Science, Creation & EvolutionSetterfield againHi Barry I fitted both data sets and there were only small differences between the results. The values from my last post derived from the bigger data set. I neglected the value from 1675 due to the missing error and the following values Auwers (Wanstead) 1727-47 Busch 1727-47 Bessel (GO) 1750-54 Peters 1750-54 since in those cases the table wasn't clear about the correct date. For the fitting I used a commercial scientific data analysis software (Igor Pro). There is a demo version available, if anybody wants to redo the fit. For data with such an high error the result is as clear as it can be. There is no possibility to deduce a variable speed of c from the data in the magnitude you proposed. There is no reason to divide the data arbitrary into different time periods. Looking at the graph of the Pulkova subset there is no obvious decay. It looks like random noise around a mean value. The best linear fit results in a 6km/s per year decay with a reduced Chi Square of around 3.5 and would be in agreement with your hypothesis. If you overlap the Pulkova data with the rest, this linear decay becomes much too high, however. Fitting with a constant results in a Chi Square of 5, which is slightly worse but would also be in agreement with the data of the subset. It's worth mentioning that in general small subsets with an high error are very sensitive to single points you include or exclude in the fit. For example you can get more or less arbitrarily any value for a c decay between 0 and +/-20 km/s per year just by the removal of some points. |
🌈Pride🌈 goeth before Destruction
When 🌈Pride🌈 cometh, then cometh Shame