I've always found all that research a little vague.
- What kind of product/art/content were used in the research?
- Did the visitor have a specific goal or task to complete?
- What is the environmental context?
These kind of factors is crucial to the understanding of the results.
In the end, it depends on a combination of the quality and/or glossiness of the content, the eagerness of the user, and the amount of surrounding distraction, among other things.
Also, you can get "impressed" by a site and then throw it away all the same.
What you're trying to do in the end is pulling the visitor/potential customer in, so it's more interesting to examine bounce rates and average time before bounce. Those will most certainly differ from site to site, depending on the landing page itself.
Also, if the bounce rate is low while the "time until bounce" also is low (i.e. the few users that do bounce does it quickly), is that a bad thing? If the bounce rate is low but the "time until bounce" is high - is that better or worse?
All sites are different and the only really valid statistics is between the old version and your new version of your site.
Comparing different sites is a little like comparing the box office sales between 'The Avengers' and 'Annie Hall'. And yes - I like them both. :-)