I read this story on my PhysOrg news feed today:
The upshot of the research with this one is that a doctor's 44-minute journey on his 13.5kg, £50 bike would be made faster with his new 9.5kg, £1000 bike. He realized that it was actually only a one-minute decrease in time to 43 minutes.
As a scientist, he did a randomized trial on his commute (flipping a coin each morning to choose which bike to ride), and found that there actually wasn't any major difference in the time that it took him.
I had been wondering the same thing: whether it was weight that was a more significant factor than components to shaving off commute time. I had kind of made up my mind that - based on the marvelously improved cycling speeds and times that I have been able accomplish with my current bike over my old Giant Sedona that it is primarily due to the components, since they both weigh about the same (and my current bike, with its panniers, often is heavier than how I normally rode the Sedona).