Steve,
Based on previous discussions, I expect you (and others) may be fascinated by this article (if you have not read it already) on expectation bias (especially the story of Einstein), which is one explanation of why the interpretation of data is important and how you can get different results using the same methodology.
[
arxiv.org]
There are many ways to arrive at differing determinations based on NFPA 921. The methodology (use of the scientific method) was not used or in error, all of the data collected was not used to test the hypothsis (selective, did not think it was important, or was not capable of interpreting the data), fundamental knowledge of the behavior of fire was not used to test the hypothesis, right analysis for the wrong fire scenario, and/or a too simplified analysis that does not take into account a significant aspect of the problem. The investigation of fire and explosion events is a difficult task at best. I think a root cause of the interpretation of data may be the over-simplification of the fire phenomenon itself.
Douglas J. Carpenter, MScFPE, CFEI, PE, FSFPE
Vice President & Principal Engineer
Combustion Science & Engineering, Inc.
8940 Old Annapolis Road, Suite L
Columbia, MD 21045
(410) 884-3266
(410) 884-3267 (fax)
www.csefire.com