1,500 out of a population of 12 million, is that a good sample size for these kinds of things?

a sample of 1,600 people gives you a margin of error of 2.5 percent

how the sample is selected is more likely to be a concern than the number of people sampled

A more expansive source polling is the Asia Foundation poll with ACSOR (who did the fieldwork for both) which ‘operationally, 648 field surveyors throughout all 34 provinces of Afghanistan interviewed 6,406 adult Afghans between June 17 and July 6, 2009′ is a better measure and unsurprisingly rather more mixed than this smaller sampled media commissioned poll (although funding still in part came from interested occupying parties).
http://asiafoundation.org/country/afghanistan/2009-poll.php

Given there is this better source than the smaller poll I wonder why choose that and why headline it as has been done in such a simplistic way.

I’m not sure why anyone would be surprised by most of this opinion poll – after all, both the Taliban and the various warlord factions are an extremely violent and crazy bunch, prone to summarily executing people on the flimsiest of pretexts. The Americans will pretty much leave you alone and focus on looking for nutters with machine guns, provided you’re not having a wedding party or making any sudden movements. A few decades of ultraviolent civil war and any semblance of order and normality looks wildly desirable, even if it’s being imposed by Godzilla, which must make the US armed forces look golden to a majority of Afghans.

The problem, of course, is that even a sizeable majority isn’t going to cut it. There’s obviously a significant minority who are still at very least willing to co-operate with the Taliban. If there wasn’t, it would be impossible for the nutters to operate any kind of insurgency at all, let alone one that has stalemated the world’s most terrifying military machine for almost a decade. That’s before we get to Karzai’s approval ratings, which strike me as weird beyond belief, given that he’s just been slaughtered for stealing the election. That implies that either a) something very fishy is going on with these figures or b) the Afghans are rather more concerned about being able to walk down the street without being executed by the beard patrol than they are about democracy. I’d go for the latter, myself.

So, some gratifying results, some deeply weird. All we need to do now is become the first occupying force in military history to wipe out a massive insurgency without having to exterminate a large percentage of the populace, and we’ll be laughing all the way to freedom, peace and ponies.