Monday, August 16, 2010

You Can Argue But You'd Be Wrong

In spite of grave misgivings, I started to read this post by Efraim Karsh, figuring that hey, it's important to educate myself on the views of others -- even if they do say ridiculous things like that Obama has "distinguished himself, in his short term in office, as the most anti-Israeli U.S. president in living memory."

And then we got here:
James Zogby himself, among others, felt compelled to attempt to rebut my article. "There are bad polls, and then there are bad interpretations of polls," he wrote in the Huffington Post. "Putting them together (i.e. a bad interpretation of a bad poll) can create a mess of misinformation."

The "bad poll" in question is a recent survey for the al-Arabiya television network, noted in my article, which found a staggering 71 percent of Arab respondents had no interest in the Palestinian-Israeli peace talks. And the "bad interpretation" is my presumed failure to recognize that this was not a fully scientific poll but rather an "online vote," which didn't refer to the Palestinian-Israeli peace talks but rather to the "Middle East peace process."

It is arguable of course that an "online-vote" by 8844 respondents (more than twice the size of the Brookings/Zogby poll), answering one straightforward question, might be more accurate and less susceptible to manipulation than "scientifically" crafted surveys purposively choosing their target audiences; or that ordinary Arabs, living as they do in one of the least democratic parts of the world, will be more candid in the relative obscurity of the web than in the presence of a pollster knocking on their front door or contacting them by phone.

And ... we're done here.

Because even I, B+ student in Statistics for Half-Witted Morons that I was, know that it is not, in fact, arguable that online polls (no matter how many people respond) are more accurate than a scientific poll. In fact, the entire polling industry got its start because George Gallup proved that scientific polling was far more accurate than even data drawn from a giant, non-random response pool (in his case, the Reader's Digest polls).

Anybody who is more statistically illiterate than I am really can't be trusted on any topic that involves gauging public opinion. Moving on.

3 comments:

N. Friedman said...

I think that Karsh was indicating that he does not think highly of polling that is supposedly conducted in a scientific manner but which, in fact, is skewed to obtain the results intended by the pollster. If that is what he intended, then he has a point although, clearly, you are correct that a non-scientifically conducted poll is not valid.

Let's be honest, you were not, as I see it, looking at what others think. You were, instead, looking to find something a knife to put into his view. Otherwise, you would have a comment on the rest of what he writes.

As for Karsh's scholarship, his latest book, Palestine Betrayed, is worth your time. It, more than any other book of recent vintage or, to my knowledge, before, depicts the views of Arabs to the dispute and the debates that occurred on the Arab side. Such is completely missing in, for example, Benny Morris' brilliant book, 1948.

So, this is not a question of opposition. It is a question of examining all sides of an issue.

David Schraub said...

Not really. Someone emailed me this article. I started reading, and stopped six paragraphs later when he talked about how maybe a non-scientific poll conducted on a website is better than actual, scientific polling. Thirteenth chime that calls all the rest in question and all that.

N. Friedman said...

David,

You fail to take into account how skeptically public opinion polling from that part of the world is viewed by scholars. And, in fact, Karsh was making fun of the polling, noting that it was no better than a non-scientific poll.

He goes on, perhaps below the point you read to - to note that his opinion was not based on a poll at all, something that, had you read on, you might have noticed. What he wrote, which is consistent with what other scholars have written over the course of decades, is:

But whatever the scientific merits and flaws of certain polling techniques, this issue has no bearing whatsoever on "The Palestinians, Alone." For, contrary to Zogby's claim, my contention that the Arab world has never had any real stake in the "liberation of Palestine" is not based on my reading of the al-Arabiya survey but on the long history of systematic Arab abuse of both the "Palestine Question" and the Palestinians themselves. A poll, even in the best of circumstances, can only give a fleeting glimpse into reality, which is what the al-Arabiya poll did; a historical survey, by contrast, can put current circumstances within their far wider and deeper context, which is precisely what my article did.

Were you to read his book, Palestine Betrayed, or Benny Morris' book, 1948: A History of the First Arab-Israeli War, or any host of other good books about the Arab-Israeli dispute or, were you to read a good history of Arab politics in general, such as Elie Kedourie's stellar work, Islam in the Modern World and Other Studies, etc., etc., you would see that Arab interest in the Palestinian Arab cause has over the years, at best, been window dressing and that while Arabs do care about the matter, they could care less about Palestinian Arabs as such.

Were you to read Karsh's book, you might be aware of just how contemptuously other Arabs held the Palestinian Arabs - and, most especially, their leadership - in the run up to the 1948 war, with the various countries fighting for their own, not the Palestinian Arab, cause.