From here: http://www.talkorigins.org/indexcc/CE/CE411.html
Setterfield chose 120 data points from 193 measurements available (see [Dolphin n.d.] for the data), and the line of best fit for these points shows the speed of light decreasing. If you use the entire data set, though, the line of best fit shows the speed increasing. However, a constant speed of light is well within the experimental error of the data.
How can Setterfield expect from us to take his work serious with this kind of research??
Some more from the same source:
If Setterfield's formulation of the changes in physical parameters were true, then there should have been 417 days per year around AD 1, and the earth would have melted during the creation week due to the extremely rapid radioactive decay. [Morton et al. 1983]
As to Mercury going faster: Perhaps the distance of Mercury from the sun changed?