perhaps because it presents a choice that genuinely conflicts many citizens, or perhaps because they prefer their own alternative (a wish that negotiations could somehow continue?).

Support for military action (combining “military action” in the first and second questions) is divided, although it is higher in some countries than recent levels of support for other military interventions, including intervention in the Syrian civil war (see 2013 and 2014 surveys here). Majorities favoring military action against Iran if negotiations fail exist in only five of the thirteen countries in the survey, although rejection of military action characterizes opinion in several important P5+1 countries, including Germany, Russia, and the United Kingdom.

Nonetheless, I think caution is warranted concerning the firmness of this support.  One reason is the high nonresponse rate noted above.  I suspect –although it is pure speculation—that many of these respondents would have preferred a response option along the lines of “keep trying…continue negotiation to find an agreement.”  If so, actual support levels are probably somewhat lower than those shown above. Further, even if the current negotiations do not succeed (or if violations of any eventual agreement are debated), further negotiation or employment of nonmilitary instruments of policy are likely to follow, and citizens have shown a clear preference for these options over the use of military force.  Finally,  the questions above refer to “military action” by the “European Union” –only in Turkey, Russia and the US are national military forces referenced.  As I showed in this paper, support for using force is always highest when the question refers to an abstract “military action” rather than to a specific action, such as sending troops.

The latter type of specific question was asked in Transatlantic Trends only in 2012 (the overall responses to the preliminary questions were very similar to those from 2014 shown above).  Respondents who chose “military action” to either preliminary question were asked further if they would support using their own country’s military forces  a.] to conduct air strikes or b.] to send ground troops.

The results are shown below for five of the six countries involved in the P5+1 talks with Iran (there is no survey for China). The chart shows clearly that support for conducting air strikes or sending ground troops is extremely unpopular in every country (although air strikes enjoy majority support in the US, confirming a similar result found by the Chicago Council on World Affairs in 2014).  Even lower levels of support characterize opinion in other countries (not shown).  The upshot of the results is that, even though respondents may support “military action” in the abstract, they shy away from supporting the use of their own country’s military forces. If the US does conduct airstrikes, it would likely act alone.




Summary and Conclusions

The surveys presented above point to three conclusions.  First, negotiations with Iran continue, consumers of polls should be cautious with the proclivity of pollsters to ask about an abstract “military action.”  As was the case in the run-up to the Iraq War, support for using force is inflated in such questions.  Second, public opinion in all countries prefers nonmilitary instruments of policy when dealing with Iran, including nonmilitary economic coercion.  So long as any type of negotiation with Iran continues –even in the absence of a final agreement—public opinion is likely to prefer nonmilitary instruments to the use of military force. Finally, even in the event of a complete breakdown in negotiations –or arguments that any agreement is flawed– support for specific uses of military force, such as air strikes or sending troops, is unpopular based on past patterns.  Whether this would change in the context of a renewed harsh tone in relations with Iran remains unknown, but the surveys from 2012 shown immediately above suggest that an attack on Iran would be a hard sell with public opinion in most countries.


Note: in the interest of full disclosure, I should note that I was an academic adviser to the Transatlantic Trends surveys from 2004 through 2014, which means that I helped write many of the questions.  However, I cannot take credit for the Iran questions presented here.  They were the brainchild of another academic adviser, Professor Pierangelo Isernia of the University of Sienna.

Thanks to Pierangelo Isernia, Bruce Jentleson, and Dina Smeltz for comments on an earlier draft of this post.

No, Americans are not misremembering their support for the Iraq War

May 27, 2015

Richard Eichenberg

Twitter: @IkeEichenberg

A recent YouGov/Economist survey seems to suggest that the American people are misremembering their level of support for the Iraq invasion in 2003.  As the figure below shows, only 38% of Americans say today that they supported sending troops to Iraq back in 2003, whereas 63% of Americans said they supported the invasion in a Gallup/CNN/USA Today poll conducted in February 2003.  YouGov’s interpretation: “Americans’ memories of their own past beliefs about the 2003 Iraq War are tinged with their current feelings about what has taken place there since and what is taking place there now.”


There is one problem with this interpretation: it is wrong.  The 2015 question from YouGov was asked after respondents had been exposed to extensive reporting and public debate about the human costs and the frustrations of the war.  In order to accurately compare results from 2003 and 2015, the appropriate question therefore is how the public evaluated a possible war in 2003 when similar considerations were included in survey questions (the Gallup/CNN/USA Today question from 2003 does not mention these considerations).

My own research, along with research I’ve conducted with Richard Stoll, show that citizens’ answers in 2003 were very close to what they think today. True, as shown below, in 49 individual survey questions that were asked from the beginning of 2003 through the day before the invasion, 62% of Americans favored some form of military action –almost identical to the 2003 numbers shown above.  However, as the chart below also shows, in the fourteen available questions that mentioned “ground troops” in the same time period, the percentage was lower – 57%.  Further, if casualties were mentioned in conjunction with any military action, support was 52%. Finally, if sending ground troops and casualties were both specifically mentioned in the question, support dropped to 42% –not far off the number that You/Gov ascertained just last week (the 3 questions were asked in the last three months of 2002).


In summary, when Americans were asked before the war began in 2003 if they would support sending troops that would experience casualties, far less than a majority said yes.  And that is about the same number who say they remember it that way in the YouGov poll.

The view that the American people supported the war in 2003 seems to be widespread, and these polls help us to understand why.  Polling organizations asked about support for the war 49 times from the beginning of 2003 through the day of the invasion of Iraq, but only 4 of those questions asked about support if casualties would be suffered.  More than half of the questions (26) referred to an unspecified, abstract “military action,” for which average support was 66%, but as the figure above shows, support was much lower when specific actions or casualties were mentioned.

The potential skepticism of the public about the war that was about to occur was underestimated for a simple reason: the pollsters rarely asked the right question.  One thing is clear however: the recent You/Gov poll shows that American citizens accurately remember they opposed the war that they got.


I am grateful to Jeffrey Berry, Debbie Schildkraut, Richard Stoll, and Janie Velencia  for comments on an earlier draft



Richard Eichenberg, “Victory Has Many Friends: The American Public and the Use of Military Force, 1981-2005,” International Security 35/1, (Summer 2005).

Richard Eichenberg and Richard Stoll, The Political Fortunes of War: Iraq and the Domestic Standing of President George W. Bush,  London: The Foreign Policy Centre, July 2004.

Do Women Dislike Drone Strikes More Than Other Types of Airstrikes?  (yes, but only a little)

February 27, 2015

Surveys in many countries show very large gender differences in approval of strikes by pilot less drones. For example, in June 2013, the German Marshall Fund Transatlantic Trends survey asked the following question (click on any image to enlarge):



















Similar results occurred in a Pew Global Attitudes Survey conducted in Spring 2013; the average gender difference in support for drone strikes was 20 percentage points:



The consistency of this finding has led to speculation as to why women should be particularly negative about drone strikes and even to some disagreement as to whether gender is really a significant cleavage in public opinion on war and peace issues.  Insightful analysis here, here, and here.

A frequently cited explanation for the large gender difference on drones is that women are more sensitive to civilian casualties, and there is some evidence for this in American public opinion (as in this survey by Pew from February 2013). However, this same Pew survey of the US revealed that women are also more likely to cite other reasons for opposing drones, including concerns about their legality, their effect on the US image, and concerns about retaliation by “extremists.”

Thus, it may simply be that opposition to drones among women is part of a broader pattern of greater skepticism toward the use of military force.

An additional question is why drone strikes should evoke stronger gender difference than other types of air and missile attacks.  After all, press treatment of drone technologies emphasizes their technological sophistication and “last minute” target acquisition capabilities.  Surely citizens have gained the impression that collateral civilian damage from such strikes, while possible and even evident, may be less than conventional air strikes from fighter or high-flying bombers?  We know, for example, that the public reacted negatively to several instances of mistaken air attacks against civilian targets during the Gulf War of 1991, the war in Kosovo in 1999, and in the wars in Afghanistan and Iraq (see the excellent Rand Corporation study of public reaction to these incidents by Eric V. Larson and Bogdan Savychof  here).

In this post, I provide additional insight on these questions. The analysis draws on my study of gender difference in opinions of the use of force in as many as 37 countries from 1990 through 2004. Complete details on my definitions and other methodological issues are provided in this paper. Briefly summarized, I evaluate “support for using military force” by including any survey question that seeks a positive or negative opinion on “the potential or actual use of military force [past, present, or future]… including questions that actively (if sometimes hypothetically) query approval or disapproval of an action involving military force as a means of policy and also including questions that ask if the action is justified, appropriate, or the right thing to do.”

The dataset includes survey measures of support for using military force in six historical episodes: the Gulf War of 1991; the ensuing confrontation with Iraq over weapons inspections (1991-2002); NATO’s intervention in Bosnia (1992-1995); NATO’s attack against Serbia in support of the Kosovar Albanians (1998-1999); the US war against the Taliban regime in Afghanistan (2001-2004); and the war against Iraq and subsequent occupation (during 2003-2004).  The full dataset includes 612 survey measures of support for using force, drawn from 37 countries (for a separate analysis of American public opinion since 1984, see this paper). In the analysis below, survey questions about air or missile strikes were asked in 31 countries; most (but not all) of these occur in NATO member countries.

In this post, “gender difference” is the percentage of women who support the use of force minus the same percentage for men.

The first chart below shows that women are less supportive of any type of military action, but air and missile strikes are indeed the type of action that evokes the largest gender difference: women are less supportive by an average of 18 percentage points (note: the average for naval actions is based on only four survey questions).  Note also that the gender difference of 18 percentage points is identical to the average for drone strikes in the GMF survey shown above, and it is only slightly smaller than the average in the Pew survey from Spring 2013 (20 percentage points).


Gender Difference in support for military actions












The chart shown below indicates that air or missile strikes elicited larger gender differences in four of the six episodes in my study.  Although I do not show the results for individual countries here, gender difference on air or missile strikes is the largest for any military action in all but five countries (the exceptions are Greece, Japan, Poland, Portugal, and Turkey).


Gender Difference in Support for Air or Missiles Strikes Compared to Other Military Actions Combined


 These results make clear that the large gender difference on the question of using drones is not unique: women dislike air and missile strikes (relative to men) by a margin that is at most only slightly lower than their dislike for drone strikes.  However, what these summary results do not tell us is why drone, air, and missile strikes should evoke such a large gender difference compared to other types of military actions.  As noted above, it may be that women are more sensitive to casualties, but the results here indicate that the issue is not casualties in the absolute: sending or increasing troops are actions that also risk casualties, but the gender difference for these actions is smaller.  It may be that air strikes and drone strikes risk higher civilian casualties (as the Pew surveys seem to find), which might imply that women are more likely than men to feel empathy or solidarity with civilians in target countries. Finally, it may be that the sudden, unexpected nature of air strikes evokes the greater sense of vulnerability to violence that women are known to experience relative to men (see this paper for an interesting discussion of this possibility). Which of these explanations is closest to the mark is something we cannot say on the basis of aggregate percentages such as those presented here, so a fuller understanding of gender difference must await additional research that explores these hypotheses in greater detail at the individual level.


Follow on Twitter:  @IkeEichenberg




No, the American Public Does Not Like Torture

December 10, 2014

With the publication of the Senate Intelligence Committee’s torture report, I have already seen claims that, morality aside,  the American public accepts or indeed “likes” torture.

This is factually incorrect.

The assertion is sometimes based on a question that was frequently asked by the Pew Center for the People and the Press.  The text of the questions reads:

 Pew:  “Do you think the use of torture against suspected terrorists in order to gain important information can often be justified, sometimes be justified, rarely be justified, or never be justified?”

From the standpoint of survey science, this question is problematic. Note that it offers 3 responses about how often torture  may be justified, but only 1 response that it is “never” justified.  The cards seem a bit stacked here (still: from 2004 to 2011, the average responses to this question were as follows:  46% often or sometimes justified; 50% rarely or never justified.)

The question has an additional flaw: it states as a premise that  torture can  “gain important information”.  But we know from the Senate’s report that this premise is debatable, to say the least.

The Pew Center does good work, but this question is not an example of it. More details on this question appear in my paper on the subject.

If we compare alternative survey question wordings on torture, the picture changes considerably.  For example, here are two questions asked by other survey organizations:


Gallup: “Would you be willing — or not willing — to have the U.S. government….Torture known terrorists if they know details about future terrorist attacks in the U.S?” (average of 2001 and 2005)

Not Willing   56%

Willing           41%


ABC/WP: “Would you regard the use of torture against people suspected of involvement in terrorism as an acceptable or unacceptable part of the U.S. campaign against terrorism?” (average of 2003 – 2005)

Acceptable         31%

Unacceptable    66%

Source for above here

If we go beyond these general questions to include very graphic questions about specific torture techniques that we now know were utilized,  we find not just rejection but revulsion.  See, for example, Table 3 in  this excellent article.

Who Likes Torture?

The following graphics provides an answer.  In 2009, following President Obama’s announcement of his executive order banning torture, the ABC/Washington Post poll asked the following. “Do you support this [Obama’s] position not to use torture, or do you think there are cases in which the United States should consider torture against terrorism suspects?” (January 2009 – June 2009).

The responses are shown in the graphic below.  When Obama announced his torture ban in January, Republican men were the only group who opposed it.  This changed later when Vice President Cheney mobilized Republicans in defense of the practices of the Bush administration.

Percent who believe “there are cases in which the United States should consider torture against terrorism suspects”

(that is,  they oppose the Obama ban on torture)   — click to enlarge




Source: ABC/Washington Post surveys (see Table 1 and Appendix 1 in this paper)



October Weekly Focus: Gender Difference in Four Crucial States

October 31, 2012


October Weekly Focus: Gender Difference in Four Crucial States

Richard C. Eichenberg

Elizabeth Robinson


(polls through October 28th)

Note: all percentages based on 2 party vote, dk’s excluded

(Update, 10/31:  Graphic corrected to show VA, not FL, in title)


Our previous posts have shown a very stable division of the gender vote for most of October both nationally and in the so-called swing states (based on polls through October 24th).  In this post, we update polls through October 28 for four crucial states: CO, OH, NH, and VA. We chose these states because of their obvious importance to the electoral college, but also because in at least three of them (CO, NH, and VA), there have been some signs in late October polling that Obama’s position may have improved slightly, and it is interesting to examine whether this improvement is due to the gender division of the vote.

There have been 69 polls in these four states during October, totalling over 58,000 respondents.  The average is 5,000 to 15,000 respondents per week. Still, even at these large sample sizes, we should treat any change in percentages of 2 percent or less with caution.  Sampling error and other statistical noise does not disappear, even at large sample sizes.

The graphic below shows the weekly average gender division of the vote in the four states combined.  The pattern is very familiar; after leading in September, Obama’s share of both the male and female vote declined in October, reaching a stable level of approximately 55 percent of women and 45 percent of men (one extra word of caution: this includes only 3 polls after Oct 28th).

Given sampling error and other noise, the best characterization is that these combined percentages have been largely unchanged since October 7th.

(click on image to enlarge)

State-by-State Summary

The number of polls for each state are relatively small on a weekly basis (monthly totals for each state are presented in other posts), so we do not show them graphically.

Nonetheless, a summary of the weekly progression of the polls in October is as follows:

CO:  Obama’s share of women’s votes in CO have been extremely stable at 54-55 percent since October 7th , with male vote stable at about 45-46%

NH  polls have been somewhat erratic. With that caveat, it is the only one of these four states that shows some signs of an increase in Obama’s share of the female vote: from 51-53 percent in the first half of October (which was a major drop from September) to 57-59 percent during the last two weeks of October (which approaches the 60 percent among women that Obama enjoyed in late September and the first week of October).

OH:  the gender race in Ohio has been rock solid stable at 56 percent Obama among women and 47 percent among men, with the overall Obama vote in Ohio between 51-52 percent since October 7.

VA:   There is a mixed pattern in VA, with the Obama female share fluctuating between 54 percent (Oct 27)  to as high as 58 percent (Oct 20), but the overall Obama vote in VA has been stable at 51 percent.

Summary Statistics for this Post

Number of polls and respondents by week (four state totals)

week               #polls    #respondents

Sept 30-Oct 6     6          6397

Oct 7-13          17        14980

Oct 14-20        13        11186

Oct 21-27        22        16512

Oct 28- Nov3     3          2192

Total    61        51267

Number of polls in October, by State

state     #polls   #respondents

CO       11         9876

NH       10          6228

OH       23        20106

VA       17        15057

Total    61        51267

Update in response to questions: polling organizations represented in this post

[number of polls and total respondents]

ANGUSREID         1    550
ARG                          6    3600
CNN/ORC                1    722
FOX                           2    2257
GRAVIS                   3    4111
GROVE                     2    1000
GarinHart                1    807
Gravis                        1    645
LAKE                          3    1400
NBC/WSJ/Marist   10    11407
NEWSMAXZOBGY    2    1674
PPP                               11    8624
PURP                             3    1800
QUINN/CBS                1    1548
QUINN/CBS/NYT    2    2542
SUFFOLK                     2    1100
SVYUSA                       3    2030
TIME                               1    783
UNH                                2    1318
WashPost                      1    1228
YOUGOV    3    2121

Total    61    51267