Originally Posted by WiredGuy

I've done some really specific benchmarking where I purposely took control sites with minimal backlinks which had really high bounce rates of 90% or so and then replicated a similiar user experience (I won't go into much detail on what I did here) where I prolonged the retention times and reduced the bounce rates from all traffic sources, Search/TypeIn/Referral traffic, and got my analytics in the 40% bounce rate and 45-60 sec hold times. The results took almost a month to realize but with all other factors being equal, it was quite apparent that the analytics information was affecting rankings.

All page content stayed the same, no increase in links. The only variable difference was the new traffic I had introduced to alter the analytic information.
WG