Flight Safety DIGEST OCTOBER 2005 Analysis of CREW CONVERSATIONS Provides Insights for Accident Investigation Flight Safety Digest Flight Safety Foundation For Everyone Concerned With the Safety of Flight Vol. 24 No. 10 October 2005 www.fl ightsafety.org OFFICERS AND STAFF Chairman, Board of Governors Amb. Edward W. Stimpson President and CEO Stuart Matthews In This Issue Executive Vice President Robert H. Vandel General Counsel and Secretary Kenneth P. Quinn, Esq. Treasurer David J. Barger Analysis of Crew Conversations ADMINISTRATIVE Provides Insights for Accident Manager, Support Services Linda Crowley Horger Investigation 1 New methods of examining recorded voice communications FINANCIAL can help investigators evaluate interactions between fl ight Accountant Millicent Wheeler crewmembers and determine the quality of the work environment on the fl ight deck. MEMBERSHIP Director, Membership and Development Ann Hill On-board Fatalities Lowest Membership Services STATS Coordinator Ahlam Wahdan Since 1984 for Large Commercial Jets Membership Services Coordinator Namratha Apparao Boeing data assembled according to a new taxonomy created by an international team indicate that controlled fl ight into PUBLICATIONS terrain and loss of control in fl ight were, by a considerable margin, the leading causes of on-board fatalities in accidents Senior Editor Mark Lacagnina 18 from 1987 through 2004. Senior Editor Wayne Rosenkrans Senior Editor Linda Werfelman Associate Editor Rick Darby Ethics Is a Safety Issue Web and Print Y Production Coordinator Karen K. Ehrlich ‘Data smoothing,’ ‘pencil whipping,’ ‘normalization of Production Designer Ann L. Mullikin deviance’ — they’re all tempting shortcuts against which Production Specialist Susan D. Reed LIBRAR aviation personnel must take a principled stand in a safety Librarian, Jerry Lederer culture. Aviation Safety Library Patricia Setze 24 TECHNICAL Director of Technical Programs James M. Burin Flight Crews Cleared for Technical Programs Specialist Joanne Anderson Near-simultaneous Takeoffs From Managing Director of Intersecting Runways Air Transport Safety Services Louis A. Sorrentino III Q-Star Program Administrator Robert Feeler A preliminary report from the U.S. National Transportation BRIEFS Manager, Data Systems Safety Board quotes the fi rst offi cer as saying that the captain and Analysis Robert Dodd, Ph.D. delayed rotation of their Boeing 737 as an Airbus A330 passed 28 Manager of Aviation overhead with ‘very little separation.’ Safety Audits Darol V. Holsman Founder Jerome Lederer 1902–2004 Flight Safety Foundation is an international membership organization dedicated to the continuous improvement of aviation safety. Nonprofi t and independent, the Foundation was launched offi cially in 1947 in response to the aviation industry’s need for a neutral clearinghouse to disseminate objective safety information, and for a credible and knowl- edgeable body that would identify threats to safety, analyze the problems and recommend practical solutions to them. Since its beginning, the Foundation has acted in the public interest to produce positive infl uence on aviation safety. Today, the Foundation provides leadership to more than 900 member organizations in more than 150 countries. Cover illustration: © Copyright 2005 Getty Images Inc. N EW ARTICLE TITLE? Copyright © 2005 Rubberball Analysis of Crew Conversations Provides Insights for Accident Investigation New methods of examining recorded voice communications can help investigators evaluate interactions between flight crewmembers and determine the quality of the work environment on the flight deck. — MAURICE NEVILE, PH.D., AND MICHAEL B. WALKER, PH.D. ockpit voice recorders (CVRs) are in- Following the investigation of a controlled-fl ight- stalled in aircraft to provide informa- into-terrain (CFIT) accident involving an Israel tion to investigators after an accident. Aircraft Industries Westwind 1124 jet aircraft, CThey provide records of fl ight crew which struck terrain near Alice Springs, Northern activities and conversations, as well as a variety Territory, Australia, on April 27, 1995, the Bureau of other auditory information. Information from of Air Safety Investigation (BASI, which became CVRs has proved useful in determining the events part of the newly formed multi-modal Australian leading up to aircraft accidents for many years. Transport Safety Bureau [ATSB] in 1999) evalu- However, there has been little discussion in the ated available methods to analyze recorded voice safety investigation fi eld about appropriate ways communications. to analyze recorded voice communications, par- ticularly in terms of analyzing the quality of the [CFIT, as defi ned by the Flight Safety Foundation interaction between crewmembers. CFIT Task Force, occurs when an airworthy FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • OCTOBER 2005 1 C ONVERSATION ANALYSIS aircraft under the control of the fl ight crew is It is acknowledged that a perfect understanding fl own unintentionally into terrain, obstacles or of cockpit activity is diffi cult to achieve without water, usually with no prior awareness by the the opportunity to interview the fl ight crew as a crew.] part of an investigation. However, the methodol- ogy described in this paper is intended to expand As a result of the BASI research, the ATSB con- the level of understanding that can be obtained tracted independent specialists in an emerging from a cockpit voice recording as a part of an fi eld known as “conversation analysis” to analyze investigation. the CVR from the Westwind accident. The proj- ect was conducted by Maurice Nevile, Ph.D., and A.J. Liddicoat, Ph.D., both then of the Australian Executive Summary National University. The independent consultants’ report provided conclusions regarding the crew ecorded voice data, such as from CVRs or interaction that were consistent with the original Rair traffi c control (ATC) tapes, can be an BASI investigation report. More important, the important source of information for accident project showed that the conversation analysis investigation, as well as for human factors re- method provided a very useful approach to search. During accident investigations, the extent identify, describe, demonstrate and explain dif- of analysis of these recordings depends on the fi culties in conversation between two or more nature and severity of the accident. However, individuals. most of the analysis has been based on subjective interpretation rather than the use of systematic This paper discusses the nature of conversa- methods, particularly when dealing with the tion analysis and its potential for use in safety analysis of crew interactions. investigation, as well as its potential for dem- onstrating the importance of appropriate crew This paper presents a methodology, called communication practices. To help explain the conversation analysis, which involves the de- usefulness of the method, information from tailed examination of interaction as it develops the original consultancy project’s examination moment-to-moment between the participants, in of the Westwind CVR is included. This paper context. Conversation analysis uses highly detailed is not an investigation of the circumstances of and revealing transcriptions of recorded voice (or the accident. video) data that can allow deeper analyses of how people interact. It is important to note that a cockpit voice re- cording provides limited information about ac- tivity in a cockpit and cannot provide a complete The paper uses conversation analysis as a tech- understanding of all activities nique to examine CVR data from the Westwind and interactions among fl ight accident. crew. It does, however, provide a good understanding of what The conversation analysis methodology provided happened and why. The analysis a structured means for analyzing the crew’s in- can be enhanced by comparing teraction. The error that contributed directly to average sound recordings from the accident — an incorrectly set descent alti- normal multi-crew communi- tude — can be seen as not the responsibility of cation and activity on a fl ight one pilot but, at least in part, as the outcome of with a recording from a particu- the way the two pilots communicated with one lar fl ight that is being studied. another. The analysis considered the following This comparison can provide aspects in particular: more detailed insights into crew activities and interactions on a • The significance of overlapping talk (when particular operation despite both pilots spoke at the same time); the lack of visual information — for example, from a cockpit • The copilot’s silence after talk from the pilot- video recording. in-command (PIC); 2 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • OCTOBER 2005 C ONVERSATION ANALYSIS • Instances when the PIC corrected (repaired) identifying the reasons for such the copilot’s talk or conduct; and, errors rather than the errors themselves (Maurino et al., • A range of aspects for how the two pilots 1995; Reason, 1990, 1997). communicated to perform routine tasks. These reasons may include In summary, the conversation analysis methodol- a range of factors associated ogy showed how specifi c processes of interaction with the task and environmen- between crewmembers helped to create a work- tal conditions, as well as the ing environment conducive to making, and not broader organizational context detecting, an error. By not interacting to work in which the crews operated. together as a team, pilots can create a context However, to identify