Archives of Personal Papers ex libris Ludwig Benner, Jr.
rainbow line   - - - - - -Last updated 7/30/12
    [Investigation Catalyst Software ] [ Investigation Research Roundtable Site ]   
[ Contact "me" at ]

New: Posted 28 Nov. 1997 Revised 17 Apr 98; Nov. 99

This project is presently on hold, while several significant differences among participants in the project are being explored.


RESEARCH OBJECTIVES: The objectives of the research project are to identify the knowledge and skill requirements for different kinds of investigations, and publish a summary of those requirements for use by investigation process designers, managers, trainers, students, practicing investigators, authors, and potentially, by legislators and investigation certification organizations.

Project Status Information

(c) 1997 by Contributors listed at the end of this document. Contents may be reproduced for educational or research purposes, provided the source is cited.


Project Status Report

Project Title: Accident Investigation Skill and Knowledge Demands

Project status

The purpose of this document is to describe progress on the project to date. Thus far, the accident, fire, and hazardous spill investigations have been reviewed. Thus fare we have been able to identify nine kinds of investigations, and examine the tasks associated with these investigations.

The investigation types identified and defined during the research thus far are listed below. The types are categorized primarily by the nature of the , investigation purpose and outputs, and the scope of the incident process addressed. the investigation tasks' complexity and rigor, and the attributes of the investigation staff. (Alternative or additional criteria would be welcomed.)
Type  Investigation project description 
1.  Produce completed forms created by others 
2.  Produce sequenced description of what happened 
3.  Produce description with explanation of what happened 
4.  Produce description with explanation of what happened, and recommend actions to improve future performance 
5.  Produce Type 3 or 4 outputs with investigation teams 
Produce Type 3 or 4 investigations with international investigation teams 
Produce litigation support investigation 
Produce investigation research investigation 
Magisterial/judicial investigations 
10  Focused investigations (1) 
10a  determine continue operation feasibility 
10b  determine restart feasibility 
11  System investigation (1) 
(1) Suggested by Zotov 7/97; not treated in this summary


Latest comments about status: (11/97)

 The active participants seem to have reached a tentative consensus to attempt to define tasks and skills for small scale (1-3 person) investigations initially.

 The following is retained, but a MANY SIGNIFICANT MODIFICATIONS are in process as a result of comments exchanged among the participants. The contents were designed as thought starters, and in that they have been used successfully among the participants. Revisions will be posted after peer review.

 Current emphasis of exchanges is on defining investigation tasks and how they are presently performed.


Investigation Task Summaries by Investigation Type.

The following describes in general terms the initial investigation tasks thought to be required for each type of investigation, and offers one or more indications of how success is measured. Note that the measure of success should be defined and applied by the investigator's "customer"

Investigation Type 1. Produce completed forms created by others

This type of investigation has as its goal the filling in of a report form with data specified by the report form author or developers. The data called for by Type 1 investigations reflect the form designers' perceptions of the data needed for further analyses. Most often the data provided are used for statistical analyses of trends, relationships, rankings or similar compilations from a statistically sufficient number of reports. Usually, the data are gathered and forms prepared by a single investigator assigned the case.

The investigator's tasks are to understand the specifications for the data to fill in the forms, identify where those data can be found, go to the data sources and observe them, interpret and transform the observations into entries for each of the blanks on the forms, and then fill in the blanks. The investigator usually must go to and make observations at an occurrence site, or wherever the people or remnant objects with the needed data may be found. When a form has a blank to be filled in with a narrative describing the accident or incident, the investigator must prepare and enter the scenario describing what happened in the blank space. Frequently the investigator must make subjective decisions about the entries, based on conclusions drawn from the observations. Primary quality assurance is by peer review. Note that data coding may or may not be a part of the investigator's tasks; this is often a contentious quality assurance issue.

Success is measured by the degree of satisfaction expressed by the forms' users or designers, and the consistency of the data provided by an investigator from one form to the other.

Investigation Type 2. Produce sequenced description of what happened.

This type of investigation involves an informal examination of "what happened" usually by a single investigator. The objective is to produce an oral or written description of the "chain of events" or scenario describing who did what, in the sequence in which they did it. The tasks generally include identifying sources of data about who or what did what and in what sequence, determining where data might be found, going to the sources and making observations. The next task is to transform the observed data into a description of who did what - actions by people or objects that produced the outcome, and then ordering the actions identified to assemble them into a logical sequence. Often the investigator will also be tasked to state the "cause" of the accident or incident. Primary quality assurance is performed by peer review.

Success is measured by the degree to which recipients of the description can follow what happened, how well their questions about what happened are satisfied, and their perceived ability to identify something to do to prevent similar future occurrences. Over time, an investigator's tasks include anticipating the likely questions that will be raised by the output recipient(s) and trying to develop answers with supporting rationale before submitting the final work product.

Investigation Type 3 Produce description with explanation of what happened.

This type of investigation is more demanding than the Type 2, in that tasks require more rigor and less speculation or judgment calls to produce convincing work products. Tasks required depend on the nature of the explanation required. If the explanation is in the form of one or more "causes" task requirements to develop a scenario are similar to Type 2, but the scenario reported must point to anything that is judged to be the cause or causes. Thus, after the actions constituting the scenario are sequenced, an additional task is to identify causes from the scenario.

If the explanation required is in a form of showing causal relationships among actions reported, the task includes application of logic tests to relationships among events to distinguish and show how each event caused the next event(s) in the scenario. Another task is to identify irrelevant events in the scenario, to winnow the essential events from the others identified during the investigation.

Typically, the scope of the task of describing what happened is broader than Types 1 or 2, in that more remote influences on actions by people or objects before or during the loss are considered, and the losses produced are also examined.

Success is measured by the degree of acceptance of the description as valid. The lack of success is indicated by the degree of controversy evoked by the description and explanation. Usually, the more plausible the explanation, the less the controversy. Parties involved in or affected by the occurrence have differing interests in the outcome, and tend to do what it takes to protect those interests, including the raising of doubts about the investigator's explanation. Therefore, another task is to prepare for such attacks with some kind of quality assurance task.

Investigation Type 4. Produce description with explanation of what happened, and recommended actions to improve future performance

This Type is more demanding than lower levels, in that the addition of recommended actions add to the potential for challenging the investigators' work products. Thus the tasks require even more rigor to ensure that the description and explanation of what happened fully support any proposed actions. Additional tasks involved in this level include the discovery, definition and evaluation of problems or needs disclosed by the occurrence, the identification, definition, evaluation and selection of the best proposed actions to improve future performance, and the development of a procedure to track the success of the actions in achieving predicted performance improvement(s). At this level, investigation tasks include quality assurance tasks, for both the conduct of the investigation and the work products it produces.

Note that the tasks required to develop the description and explanation of what happened are retrospective and involve the development and use of historical data. The demand to produce recommendations imposes predictive tasks on the investigator, in that the investigator must predict future events in evaluating needs and proposed actions, as well as the performance tracking.

Success can be measured ultimately by the degree to which the definition of the problems or needs is accepted, by the acceptance of the validity of the recommendations, by the demonstrated improvement in the performance achieved by the recommendations actually implemented, or by the efficiency with which the tasks are performed.

Investigation Type 5. Produce Type 3 or 4 outputs with investigation teams.

Team investigations involve project management tasks in additional to the investigation tasks described thus far. They require whoever is in charge of the investigation to plan, organize, staff, direct and control all the investigation tasks performed, so they produce the quality of outputs demanded of the team, on time and within budget. Planning tasks include the selection of the objectives, scope, methodology, schedules, deliverable specifications, and quality assurance specifications. Organization tasks include the divisions of labor and tasks designs required to perform the investigation tasks to meet the planning needs. Staffing tasks include selection of qualified personnel, their training or indoctrination, and the funding of their time. Direction tasks include the day to day task assignments to each team member, and the progressive integration of their findings as the investigation proceeds. They also involve determining the value of additional data and weighing of tradeoffs involved in spending the money to get the data. The control tasks require tracking work of others to ensure that planned task objectives and outputs are achieved and produced on time and on budget, while simultaneously achieving a satisfactory quality level.

Success is measured by the quality, efficacy, efficiency, timeliness and consistency of the work products produced by the Team.

Investigation Type 6. Produce Type 3 or 4 investigations with international investigation teams

International team investigations add special tasks to Type 6 Team investigations. The tasks may involve serving on a team formed by another nation, or directing or serving on a US team with foreign nationals representing their interests in the people or objects involved. Added tasks include preparations for functioning on a team operating in a foreign nation, to ensure national laws or regulations, unique protocols or procedures, or national customs or sensibilities are not transgressed. Another task is the planning and execution of international travel arrangements, including credentials, passports, accommodations, interpreters, communications during on site and subsequent exchanges, among other items.

Success is measured by the degree of harmony achieved during the field investigation and during the development of the findings and the report preparation, and by the results achieved.

Investigation Type 7. Litigation support investigation

This type of investigation differs from the others in that litigators need investigation inputs to support or disprove their legal theories, and meet the prerequisites for admissibility in the venue in which the case is being argued. The investigator's task is to develop forensic evidence on which to base opinions about what happened. Another task is to help rebut arguments made by the other side in an adversarial process. Thus the task is similar to Type 3 investigation tasks, with the added task of helping to integrate the investigation findings into the case and testimony within the rules of the jurisprudential rules which govern.

Success is measured by the ability of the investigator's work to withstand cross examination without having to acknowledge error or unreported uncertainties.

Investigation Type 8. Investigation research investigation

This is the most sophisticated type of investigation, in that the investigator must understand what happened and how the occurrence was investigated. The investigation research investigator must observe what the incident investigators do and why they do what they did. Concurrently and independently, the investigator must be able to find out what actually happened and why it happened, to have a valid baseline from which to critique the investigation process. Success is measured by the increased understanding of investigation needs and tasks required to satisfy those needs, and the translation of that understanding into improved investigation efficacy, efficiency, utility, replicability and value.

Investigation Type 9. Magisterial investigations.

This is a special type of investigation, usually conducted under the direction of a noted jurist or prominent official, conducted under the rules of procedure of the nation in which it occurs. These requirements and the subsequent tasks vary widely. This type has not yet been studied thus far to identify any task requirements. 

Type 10 Focused investigations. Suggested by Zotov (Discussion to follow) 

Type 11. System investigation Suggested by Zotov (Discussion to follow)

Investigator capabilities.

A vacancy announcement for an NTSB investigator position is provided to indicate the qualifications required for positions as field air safety investigators. Position descriptions or job announcements are invited to help define investigation tasks for this research project.

View vacancy announcement.

Table 1.

This table begins to list the tasks required for each of the types of investigation considered thus far. When this task list has been completed, the task knowledge and skill requirements will be attached to each task for each investigation type. The tasks need to be defined because it is anticipated that some tasks may require similar knowledge or skills. This will likely require numerous iterations as the task list and knowledge and skill needs are developed.

Inputs in the form of new tasks to add to the list, or restatements of the tasks listed are invited. For modifications, it will help if you list the task number (1 through 1100) and indicate to which type of investigation (1-11) your comments apply.

View Table 1 now

Contributors to date include

Ludwig Benner
Hughes Chicoine
Jeremy Dawson
C O Miller
Ira Rimson
Terje Sten
James Stewart
Frank Taylor
Dmitri Zotov
William Waldock


[B] [C] [D] [E] [F] [G] [H] [I] [J] [K] [L] [M] [N] [O] [P] [Q] [R] [S] [T] [U] [V] [W] [Z]

 This is an index of topics addressed in exchanges among project participants to date. Discussion is found in files distributed periodically to participants.

- - A - -

AAFS, Code of Ethics
acceptance of the recommendations
accident causation
accident, carrier
adversarial habits
- causal
- enforcement
- investigation

Alkov R
Ames human factors paper, 1928
analysis- accident- Cali- causal- causal vs why- because- cognitive failure- comparative- deontic- failure- failure, Pterodactyls- fire- from accidents vs reports- hazard- human failure- in report- inconclusive- industrial accident- lack of standard technique- linguistic- logical- mechanical- opponent's- performed on any facts- predictive- progressive- proactive- quality- risk- root cause- stage- structural- textual form- traditional- W- B graph- Why- Because- withstanding criticism/analysis causal
analysis Why- Because
analytical templates
analyze, ability to
approach- ASSE- consensus- empirical- fact, analysis, conclusions- group- judicial- Ladkin- management- multiple time lines- quality assurance- scientific- shrinking from- system safety- systematic- to investigations- to losses Arrow Air arson
ATS Guide
audit trail
investigation, organization

- - B - -

blame, allocating
Boscombe Down
Brief of Evidence
bureaucratic inertia

- - C - -

Cali crash
case selection
casualty- definition- priority overcausalty definition
cause- complex- determining- insignificant- joint- notion- partial- primary- probable- reversal- sequential- specific- suppositions- underlying Causes
chain of custody
chains of causation
check and balance
Chernowitz's First Law of Mathematical Utility
classification- human errors
client perspective
Coding, uniform fire investigation
cognitive failure
concept trees
conclusions- official- unjustified- W- B graph
conflict, unresolved
consensus, on investigation functions
context of investigation tasks
context, institutional
contributory factor'
controversy, sources
Courts of Inquiry
creative insight
criminal proceedings

- - D - -

data- format and order- irrelevant.
data sets
data, for analysis- data for
decision criteria
decision makers
define what is being caused
delay- implementing recommendations
description, narrative
description, scenario
description, symphony
design- human- machine- instructional
diagrams, entries
discovery process
documentation, investigation
due deference

- - E - -

ethics- investigator- ISASI- manager- moral content- research project
evaluation process
evaluations- after- market
event- counterfactual- identifying- reconstructed- segment
events- selecting for analysis
evidence- circumstantial- conflicting- contamination- contradictions in- estimated- fuzzy- hard- inferential- interpretation- missing- partial- physical- preserving- psychological- resolve controversy- spoilation- types- what is
evidence preservation

- - F - -

fallacies- list of Logic
fault- tree
fire propagation
fire suppression
focus, research
Focused investigations
free association
functions- management- of research- recommendation
fundamental questions
fuzzy logic

- - G - -

GBS Sleipner
Guam Accident

- - H - -

Hazard Prevention
hazards, mitigating
Heron case
Holmes, Sherlock
human error, workshop
Hunt D
hypothesis- generation- importing into investigation process- never mind the facts- premature- premature- proving- randomly chosen- rejected by observation- scientific- self criticism- testing- unrecognized

- - I - -

ideas, organization of
illusion of knowledge
incompleteness- detecting in reports
inquiries, framing
instruction- designing- human factors investigation- technique
intervention placement
investigating authorities, province of
investigation- accident- costs- customers- developing standard for- failure- field, work products- fire- funding- incident- insurance- Level- levels- manuals- matrix for- of investigations- police- process- program- quality- research- sponsors- tasks- types
investigations- human factors- proactive
investigations, multi- national
investigative process
Investigators, ATS
investigators, fire and explosion
investigators, government
ISASI- politicization of- ATC- editorial- policy- policy re guides
ISD model
issues, in investigations

- - J - -


- - K - -

Knowledge, prerequisite for fire investigation

- - L - -

litigation, litigators
logic types
Lusaka crash recommendations

- - M - -

Mahon Report
Mahon's book
management- influences on accident- experience- investigation- model, ideal for investigators- of assessment- of instruction- of risks- of safety critical systems- reviews
management process- ignoring
management style
managerial concerns
Manchester, research
merits of Recommendations
methodology- competent- examining- fire investigation- progressive analysis- standardized
modeling method
Muir, Helen

- - N - -

Namibia midair
need for improvement
NFPA 921- critique
Nihil illegitimi carborundum
Norwegian AAIB
notations, on diagrams

- - O - -

objective- quality assurance
objectives- investigation task- measurable
objectives of sponsors
observation- interpreting- photographic- recording- retrieving
open- ended commitments
organizational factors

- - P - -

peer review
Petri- Net
Politically Correct
practices, ISASI Standard
prevention- accident
principles, EIR guiding
probability, 0
Probable Cause
problem- defining- defining investigation- diagnosis- fixing- known
problem, understanding a
Product Liability
progressive analysis
project- all differ- every investigation is a
public statements

- - Q - -

Qualification, Professional, Fire Investigation
qualifications- for investigating management
quality analysis
quality assurance
quality control

- - R - -

Reason J
Reason model
reasoning- ability to understand- Boolean- causative factor- deductive- differentiating tasks- faulty in reports- from things observed- law vs science- logical- necessary and sufficient- non- monotonic- scientific- sequential
recommendation- consequences, of- development- first is spend more money- functions- implementation- options- ranking- response to
recommendations- delayed- retrofits- safety- what kind?
references, in investigation guides
requirements- investigation
research, investigation as
research- investigation project
research, investigation process
resistance to commentary
resources, investigation
responsibility- for objectivity
responsibility, investigator
restart issues
Richardson J
risk- high, training for
risk acceptance
risk of partial falsehood
root cause
Root Cause Analysis- RCA
Rules of Court

- - S - -

safety management
scientific investigation
scientific method- goal
semantics- causal- Lewis
silver bullet
skill, writing
sources, for investigation guides
Spatial Disorientation
standards, ISASI
subjective baggage
success, measure of
synthetic faculty
System Safety
Systems investigation

- - T - -

task- unavoidable investigation
task order
tasks- investigation
techniques- hazard analysis- modeling- need for logical rigor- refining
thinking skills
thought process
thoughts, organizing
training, investigation
training, major accident investigation
training, standardization
Transport Accident Investigation Commission
TWA Flight

- - U - -


- - V - -

validation method
validation process
viable remedial action

- - W - -

Warwick explosion
WB- graph
Wider Twin Otter accident
witness estimation
witness plates
writing, linearity

- - Z - -