The Transformation of CRM-
based Training and it Effects on
Aviation Safety in the U.S. from
1960 to 2013
Lukas Rudari, Tyler Spence, Lauren Sperlak, Robert Geske, Micah
Walala, Paul Brown, Christopher Morris, Mark Fernandes,
Joseph Krsek, Mitchell Hart, Michael Thanos
Purdue University Student Chapter Flight Safety Foundation
Agenda
(1) Introduction
(2) History of CRM
(3) Methodology
(4) Statistical Analysis
(5) Results
(6) Discussion
(7) Conclusion
Introduction – Purpose of
Study
The purpose of this study was to analyze the NTSB
accident database and assess accident reports in order
to determine if there was a significant change in Crew
Resource Management (CRM)-related accidents
associated with the progression of CRM training
History of CRM
1970s
Aircraft systems became more complex
1980s Rate of catastrophic incidents did not decline as anticipated
It became apparent that accidents tended to be “attributed
less to mechanical failure and more to human error” (Alkov,
1990s 2009, p.3)
NASA conducted interviews which revealed that
Many near miss accidents were result of a breakdown in
2000s
communication and that
Critical information was not relayed to pilot in command
Present
History of CRM
1970s
Between 1980 and 1989, causal factors for accidents
attributed to flight crew were still in excess of 70% of all
1980s
accidents (Wiener, Kanki, & Helmreich, 1993)
United Airlines was the first U.S. airline to introduce a CRM
program in 1981, and many airlines followed by the mid 1980s
1990s
All cockpit crew members received training on team-
building, leadership styles, and decision-making
2000s
United reported success and change in corporate culture
By the late 1980s, Cockpit Resource Management was
replaced by Crew Resource Management
Present
History of CRM
1970s
AC 120-51: Cockpit Resource Management Training was
issued by U.S. FAA in 1989
1980s
Mandatory compliance for scheduled airlines
Included flight attendants, dispatchers and maintenance
personnel
1990s
FAA introduced Advanced Qualification Program (AQP), which
allowed carriers to develop customized training
2000s
Scope of CRM was broadened to include human factors issues
CRM became integral part of flight training
Present
History of CRM
1970s
Human behavior is inevitable but can provide valuable
information for future focus and training
1980s
Focus on the limitations of human performance and ways to
reduce error
Organizational culture is still an obstacle to effective
1990s
implementation
CRM is regarded as a countermeasure with three main goals:
2000s
error avoidance
trapping emerging errors before they amplify risk
Mitigating the consequences of any errors committed
Present
History of CRM
1970s
Transition from “error management” to “threat management”,
from reactive to proactive
1980s Line Oriented Flight Training (LOFT) allows the use of
simulators to expose crews to non-standard or
emergency situations
1990s Identification of systemic threats to safety
Implementation of CRM into critical areas of flight like
maintenance and dispatch
2000s
Beyond flight into medicine, firefighting, and railway
systems
Present
Methodology
The aviation accident and incident data was collected from
the NTSB online database
A list of keywords identified as essential in CRM failures was
created by the researchers based on identified core
components and standard practices of current CRM theory
The keywords were used as the query list in the database to
search for accident and incident probable causes
Methodology
The list of keywords used to
query the database
Statistical Analysis
Descriptive statistics were conducted first to
gather averages and standard deviations of
generations of CRM policies
The accidents and incidents were tallied by year
The total number of commercial departures for
each year were collected from Department of
Transportation (DOT)
In order to empirically analyze the conditions of
accidents and incidents related to CRM, a
segmented regression of interrupted time
series was used
The outcome variable for this analysis was
accidents and incidents per one million
departures
The predictor variable was time
Two-tailed p-values of 0.1 were chosen
Results
CRM Accidents and Incidents per million departures
12
10
Number of accidents/incidents
0
1962 1972 1982 1992 2002 2012
Continual increase until the early 2000s
Peak occurred in 1997, with 10.11 accidents/incidents per
one million departures
Results
Mean Accidents and Incidents per million departures after CRM policy changes
Results
A segmented regression interrupted time series analysis was
performed to determine if different trends existed before and
after CRM policy implementation
There were noticeable differences in terms of the accident and
incident rates between three policy periods
However, the second series of iterations in Policy Intervention
Period II did not appear to have any added benefits beyond the
general effect of CRM policy already being established
Results
CRM accidents and incidents per million departures trend line
with government policy interventions
10
9
Number of accidents/incidents
8
7
6
5
4
3
2
1
0
1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013
Pre 1990 1990-1998 1999-2013
Discussion
Based on the analysis, there appeared to be a significant
effect on accident and incident rates among commercial
aircraft with the implementation of CRM policy
Initial increase after the first adoption of CRM policy
By the last series of iterations, accident and incident rates began to
reduce significantly
Based on the analysis it is difficult to tell if the latter
evolutions of CRM are as important in reducing CRM
accidents and incidents as ensuring that there is a focus on
CRM training among the airlines and employees
The initial rise and following decline of accidents and incidents
could be indicative of an initial compliance with the initial CRM
policies, but a delayed application in practice among airlines
Conclusion
Using the NTSB accident database, the researchers assessed
accident reports in order to determine if there was a change in
CRM-related accidents associated with the progression of CRM
training
Based on the analysis, there appeared to be a significant effect on
accident and incident rates as a result of the CRM training
The results suggest that the focus on CRM training in the aviation
industry has made an impact in reducing human error
As CRM understanding and implementation continues
to advance in international and airline policies it will be
necessary to ensure a full comprehension and practical
application of knowledge gained by industry, and
scientific research to ensure the best possible gains in
safety
