Welcome to WebmasterWorld Guest from 188.8.131.52
As we have been finding out, it depends a lot on the statistical base that you are working with.
I agree with a post made by europefor visitors earlier that you need several thousand impressions to base a conclusion on.
How many (order of magnitude) impressions are the banner and skyscraper examples based on.?
Btw, has anyone noticed a little decline in the overall pageviews per day of your site. Previously, if a surfer visited 4 pages, now its cut down to 2 or 3, probably because the ads are really targeted and the surfers are incentived to click on them. Not a problem really, but I wish they would provide the option of target=_blank :)
Purely based on conjecture:
Do you think the difference between banners and skyscrappers is that more people are "ready" to leave a page half way down than are willing to leave as they arrive at the new page (im assuming your banners are at the top of the page)?
Ie, as they arrive, they havent found the info they are looking for, so unless a banner at the top is highly relevant (more relevant than your page appears), then they are less likely to click through.
But, halfway down, as they have begun to digest your content they may be more inclined to "impulse click".
Granted, a wild theory, entirely dependent on subject area and relevancy of ads being displayed.
Interesting theory, that sounds as if it might be plausible.
If it were true, then moving banner ads progressively down the page should give you the statistics...problem is, I can see no way to do this, and get valid results, unless you only have the ads on only one page and that page delivers at least a couple of thousand impressions per day.