| The Investigation Process Research Resource Site |
A Pro Bono site with hundreds of resources for Investigation Investigators
|Home Page||Site Guidance||FAQs||Old News||Site inputs||Forums|
to advance the
Search site for::
Launched Aug 26 1996.
A comment on: Integrating Human Factors: The Future of Accident Investigation by Captain Robert L. Sumwalt, III (F03396), Air Line Pilots Association, Team Leader, Working Group for Human Factors in Accident Investigation.
The paper may be read at the following address:http://www.awgnet.com/safety/library/isasilib.htm
One of the keys to this paper is given in the conclusion of another paper posted at the same address, same date, by an author who is quoted in Sumwalt. It sets the perspective of Human Factors : Aviation Human Factors and the Safety Investigation Process, by Captain Daniel E. Maurino (FO3290), Coordinator, Flight Safety and Human Factors Programme, International Civil Aviation Organization (ICAO):
If error is indeed a cause, it needs not be a causal "factor". If error is but a factor, it needs to be qualified and hierarchized against other factors. But if error bears directly on the accident, then it is the foremost element to be remembered concerning an accident. According to the sources referenced by the author, the latter prevails since this [error = causal factor] occurs in a large majority of aircraft accidents and incidents [...] in 60 to 80% [of] aviation mishaps.
Now in addition to error = causal factor, the author prepares a slip between mishap and accident where error characterizes so many accidents, on the following scale:
The second founding statement in this paper states that almost any accident investigation is, in essence, a human factors investigation and almost all accidents have causal links to human error. At this point, the author is certain to hit the target in 60 to 80% of occurrences if the error = causal factor is allowed to hold true.
Note: For example, if a crash was initiated by structurally worn front cargo doors, to what significant extent could the author associate a human error = causal factor?
The author goes on to mention occurrences that bear - in my understanding - signatures of error such as design error, [...] inspection error [...] installation error [...] weather dissemination errors [...] errors of decision-making [...], including a terrorist act which is often the result of some error some where in the system.
The signatures of error notion, if it is intended to qualify human error, is really a matter of personal style, and one can wonder if the notion has any future in professional investigation (unless signatures of error is another way of evoking lie detectors), especially if the error(s) are not obvious at all [and/or] may be deeply embedded within the system, perhaps far, far away from scene of the accident site.
Those who wonder how to interpret any set of directives for investigation will be enlightened at the following mission statement, which is not a task description: An accident investigator's mission is to identity the factors [...] root causes which may not have been recognized previously.
When the author digs deeper into human error, human factors, scientific methodology, objectivity, insight, etc., he puts them in opposition with the "operational perspective": Human factors has a basis in scientific methodology that can lend objectivity to the investigation and provide insight not obvious from a purely operational perspective...
However, when advocating the human error comprehensive approach, the author seems to limit its scope to mishaps and causal factors: human factors must be fully integrated into accident and incident investigations to ensure identification of all relevant factors leading to the mishap [...] providing a very basic human factors education for all involved in mishap investigations...
As is the case in many other investigative specialties, the human error approach offers an advantage over other approaches. Arson investigators know what they are looking for at the outset of an assignment, and so do human error seekers: ...without a clear description or knowledge of exactly what you are seeking [...] the task of finding something becomes more assured when you have a clear understanding and knowledge of what you are seeking. The same applies to mishap investigation. [...] meaning that the staff sees its value towards logically explaining mishaps.
With little warning, Captain Robert L. Sumwalt, III switches from mishaps and the intricacies of human error and the collection and analysis of potentially relevant human factors information to a crash site: Consider the field investigator who picks up a bent piece of metal on the accident site. She or he may be able to examine it in the field and conclude... It should be agreed that such an occurrence is not impossible, but it is not the rule and certainly not a recommended method. A bent piece of metal - one piece, that a single person could pick up and handle - cannot yield a cause or a causal factor...
Had the paper remained focused on mishaps, for example, it would have been correct to say that the discovery of human error should be considered as the starting point of the investigation (ICAO's Dan Maurino suggests...).
The latent accident / mishap
The word latent, which seems to be quoted from the Reason Model has a meaning of its own, and the meaning is well associated with definite notions - and latent is but an adjective (in the American Dictionary of the English Language, 1981, p.739):
The "latent image" which is of concern to photographers is not mentioned in this edition of the American Dictionary of the English Language.
Sumwalt associates latent and dormant with conditions, and he further associates latent with failures: Reason Model [...] says that accidents often result from the interaction of a series of flaws known as latent conditions that have been embedded in the system. [...] Latent conditions or latent failures are systemic flaws whose consequences may not surface until long after being introduced into the system. These dormant conditions...
In any case, if the Reason Model or theory is to be accepted in the future, it should provide, - each time such latent condition or failure is discovered -, the three following components in its demonstration:
.1 the initial stimulus that has embedded the condition or failure in the system
Hugh Chicoine tpi, CFEI, CFII (NAFI, NFPA)