Optimizing Wikipedia's Fundraising Banner Ads

I've found them far more obnoxious over time, personally. As the first commenter notes, this is also a legitimate concern--but there's no easy metric for obnoxiousness. (Clicks on the "hide" button perhaps?)

I have written a blog post on this with detailed explanation of the 'anchoring heuristic'.<p><a href="http://saperduper.org/post/293243288/wikipedia-donators-anchoring-heuristic" rel="nofollow">http://saperduper.org/post/293243288/wikipedia-donators-anch...</a><p>I also posted it on HN <a href="http://news.ycombinator.com/item?id=1007587" rel="nofollow">http://news.ycombinator.com/item?id=1007587</a>

I still don't understand why Google doesn't just support Wikipedia. That whole "knowl" thing didn't work out that well... it would almost be a feel-good thing for them. Hell, google.org could do it on behalf of the parent corporation as a charity move.

a) statistical significance?<p>b) average is probably not as good a measure as median, since I doubt it's normally distributed.

Are these banners really better than advertising? (What revenue would half the pixels, devoted to tasteful ads instead of a pledge drive, yield?)