The link below has info related to the approach for testing/verifying the assumptions behind the Arrhenius model.
(Similar to Meeker's approach.)
One of the important bits is to conduct tests of the material at several temperatures, as the Arrhenius relationship assumes a (transformed) linearity, which can't be verified with only two points (temperatures).
http://www.weibull.com/AccelTestWeb/arrhenius_relationship_chap_.htm
See also Meeker section (tbd) and Nelson (1990, pages 521-544)
See also NIST material on Degradation at http://www.itl.nist.gov/div898/handbook/apr/section4/apr423.htm
This method is related to, and uses principles of, regression.