Partly Facetious: how can you measure popularity under emergency?

"Brown's ratings are as low as 23%."
"So is Bush's."
"That's true but they are high by the standards of our political leaders."
"I don't think even Musharraf's is as low as 23%."
"Well, you can't take a survey during emergency."
"Who told you that?"
"A little bird told me that."
"Don't be facetious."
"I am not being facetious. I mean you have the Pakistani educated and the liberals out on the streets protesting against Musharraf's emergency and the firing of the judges, press censorship and all and you really think he will allow results to be broadcast that would show his true popularity rating in this country!"
"Oh, I see where you are going with this. But in any case you are wrong. I read somewhere that Musharraf's popularity ratings are as low as Bush's - maybe a couple of points lower."
"Well, that's surprising. Who did they survey?"
"Am not quite sure and furthermore am not sure where the survey was carried out. See it's normally a randomly selected sample but in this country nothing is random. I mean are the bureaucrats truly random - they are so used to voting for the 'right' person with no means of knowing whether the election officials have not been instructed to check who voted for whom."
"And you may think the same applies to members of the armed forces?"
"Well, they are state employees too, aren't they? I mean they may think they have a right to rule us but it's our tax money that pays their salaries."
"But even then state employees form a very small part of our population."
"Right and then you have outright rigging."
"You reckon rigging even in a survey?"
"Ah, there you have the new and improved democracy at work."
"Don't be facetious."