This is my second post on this topic, wherein is a discussion on why survey wording matters. The last post dealt with the “Obama is quantitatively ahead of his GOP rivals spin,” which wasn’t technically true and would be irrelevant if it was. This one deals with an issue that’s a little more tangled and open to interpretation – but not much. The question is whether Israelis favor an attack on Iran without the prospect of gaining U.S. support, which they don’t.
This finding is being spun to show they don’t favor an attack over current U.S. public objections. The data shows, bluntly, the opposite.
The poll asked respondents: “There has been increased talk of a military strike by Israel against Iran’s nuclear facilities, even though the United States, the UK and Germany have advised against it. What do you think Israel should do?” Putting aside the negatively valenced wording (that “even though” could get dicey in a low-information environment, but maybe it’s a translation quirk), the results were:
“Strike Iran’s nuclear facilities, even without the support of the US” – 19 percent
“Strike only if Israel gains at least American support” – 42 percent
“Do not strike” – 34 percent
The spin is that under the current configuration – with the U.S. “advising against” a strike – only 19 percent of Israelis support the action. But that’s not what the question asked. It asked about U.S. “support,” against the backdrop of current U.S. “advi[ce].” It’s entirely possible that most Israelis believe the U.S. would still “support” a strike despite Obama’s current position, in which case the number in favor would be closer to 61 percent. And it turns out, per a subsequent question in the very same survey, that’s exactly what they do think!
Respondents were asked: “Given America’s recommendation that Israel not strike Iran, what do you believe the U.S. government’s reaction would be if Israel strikes anyway?”
“It would join the war on Israel’s behalf” – 27 percent
“It would support Israel diplomatically, but not provide military assistance” – 39 percent
“It would stay neutral” – 14 percent
“It would punish Israel by reducing its current support to Israel” – 15 percent
So most respondents would be reluctant to support an attack unless the U.S. ended up supporting Israel, but most respondents think the U.S. will end up supporting Israel. Unless there’s something very strange going on inside the crosstabs, where the people who think the U.S. would support an attack are also the ones who categorically oppose the attack, then there’s a clear majority for a strike.
The best way to aggressively spin the question, then, is to extensively describe the first question but not tell readers how the second question was worded. Want to guess how Media Matters posted the poll?
Israeli officials, including its foreign minister, have hinted that such an attack would be their decision and their decision alone. But a poll released today by the University of Maryland showed that Israelis don’t support that policy… Only 19 percent of Israelis polled expressed support for an attack without U.S. backing… while 42 percent endorsed a strike only if there is at least U.S. support… More than a quarter of those surveyed think the U.S. would join an Israeli war, and nearly one in four said the U.S. would give Israel diplomatic but not military support.
This is nicely done. It establishes plausible deniability by acknowledging that the second question exists, but it doesn’t give the full wording, which would contradict the rest of the post. Nonetheless, Israelis think the U.S. will support its democratic ally against a declared enemy of both countries, and on that basis they support a strike against Iran. That might be naive, but it’s what the survey shows.
Caveat: it could be the wording is different in the Hebrew. This criticism relies on Israelis functionally imagining that the actual American position is – stated or unstated, but in reality – “we recommend/advise you not to do this, but we’ll support you if you do.” That’s the interpretation most consistent with the question wording in English. If the first question in Hebrew somehow equated “advice” with “support,” or if there’s something going on with “gains” that implies prior support, then that would weaken this particular criticism. But equating advice and support creates other problems – e.g. given that 20 percent of Israelis are undecided on Iran, it’s bush league polling to frontload a question with global opposition, implying there’s probably a good reason for that opposition.