Citation

Etienne J. Safety Sci. 2008; 46(10): 1420-1434.

Copyright

(Copyright © 2008, Elsevier Publishing)

DOI

10.1016/j.ssci.2007.10.002

PMID

unavailable

Abstract

The field of organisational reliability analysis lacks proper qualitative methods in order to interpret data on normal operational organisations and produce diagnoses regarding their reliability. Due to the peculiarities of catastrophic events, and especially their low frequency and multi-causality, such interpretations and diagnoses could not rest on the sole data one may collect about normal operations. Rather, it can only rest on the theoretical knowledge stemming from post-accident studies. Hence normal operations studies require methodological tools in order to regulate knowledge transfer from post-accident cases to normal operational cases. Such tools need not only to be practical, they should also fit with the peculiarities of organisational factors and processes. I argue that comparisons between actual and counterfactual normal operational organisations can be a fruitful although not ideal means to interpret data on normal operational organisations and hence to assess their reliability. This procedure enables to import arguments from post-accident studies and to combine them with empirical data on normal operational organisations in order to produce relative measures of reliability. Organisational complexity sets inescapable limits to this method, but the latter consists nevertheless in better epistemological criteria for qualitative reliability analyses than there existed before.