Last week, Hillary Clinton recounted Vladimir Putin's actions in Ukraine and commented "Now if this sounds familiar, it's what Hitler did back in the 30s." This was duly reported as a gaffe and Clinton "clarified" her comments the following day: "I just want people to have a little historic perspective. I'm not making a comparison certainly, but I am recommending that we perhaps can learn from this tactic that has been used before."
Most polls, including ours, don't attempt to put questions into "historic context." We tend to ask questions and let people put them into whatever context they want. We asked whether "the U.S. should get involved with Russia's dispute with Ukraine" without asking whether they know where Ukraine is, what the dispute is about, or anything else. Would it make any difference if some comparisons – of the sort made by Hillary Clinton – were suggested?
We tried a simple experiment this week. We asked one half of our survey takers the standard sort of questions about U.S. involvement in Ukraine: "Do you think the U.S. should get involved in Russia's dispute with Ukraine?" followed by some specific things that the U.S. might do (hold diplomatic negotiations with Russia, impose economic sanctions, provide economic aid, and some military options).
To the other half of the sample, we first asked "Do you think Vladimir Putin's actions in Crimea today are similar to what Hitler did in Austria and Czechoslovakia in 1938?" and "Would you consider it 'appeasement' for the U.S. and other western democracies not to take strong action to defend Ukraine?". These aren't our normal types of questions. They are quite a bit less nuanced than Secretary Clinton's comparison (or non-comparison) of Hitler and Putin, but we weren't trying to be nuanced. Then we asked these respondents the standard "balanced" questions about U.S. involvement in Ukraine. (We also asked these questions of the rest of the respondents, but only after they had told us what they thought the U.S. should do in Ukraine.)
Does context make a difference? Well, yes it seems to. Only 21% of those asked in the conventional way favored U.S. involvement in the Ukraine. When this question was preceded by the questions about appeasement and comparing Putin to Hitler, support for U.S. involvement rose to 29%. It didn't change the overall result -- a majority of Americans still oppose getting involved in the Ukraine even after the parallel to 1938 is mentioned -- but it does make a difference of about 8%.
It's also interesting to see how the comparison with Hitler raises (or doesn't raise) support for some specific actions. It does not, for example, appear to have much impact on support for military intervention. Even after mentioning Hitler, only 7% favored sending troops (compared to 5% in the normal poll) and 13% favored providing weapons (compared to 10% without the cue). The suggestion of appeasement did appear to raise support for economic sanctions and financial aid (by 7% and 8%, respectively) and slightly dampen support for negotiations.
We try not to take sides on issues or to argue with respondents in our surveys. We didn't say that Putin's tactics were like Hitler's or that Obama was an appeaser. But merely asking such questions is enough to move public opinion as it is conventionally measured.
Full results can be found here.
Economist/YouGov poll archives can be found here.