Varying radioactive decay constants: Are we missing something ?

From Wikipedia:

The current scientific explanation for the Earth's temperature gradient is a combination of the heat left over from the planet's initial formation, the decay of radioactive elements, and the freezing of the inner core. Another possible explanation includes the georeactor hypothesis, though this theory has not gained acceptance in the scientific community.

Hence, if the radioactive decay constants vary with the location in the solar system, then the core of our planet is heating anisotropically: i.e. there are places that are colder than others, most probably inducing convection movements at the center of the Earth. Wouldn't it have been noticed by now ? Let us also remember that Radium-226 that is studied in this article is a decay product of U-238, and we know there is something fishy about the ratio U-235/U238.

A reader on Slashdot suggests that a bias might come from a shift in frequency due the power load on the electrical grid:

...As someone who made the equipment that the scientist probably used to do the counting, I have one possible explanation. Most Multichannel Analyzers (MCAs) of the time used a line clock to determine the time. They assume that the power company delivered 60Hz power (or 50 Hz in Europe), This frequency was almost never precise but varied by .1 to .2% (one plant where I measured the frequency put out 58.8Hz for example, a real mess for us) from time to time. A systemic variation due to power loads (heating in winter/ AC in summer) could easily bias the power frequency by about the right amount with the right periodicity. The universe might well be safe...

This is not entirely far fetched. One of the thing that one can notice in the graphs of the article are the apparent shifts between the oscillations. In order to get a better idea, I went ahead and downloaded some of the data for the French electrical grid in 2007 to check this potential bias mentioned by the slashdot commenter and as one can see there is indeed some peaking occurring away from January 1.

Furthermore, in the article, one can see the following item:

This possibility is supported by the data we report in Ref. [18] in which we present evidence for the possible detection of a change in the decay rate of 54Mn during the solar flare of 13 December 2006. As noted in Ref. [18], the coincidence in time between the change in the 54Mn counting rate and the solar flare, along with other observations, is consistent with a mechanism based on a change in phi_nu during the solar flare.

Again, this is also entirely consistent with how geomagnetic storms influence the power grid and one of the reasons, the NOAA provides information on this to electric utilities.