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
. . . .. . . . . . . . last updated 3/23/07


to advance the
State-of-the-Art of
investigations, through
investigation process

Research Resources:

Search site for::

Launched Aug 26 1996.

Critiques of Lekberg paper
Posted 26 Jan 1998

The following are reviews/comments/critiques about the Lekberg paper.

  1. H Chicoine
  2. L Benner

Back to Lekberg paper

Comments by H. Chicoine

ANALYSIS is, for many,
all there is to investigation.

A comment on: Different Approaches to Incident Investigation -
How the Analyst Makes a Difference

The root ANALYS* (standing for analys*is, analys*t, etc.) occurs 66 times in a text of approximately 3109 words.

And the words investigation / investigator occur once each, in the title and in one sentence.

The title suggests that an INVESTIGATION is conditioned by the individual APPROACH of the different ANALYSTS:

1) in the title:

Different Approaches to Incident Investigation - How the Analyst Makes a Difference

There may be different approaches conveyed in the methodologies used in the overall investigative process in incident or accident investigation. Indeed the different disciplines contribute their methods or methodologies, but why should THE ANALYST MAKE A DIFFERENCE ? This statement seems to offer a passport to arbitrary interpretation and cannot serve the quest for corroborated facts or for a stream of coherent links. In the investigative process, analysis is often an explanation between two established nodes or facts. An explanation will rest on established benchmarks of the investigation in question. Analysis is not interpretation. Of course, one analyst's view may be different from the opinion of another analyst or from the view of an engineer, their perspectives may well diverge because their interpretations belong to their respective disciplines. But, this is not investigation.

2) in a sentence:

Accident and incident analyses will be guided by several different factors: one is the purpose of the investigation;

Following several comments, by Ira Rimson especially (a leading IRR-EIR participant), and following questions often raised about the "approach", "purpose", "objective", etc., of an investigation, i.e. all that is likely to impart a direction to, or set an investigation in a "perspective", managerial, judicial, financial or other, I have come to conceive that all that is not productive methodology in the investigative process should be estranged from the investigative process and/or should be considered a constraint, recognized as such, and factored in. Granted, some constraints are unavoidable and must be adapted to, but we nevertheless need to identify constraints and acknowledge them for what they are. Recognizing constraints early in the investigative process is not different from recognizing gaps in a time line later in the process. Constraints can be factored in without distorting the entire process.

The single sentence in Lekberg's paper which addresses investigation says that the purpose of an investigation is one among several different factors, and that the analyses will be guided by such factor or purpose. I think that investigators are reluctant to superpose such guiding constraints on their work unless they nurture a preconceived notion of the result that will emerge.

The most common guiding principle invoked is prevention, which invites to bring some change to a piece of standard or regulation, or to procedures in air safety, fire safety, public safety, etc. Specialized personnel and experts such as in law enforcement organizations, university research centers, etc., who advocate their discipline yet address investigation in that perspective introduce biases that tend to shape or conform things instead of taking an imprint of the event.

As presented by Lekberg, analysis is not defined or qualified, but it imposes a "perspective". And the word "perspective" occurs 30 times in Lekberg's paper.

This is not an uncommon pattern because a similar twist exists in fire investigation and is sometimes (often) factored in the ignition sequence as follows:

the action / occurrence that brings together fuel and ignition source in the presence of air MAY BE the result of human action or omission, therefore...

The proponents of this approach often argue that the event should/must be investigated separately, at least in that perspective.

Lekberg seems to suggest that the human error perspective should impart an entirely different perspective to an investigation. I suggest it could add a dimension and should subtract none. Investigation is a metadiscipline and it can accommodate thehuman error perspective. Human error and failure analysis are not guiding principles, they are not in opposition, they do not constitute an either / or option, and they are not face to face alternatives. Hugh

L Benner 1/24/98

Chicoine's comments point to confusion between investigation and analysis. The paper focuses on investigation into the ANALYSIS of REPORTS, not the investigation of accidents. That confusion is interesting. in that it demonstrates once again what mischief the allocation of accident "causes" from reports can create, regardless of the purposes of the investigation or analysis. In that respect, it seems to be similar in nature to the research by Johnson, Ladkin and others, who also analyze accident reports, rather than the investigations which produced the works being analyzed. Vaughan's research shows clearly why that indirect approach is less desirable than direct observations from original sources.

Lekberg demonstrates that analysis of reported accidents for the attribution of cause is a subjective exercise and dependent on the backgrounds of the analyst. This is a useful contribution to the debate about the utility of the entire cause concept when applied to accident reports and investigations.

The the definition of the problem (inconsistency) and the solution offered (integrated miltidisciplinary approach) while not unique, do not seem valid. Re the problem, an obvious question that was not answered was whether the inadequacy of the accident data contributed to the differences. Another question is why the AEB methodology did not produce greater consistency. Lekberg's solution suggests that Lekberg views the problem as attributable to the backgrounds of individuals charged with determining causes. A more logical conclusions seems to be the impossibility of the task, namely objectively and consistently satisfying the demand to select an accident cause or causes from accident reports.

As demonstrated by this project, it appears to me that applying a technically bankrupt concept to flawed data will lead to predictably inconsistent results, and that adding more people to the task will not improve the results..

Your comments are invited.