While the excellent analyses above stand on their own, I should point out that an experiment of this nature was actually carried out in
Britain in 2001.
1 Schoolchildren across the country all jumped (from the ground, to avoid the distance problems of uneven chairs) at the same time, to determine what result (if any) there might be. While the
Earth was not noticeably affected, researchers in Britain noted that local
seismographs recorded the event, with the strength of (approximately) an earthquake of magnitude 2 on the
Richter Scale.
(Note: m_turner informs me helpfully that a M2 earthquake is equivalent to the energy release (roughly) of 1 ton of TNT.)
We might, therefore, attempt to scale up from the number of participating children in Britain (something around one million) to the number of Chinese. There remains the problem, however, of arc coverage and of coordination. In our case, we are looking at the local vibratory effects of the Jump, not the aggregate effect on the Earth's orbit/spin. This makes things harder; we would need to know the number of Chinese, where they were jumping in relation to each other, and how much they weighed! This is because of the effects of interference.
Assume the speed of sound in earth is speed s. This is the speed at which the vibrations from each Jump would propagate outwards from the spot of each Jumper. Now, if the jumpers are the correct distance apart, it is quite possible for the vibrations to meet when precisely out of phase with each other (that is, when the period of one wave is 180 degrees from the other). If this happens, physics tells us that for the zone of interaction where this condition occurs the net result will be zero - it will be as if no wave passed at all.
So, to really be sure, we'd need to know the average separation distance of the Jumpers' feet, the speed of sound s in their floors, and where (exactly) they were all standing. Otherwise, there is a very small but non-zero chance that when they jumped, they would cancel out each other at the surface point of measurement!
1: The reference is from Slashdot, posted September 7, 2001. Unfortunately, it was just a pointer to a Yahoo! news page which has expired; I'll endeavor to hunt down the original story.