How Much Does Public Opinion on Syria Matter? How Much Will It?


The chart above comes from Google Trends, and the message is clear: even at the height of the US’s ostensible march towards military engagement in Syria – remember, all signs pointed towards military action last weekend until Obama’s surprising decision to seek Congressional approval – more Americans were using Google to get information about Miley Cyrus than Syria.

I was moved to compare the two after some discussion on Twitter this evening prompted by Larry Sabato’s reporting of a Reuter’s poll showing support for military action in Syria hovering somewhere around 20-30%.

The question I want to raise is the extent to which this might ultimately matter. There are two ways to think about this. The first is whether public opinion is going to influence whether the US actually launches hostilities against Syria. Here, I think public opinion mattered in so far as it may have played a role in getting Obama to seek Congressional approval in the first place – although personally I think the lost vote by Cameron in the UK was probably more important – but at this point the ball is probably in Congress’s court. It is certainly possible that a huge swing in public opinion could have an effect on the forthcoming vote, but my guess is 20% support vs. 40% support doesn’t make all that much difference at this point. The Senate is likely going to vote to approve in any case, and the House dynamics are going to follow district level concerns more than national ones. I’m sure we can find a few people who might flip because of trends in national public opinion, but I’d need to be convinced by someone who knows more about individual US legislators than I do that this could actually swing a vote.

But the more interesting question is whether the low public support for military action would actually have an effect down the road on either Obama’s ability to govern or the political fortunes of individual legislators. Here I am skeptical – conditional on this being the limited, aerial engagement that is being discussed now and not having some unexpected escalation occur – that Americans will actually not care all that much in the future if Obama launches a limited number of missiles at Syria. (hence the teaser figure above).

Consider the following thought experiment: if Benghazi had not taken place, then would there be any discussion of the rest of the Libyan events playing a role in American public opinion today? And Libya will – most likely – end up involving more military involvement than seems likely in Syria.

Then perhaps ironically, from a public opinion standpoint what now seems to matter here is the fact that a vote actually has to be taken in Congress more than the action that might result from that vote. As Sabato pointed out to me in a conversation we continued off of Twitter, it is possible that the vote could cost certain incumbent House Republicans in the coming 2014 primaries. And the point has been repeatedly raised that it would be a major political setback for Obama were he to lose this vote in Congress. So in a sense, we go through the looking glass: foreign policy has the potential to “matter” from a public opinion standpoint only because it has been converted into domestic politics: Is Obama losing strength in the second term? Are certain conservative members of the House conservative enough? My guess is these questions will reverberate in American political discussion longer than whether it was appropriate for Congress to authorize (should it choose to do so) and the President to execute low-scale military action despite underwhelming support on the part of the American public.

11 Responses to How Much Does Public Opinion on Syria Matter? How Much Will It?

  1. Andrew Gelman September 4, 2013 at 3:32 am #


    There’s one thing I don’t understand in your post. You write, “It is certainly possible that a huge swing in public opinion could have an effect on the forthcoming vote, but my guess is 20% support vs. 40% support doesn’t make all that much difference at this point. The Senate is likely going to vote to approve in any case, and the House dynamics are going to follow district level concerns more than national ones.”

    But a national difference of 20% in support would correspond to large differences in individual districts, no? An average of 20% per district, which could make a difference in a lot of places, I’d think. So I don’t see how your argument works, when you say that changes in public opinion wouldn’t matter for Congress.

    • Joshua Tucker September 4, 2013 at 9:05 am #

      Andy: My main thought was that – like Tom M. noted – we have so few House districts that are sufficiently competitive these days that legislators are more likely to be motivated on an issue like this by views of likely primary voters than by district as a whole. So if more moderates in a heavily conservative district swing behind the attacks, is that going to be enough to compensate for accusations of collaborating with Obama for a R legislator potentially facing a Tea Party challenger?

      Also, I wonder just how much legislators think voters will weigh a vote on a foreign policy decision. If the Senate is already likely to support at the 20% support levels Reuters is reporting, then that suggests there is already some discounting going on. It may be the case that there are legislators carefully watching opinion swings in their districts, but my guess is (a) number is small and (b) it would be less than for a domestic politics issue. But the point I was trying to make at the end was that to the extent that it may be going on (thus as you noted correctly undercutting my point about national swing not mattering), it would be because this has somehow been redefined into a domestic issue, i.e. “stopping Obama”.

  2. Jon M September 4, 2013 at 3:41 am #

    I’d be careful about reading anything about public opinion into Google Trends data unless the particular term you’re using has been validated against a representative source of data such as surveys.

    I’ve found that many plausible search terms are completely uncorrelated with public attention on issues (see

    Your conclusions may be correct but it could also be that information on Miley Cyrus is more accessible through Google searches than information on Syria.

    • Joshua Tucker September 4, 2013 at 9:09 am #

      Jon M: You may be correct, although I would guess it was the opposite – that there was more information on TV and other sources about Miley Cyrus than Syria. But just to be clear, I was using Google Trends as an attempt to measure interest or attention, not “opinion” in the sense of being for or against intervention. Thus the measure is not comparable to the Reuters data.

      But point is well taken that Google searchers are likely not a representative sample of the population. But they probably are a large segment of the population. Have you ever looked for data to see the numbers on this? I suspect there must be a paper on the representativeness of Google search users somewhere. And I would suspect in this regard, Google searchers probably look like Bing searchers as well.

      • Jon M September 4, 2013 at 5:13 pm #

        I remember reading some figures somewhere that said that Google searchers are fairly representative.

        The problem’s actually not so much that Google users are unrepresentative but that you can’t distinguish between heavy searching by a small subgroup. Usually the issue is with news junkies and monkey cage readers searching heavily for issues that rest of the world doesn’t care about, but here it’s perhaps the celebrity gossip junkies we need to worry about. The data are consistent with either a small portion of Americans being hugely obsessed with the Miley Cyrus story or the whole country following avidly.

    • Jon M September 4, 2013 at 5:49 pm #

      This discussion prompted me to finally write up a blogpost about a simple fix Google could make to Google Trends: measure the proportion of people who search for a term not the number of searches for it

  3. Tom M September 4, 2013 at 8:23 am #

    There may be some validity in your statements re: an uptick of 20%, but a LOT of districts are so polarized it may not make as large a difference as you think.

    I agree with your concerns, although I suspect information on Syria is fairly plentiful on the Web, too.

    What I find telling is that ANY “Conservative Republican” Congress(man) could in “good faith to his ‘constituency’ ” NOT vote to approve the measure. The Right has been trying to balance the budget on the backs of social programs while maintaining a military budget that outstrips how many of the next countries in the world? They are pro Big Business and Big Military.

    As for me, it is a lose-lose, but hope to win in the long run decision. We cannot repeat the Herbert Hoover and Neville Chamberlain Isolationist and Appeasement policies that led us to WWII. The War to End All Wars (WWI) showed us the horrors of chemical warfare, and the holocaust showed us the horrors of sticking our heads in the sand.

    Yes, this is a “Civil” War – an ongoing blood feud of one “family” against another that has gone on for millennia. That is not what we are entering into. We are exacting a price for daring to use Chemical/Biological weaponry against innocents – because it exemplifies a willingness to go beyond – to use or sell to those who would use same against “the great infidel evil”.

    To the author – you raise good points. As someone who has critiqued High School Forensics debates, and proof read both papers and stories for people, I winced and chuckled a bit as I read one line in particular – “Here I am skeptical – conditional on this being the limited, arial engagement that is being discussed”

    I know the Pen is Mightier than the Sword, but I don’t think the Armed Forces are going to use a True Type Font to wage war. “Arial” is not the correct word. Aerial is. I thought perhaps MS Word auto-corrected for you, but the online Spell Checker on this site is not happy with “Arial”, either.

    • Joshua Tucker September 4, 2013 at 9:12 am #

      Tom M: Thanks for the comments and for noting the typo – embarrassing (although the idea of a war fought by fonts may have some relevance in today’s cyber-warfare world)! It is corrected…

  4. Chris September 4, 2013 at 9:18 am #

    Love the Miley Cyrus data, but I agree with the point made above that the political implications of google trends search data are not very clear.

    On the question of whether public disapproval will matter if Obama goes forward. I would say “yes.” For example, here is a link to my APSA paper (co-authored with Joe Grieco) which demonstrates that low public approval for the President’s handling of Militarized Disputes significantly reduces his ability to get domestic legislation through Congress. I would acknowledge, however, that in this particular case our estimated effect may fall victim to floor effects.

  5. Chris September 4, 2013 at 9:29 am #

    Ooops – here is the link:

  6. Todd Phillips September 4, 2013 at 9:24 pm #

    It seems silly to ask whether Americans support hostilities against Syria without first asking whether Americans understand the situation well enough to make an informed decision on the issue. If they don’t fully understand the issue it is a GIGANTIC mistake to put any weight on their decision. Any expectation that uninformed people can or should make decisions about important issues is a recipe for disaster and it is against every American’s interest.

    It follows that it is a mistake to expect Americans to make decisions about political leaders in elections that they don’t truly understand. If they don’t understand them, and 99.999% of people don’t, it’s not because they’re dumb or lazy, but because it is an impossibility in a political world that is almost infinitly complex. Therein lies the problem with our system of government and why we have so many problems.