The total time between when the actual research completes and the end of the increment is all wasted. Its not so important at low tech levels but becomes much more important later on.
I'm watching a let's play where the guy is continually using 30 day jumps. He almost always has no research queued other than the current set of projects.
It seems to me like he's losing research points that could be going toward projects if only he'd queue them up. Is that correct or does the interrupt scheme save him?