"The average Canadian has only one testicle"

It's an example of how misleading it is to average polling answers -- half of Canadians (the men) have two testicles, while the other half of Canadians (the women) have none. So when you average them, you get the ridiculous "answer" that the average Canadian has one testicle.
This news story uses this example to describe a critique of recent polling results about Canadian attitude toward the Israel-Hezbollah war.
Here's the dispute:

. . . the Strategic Counsel firm is standing by a recent survey that suggested only one-third of Canadians shared Harper's staunch pro-Israel stand. In the other corner, the head of the Compas firm says the prime minister enjoys twice that much support and accused his rivals Monday of conducting a "misleading anti-Harper poll.". . . . . . Compas said its rival invited an anti-Harper response by asking about "Israeli actions" - a term it decried as a hostile-sounding statement that swayed respondents.
. . . But the Strategic Counsel defended its two-week-old findings, and several other industry observers agreed the company had conducted a solid survey.

You be the judge:

Compas arrived at its conclusion that Canadians supported Harper after asking the following four questions:
-Does Israel have a right to defend itself? (82 per cent responded in the affirmative)
-Was Iran wrong to arm Hezbollah and call for the destruction of Israel? (69 per cent agreed)
-Was Syria wrong to arm Hezbollah and disobey the United Nations resolution requiring Syria to keep guns out of Lebanon? (68 per cent agreed)
-Did Hezbollah in Lebanon start the war? (Just 38 per cent agreed)
Compas then took those four responses, averaged them out, and concluded that 64 per cent of Canadians supported Harper's policy.
One industry insider sneered at that methodology. "That's certainly stacking the deck," he said. "Those four policies can't be (averaged). That's like saying the average Canadian has only one testicle." . . .
The Strategic Counsel concluded that only 19 per cent of Canadians believed Harper's position was a principled one, while 53 thought it was designed to mimic the U.S. stand.
Compas came to a very different conclusion. The firm asked respondents whether they believed the government's policy was designed because:
-It wanted to earn U.S. goodwill and protect Canada's economic interests. (21 per cent agreed)
-President Bush is a role model (12 per cent)
-Israel has a right to defend itself (19 per cent)
-Arab extremism is a problem (12 per cent)
-Hezbollah is terrorist (12 per cent)
-Syria and Iran are problems (4 per cent)
Compas then proceeded to add up the final three of those responses and come up with the figure 28 per cent - while leaving separate the President Bush/U.S. responses, which would have reached 33 per cent had they been lumped together.
Butler concurred with the industry insider's view that such methodology was unorthodox: "I agree with that remark entirely," he said.

The other problem is that the Compas poll didn't survey as many people:

The Compas poll of 500 respondents is deemed accurate to within 4.5 percentage points, 19 times out of 20. The Strategic Counsel polled 1,000 Canadians between July 27 and 30, and its findings are considered accurate to within 3.1 percentage points, 19 times out of 20.

I wonder which firm will be hired by the Conservatives to do their next survey?