1. Report No. 2. Government Accation No. 3. Recipient's Catalog No.

NASA TP-1875 1
4 Title and Subtitle 5. Repon Date

INFORMATION TRANSFER PROBLEMS IN THE September 1981

6. Performing Organization Code
AVIATION SYSTEM
7 Authorls) 8 Performing Organization Report No

C. E. Billings and E.S. Cheaney, Editors A-8567

10. Work Unit No
9 Performing Organization Name and Address
- 505-35-41
Ames Research Center, NASA 11 &tract or Grant No
Moffett Field, CA 94035
13 Type of Repon and Period Covered
2 Sponsoring Agency Name and Address
Technical Paper
National Aeronautics and Space Adnunistration 14 Sponsoring Agency Code
Wastungton, D. C. 20546
5. SugplemeGry Notes

C. E. Billings: Ames Research Center.

E. S. Cheaney: Battelle's Columbus Division, Mountain View, California.
6 Abstract

This technical paper is a compilation of seven related studies of information transfer problems. The first
describes the dunensions of the problem in the aviation system, based on screening of a sample of 12,373
ASRS reports. Two studies discuss communication problems within and among air traffic control work
stations. Two papers explore air-ground communication problems under normal and emergency conditions.
A report of communication patterns and problems within the cockpit presents experimental and anecdotal
data to show the relationship between communications and flightcrew performance. A final paper sum-
marizes the findings of these studies, discusses the implications for safety, and offers conclusions.
It is concluded that both human attributes (distraction, forgetting, failure to monitor, nonstandard proce-
dures and phraseology, and complacency) and system factors (nonavailability of traffic information,
degraded information, ambiguous procedures, environmental factors, high workload, and, less commonly,
equipment failure) contribute to information transfer deficiencies. Various facets of the air traffic con-
troller's role are discussed in the context of the enabling factors found in these studies. The role of the
pilot i s considered, especially when flying under visual flight rules, who could do more to enhance informa-
tion transfer within the system. Automation also appears to be a factor in information transfer in that
automated infomation systems may lead to delayed decisionmaking. Behavioral impediments to infonna-
tion transfer are discussed. Several ways of enhancing information transfer are presented.

7. Key Words (Suggcrted by AuthorW) 18. Distribution Statement

Communications Unclassified - Unlimited
Air traffic control
Flight operations
Human factors
Subject Category 03
3. Security Classif. lof this report1 20. Securitv Classif. lof this pagel 21. No. of Pages 22. Rice

Unclassified Unclassified 98 A05

For sale by the Natlonal Technlcil lnlormatlon Service. Springfield, Virginia 2 2 1 6 1

NASA-Langley, 1981

I
NASA Technical Paper 1875

Information Transfer Problems

in the Aviation System

Edited by
C . E. Billings
Ames Research Center
Moffett Field, California
E . S . Cheaney
Battelle’s Columbus Division
Mountain View, California

NASA
NalIonal Aeronautics
and Space Administration
Sciontific and Technical
information Branch

1981
 TABLE OF CONTENTS

Page

SUMMARY ................................................................ 1
INTRODUCTION ........................................................... 1
LIST O F CONTRIBUTORS ................................................... 3

ABBREVIATIONS AND ACRONYMS .......................................... 5

1 . DIMENSIONS O F THE INFORMATION TRANSFER PROBLEM ................. 9

Charles E. Billings and William D. Reynard

2. INFORMATION TRANSFER IN THE SURFACE COMPONENT O F

THE SYSTEM: PROBLEMS ASSOCIATED WITH BRIEFING O F
RELIEF CONTROLLERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Ralph L. Grayson

3. INFORMATION TRANSFER IN THE SURFACE COMPONENT O F

THE SYSTEM: COORDINATION PROBLEMS IN AIR TRAFFIC CONTROL ...... 25
Ralph L. Grayson

4. INFORMATION TRANSFER BETWEEN AIR TRAFFIC CONTROL

AND AIRCRAFT: COMMUNICATION PROBLEMS IN FLIGHT
OPERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Ralph L. Grayson and Charles E, Billings

5 , INFORMATION TRANSFER WITHIN THE COCKPIT: PROBLEMS

IN INTRACOCKPIT COMMUNICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
H.Clayton Foushee and Karen L. Manos
6. INFORMATION T R A N S F E R DURING CONTINGENCY OPERATIONS:
EMERGENCY AIR-GROUND COMMUNICATIONS ............................. 73
Richard F. Porter

7. THE INFORMATION TRANSFER PROBLEM: SUMMARY

AND COMMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
C.E. Billings and E. S. Cheaney

...
111
 INFORMATION TRANSFER PROBLEMS IN THE AVIATION SYSTEM

SUMMARY

Problems in the transfer of information within the aviation system were noted in over 70% of
28,000 reports submitted by pilots and air traffic controllers to the NASA Aviation Safety Report-
ing System (ASRS) during a five-year period 1976-1 98 1. These problems are related primarily to
voice communications, although many deficiencies in visual information transfer have also been
described.

Much of the information on which the safety and efficiency of the aviation system depends is
highly dynamic. Without current, unambiguous information, neither pilots nor controllers can make
appropriate decisions in support of their respective missions. Information transfer deficiencies,
therefore, have obvious safety implications.

This technical report is a compilation of seven related studies of information transfer prob-
lems. The first describes the dimensions of the problem in the aviation system, based on screening
of a sample of 12,373 ASRS reports. Two studies discuss communication problems within and
among air traffic control work stations. Two papers explore air-ground communication problems
under normal and emergency conditions. A report of communication patterns and problems within
the cockpit presents experimental and anecdotal data to show the relationship between communica-
tions and flightcrew performance. A final paper summarizes the findings of these studies, discusses
the implications for safety, and offers conclusions.

It is concluded that both human attributes (distraction, forgetting, failure to monitor, non-
standard procedures and phraseology, and complacency) and system factors (nonavailability of
traffic information, degraded information, ambiguous procedures, environmental factors, high
workload, and, less commonly, equipment failure) contribute to information transfer deficiencies.
The air traffic controller plays a pivotal role in information transfer within the aviation system;
various facets of this role are discussed in the context of the enabling factors found in these studies.
The role of the pilot is considered; the study suggests that pilots, especially when flying under
visual flight rules, could d o more t o enhance information transfer within the system. Automation
also appears to be a factor in information transfer in that automated information systems may lead
to delayed decisionmaking. Behavioral impediments to information transfer are discussed. Several
ways of enhancing information transfer are presented.

INTRODUCTION

The NASA Aviation Safety Reporting System (ASRS) was implemented in April 1976, under
the terms of a Memorandum of Agreement between the National Aeronautics and Space Admin-
istration (NASA) and the Federal Aviation Administration (FAA). During the ensuing 5 years, over
28,000 reports weTe voluntarily submitted by pilots, air traffic controllers, and others who work
within the U.S.national aviation system.
 Throughout this period, skilled aviation professionals have scrutinized incoming reports for
evidence of deficiencies or discrepancies in the system, and for human behavioral problems that
may negatively affect the system's safety and efficiency. From the early months, it has been clear
that deficiencies in the handling of information are frequent concomitants of the occurrences
reported t o ASRS. As a result, research performed on this large data base has tended t o focus on
problems in the management and handling of information. This technical report deals with informa-
tion transfer as a generic problem; it examines information transfer problems in several settings.

Close examination of ASRS reports led to the finding that information transfer problems, as
we have come to call them, did not ordinarily result from an unavailability of information nor
because the information was incorrect at its source (although there are certain exceptions t o this
generalization). Instead, the most common findings showed that information was not transferred
because (1) the person who had the information did not think it necessary to transfer it or (2) that
the information was transferred, but inaccurately. A host of other problems was noted, but less
frequently. The distribution of information transfer problems in a large subset of aviation safety
reports is discussed in chapter 1.

The air traffic control system may be thought of as an information management system in
which nearly all the infomation changes rapidly over short periods of time. The ATC system con-
tains many information transfer nodes or interfaces, but the tactical management of air traffic
depends almost entirely on the timely transfer of information among controllers and between con-
trollers and pilots. Among controllers, two interfaces appear to be associated with significant
potential information transfer problems: problems associated with the transfer of information
between a controller responsible for a sector and another controller who is about t o relieve him
(discussed in ch. 2) and problems associated with the transfer of information between sectors - the
coordination of air traffic (examined in ch. 3 in the context of the often severe operational prob-
lems occurring therein).

The exchange of information between controllers and pilots is presently carried out entirely
by means of radiotelephone. Voice messages carry virtually all clearance, advisory, and warning
information from ground to air, and provide the medium by which pilots normally respond t o such
information. Many psychological studies over the years have examined the types of problems
involved in the oral transfer of information; chapter 4 examines the communication problems
reported to ASRS.

The management of information within the cockpit of modern complex aircraft is as impor-
tant t o the safe and efficient conduct of flight operations as is the management of information on
the ground, The competent aircraft commander in today's environment must collate and process a
great deal of information from aircraft instruments, from manuals and charts, from other flight-
crews, and from the ground. This task is not always a simple one, particularly when some of the
information received is contradictory, ambiguous, o r otherwise misleading. The information inter-
faces among cockpit crewmembers are examined in chapter 5 .

Chapter 6 contains the results of a study performed to determine whether information is

efficiently transferred in situations in which the system is not performing normally. It has long
been known that people function differently in emergencies than under other circumstances.
Narrowing or channeling of attention occurs under such circumstances; peripheral information may
easily be missed. Subjective time compression is frequently observed, and essential functions may be

2
neglected. Although aviators and air traffic controllers are trained to cope effectively with emer-
gencies, information transfer during such emergencies is often a secondary task. It was thought that
information transfer failures in such circumstances could compound the original problem, whereas
under less trying circumstances, the problem might arise from a deficiency in information manage-
ment. This proposition was examined as part of a larger effort t o study behavior in emergency
situations.

The final chapter represents an attempt to draw generalizations from the data studied in the
course of these six research tasks. It summarizes the conclusions drawn in the course of the studies
and suggests possible approaches to the problems identified.

Each of the chapters of the report presents the conclusions of its author and should be read as
an individual research effort. The conclusions presented are also the opinions of the authors and d o
not represent an official position of NASA, Battelle’s Columbus Division, or the FAA.

The editors acknowledge with gratitude the insights of ASRS staff members that led t o this
series of studies. They are most grateful to Dr. Earl L. Wiener and Dr. Alan B. Chambers for their
thoughtful and critical reviews of the final document.

EDITORS

Charles E. Billings Ed S. Cheaney

Ames Research Center, NASA Aviation Safety Reporting System Office
Moffett Field, California 94035 Columbus Division
. Battelle Memorial Institute
. Mountain View, California 94043

3
 LIST OF CONTRIBUTORS

Charles E. Billings Ralph L. Grayson

Assistant Chief for Research Principal Research Scientist
Man-Vehicle Systems Research Division Battelle ASRS Office
Ames Research Center, NASA 625 Ellis Street, Suite 305
Moffett Field, California 94035 Mountain View, California 94043

Ed S. Cheaney Karen L. Manos

ASRS Project Manager Cadet First Class
Batteile ASRS Office United States Air Force Academy
625 Ellis Street, Suite 305 Colorado Springs, Colorado 80841
Mountain View, California 94043

H. Clayton Foushee Richard F. Porter

Postdoctoral Research Associate Principal Research Scientist
Man-Vehicle Systems Research Division Battelle Columbus Laboratories
Ames Research Center, NASA 505 King Avenue
Moffett Field, California 94035 Columbus, Ohio 43201

William Reynard
Chief
Aviation Safety Reporting System Office
Ames Research Center, NASA
Moffett Field, California 94035

4
 ABBREVIATIONS AND ACRONYMS

AIC Aircraft

ACR Air carrier aircraft

AIM Airman’s Information Manual

AR Approach radar controller position

ARTCC Air Route Traffic Control Center

ASDE Airport surface detection equipment (radar)

ASR Airport surveillance radar

ASRS NASA Aviation Safety Reporting System

ATC Air traffic control

ATIS Automated terminal information service

B747 Boeing 747 widebody transport aircraft

BFF Scottsbluff, Nebraska

BOR Briefing of relief

BOS Boston, Massachusetts

BRITE Bright radar indicator (control) tower equipment

CRT Cathode ray tube or scope

DAB Daytona Beach, Florida

DCA Washington National Airport, Washington, D.C.

DEN Denver, Colorado

DME Distance measuring equipment (airborne)

DR Departure radar control position

E East

FAA Federal Aviation Administration

5
FAM Farmington, Missouri

FE,F/E Flight engineer

FL Flight level (standard altitude in hundreds of feet)

F/O First officer or copilot

F4 MacDonnell-Douglas fighter-bomber

GA General aviation

IAD Dulles International Airport, Washington, D.C.

ICT Wichita, Kansas

IFR Instrument flight rules

ILS Instrument landing system navigation facility

IMC Instrument meteorological conditions

J Jet airway (with identifying number)

J FK J. F. Kennedy International Airport, New York

L Left one of a pair of parallel runways

LAX Los Angeles, California

LGA LaGuardia Airport, New York

LHR London Heathrow Airport, England

LIT Little Rock, Arkansas

LTSS Less than standard separation

Mayday Internationally recognized distress call or message

MIA Miami, Florida

Mode A Airborne radar beacon transponder without altitude reporting

Mode C Airborne radar beacon transponder with altitude reporting

MSL (Altitude with respect to) mean sea level

6
N North

NAS National Airspace System

NASA National Aeronautics and Space Administration

NE Northeast

NM Nautical miles

NMAC Near midair collision

NTSB National Transportation Safety Board

Nw Northwest

ONL O’Neill, Nebraska

ORL Orlando, Florida

R Right one of a pair of parallel runways

RAPCON Radar approach control facility (military)

RVR Runway visual range

S South

SE Southeas t

SFO San Francisco, California

SID Standard instrument departure route

SMA Small aircraft (less than 5,000 lb gross weight)

SMT Small transport (5,000-12,500 lb gross weight)

Stage 111 Radar control and separation services in terminal area

sw Southwest

TACAN Tactical air navigational aid (navigation facility)

TRACON Terminal radar approach control facility

U.S. United States

7
UHF Ultrahigh-frequency (radio communications unit)

V Victor airway (with identifying number)

VFR Visual flight rules

VHF Very-high-frequency (radio communications unit)

VMC Visual meteorological conditions

VOR Visual omnirange navigation facility

W West

WX Weather

xxx Code letters used to avoid identifying a location or navigation facility

XYZ Code letters used t o avoid identifying a location or navigation facility

8
 ~~ ~~

1 . DIMENSIONS OF THE INFORMATION TRANSFER PROBLEM

Charles E. Billings and William D. Reynard

INTRODUCTION

This report describes a class of problems that in one form or another is involved in the vast
majority of the occurrences reported to ASRS. Over 70% of reports contain evidence of human
error, either in the air or on the ground. An even larger fraction, however, contain evidence of a
problem in the transfer of information within the aviation system. Several facets of the information
transfer problem have been examined in previous ASRS reports (refs. 1-3). Others are addressed in
this compilation. ?'his chapter presents a general analysis of the problem and examines the implica-
tions of some proposed solutions.

INFORMATION TRANSFER IN AVIATION

There are presently two primary methods of transferring information in the aviation system.
Relatively static information is stored in the form of books, manuals, and charts. Some relatively
dynamic information, such as weather charts, is also stored in hard copy. The transfer mechanism
for this information is, of course, the visual perceptual channel.

Dynamic information concerning aircraft status is portrayed in analog or digital form on elec-
tromechanical devices (instruments) in the cockpit. Information concerning aircraft identity, loca-
tion, speed, and intentions is displayed in alphanumeric and symbolic form on cathode ray tubes in
ATC facilities. Again, the information transfer mechanism is the visual perceptual channel.

Information about weather and other aircraft in the immediate area, under visual meteorologi-
cal conditions, is also received by the aircrew by means of the visual perceptual channel.

Highly dynamic information (clearances, traffic avoidance data, changes in weather, notifica-
tion of unplanned changes in status) is a t present transferred almost entirely by means of simplex
radiotelephone. It is transmitted by voice and received through the aural perceptual channel.

Other information concerning terminal-area weather and related information is also trans-
mitted by voice communication channels. Nonverbal aural information transfer includes Morse code
identifiers and alerting and warning sounds.

APPROACH

Each incoming report was screened by experienced aviation professionals t o determine

whether, in their view, an information transfer problem was present. If their finding was positive,
they provided information in four coded fields t o characterize the problem. These fields formed the

9
basis for the data summaries presented in this report. The instructions are summarized and the data
fields discussed briefly here,

1. Instructions: Analysts were instructed t o complete a set of data fields for each information
transfer event deemed significant t o the occurrence or situation described in a report. Significance,
in this context, was to be based on the contribution of an information transfer event (or its
absence) t o the unsafe occurrence or situation. Thus, significance has the negative connotation of
“fault”: it conveys that a message o r information transfer event had something wrong with it.

2. Message origin: In this field, the analyst indicated whether the message originated with a
controller, a pilot, o r some other person or device. When information believed to be pertinent was
not transferred, the analyst indicated by whom such information should have been provided.

3. Message type: The analyst indicated whether the information t o be transferred was a clear-
ance, coordination message, request, warning, other control message, statement of intentions, data,
advisory, or confirmation, o r other type of instruction.

4. Message problem: Here the analyst expressed his view of the nature of the information
transfer problem (see below). Messages could be categorized as absent, incomplete, inaccurate due
to phonetic similarity, transposition or other cause, false, ambiguous, untimely, garbled in trans-
mission or presentation, not transmitted because of equipment o r device failure, or not received
because of a failure t o monitor by the intended recipient.

5 . Message medium: Here the medium used, or that would have been used, t o convey the
information was indicated. Available choices included publication, radio, interphone, video, tape
recording, chart or similar graphic, telephone, direct voice, o r visual (instrument, etc.).

It will be noted that analysts were t o decide whether a failure to transfer information contrib-
uted t o an unsafe occurrence. This, of necessity, involved a judgment by the analyst, and the data
are heavily influenced by this judgment, as will be seen.

The coded data from a sample of 12,373 consecutive reports received between May 1, 1978
and July 3 1, 1980 were used in this study.

RESULTS

Grayson (ref. 4) in a study of air-ground communications, found evidence of an information

transfer problem of some type (aural o r visual) in over 70% of ASRS reports received between May
1978 and August 1979. The data have been rescreened based on intake since that date. The classi-
fication of the problems found by ASRS analysts in these reports is shown in table 1.1.

These data illustrate each of the generic problems within the information transfer process. A
common problem was failure t o initiate the information transfer process. It is necessary both that
the information exist (and it almost always does exist) and that it be made available t o those who
need it. In a larger number of cases, an attempt was made t o transfer information. The common
failure points in the transfer chain were inaccuracies in the transmitted message, and failures to

10
transmit the message at the right time. TABLE 1 . 1 ,- INFORMATION TRANSFER PROBLEMS
Many of the latter failures involved the
issuance of advisory messages concerning Number Percent of
Nature of problem
other air traffic. of citations citations

Message not originated 4,039 37

In a substantial fraction of these
Message inaccurate, incomplete,
reports, a message was not heard, o r if ambiguous, garbled 3959 37
heard it was misunderstood or misinter- Message correct but untimely 1,356 13
preted. The reports indicate that expec- Message not received or
tation plays a major role in this facet of misunderstood 1,165 11
the problem. Misunderstandings often Message not transferred due to
occur because the recipient of the mes- equipment failure 296 3
sage, usually a pilot, hears what he
expects t o hear based on familiarity with
the “usual” way of doing things. This
I
10,815’ 1 00%

error is then compounded when the ‘10,815 citations in 9,046 of 12,373 ASRS reports, May
originator of the message, usually a con- 1978-July 1980; 73.3% of reports contained such a citation.
troller, hears what he expects to hear:
the clearance read back as he gave it. If a message is not received, it will not be acknowledged, and it
will usually be transmitted again. If a misunderstood message is acknowledged as correct, everyone
thereafter is operating on wrong presumptions.

It is noteworthy that few of these informa- TABLE 1.2.- MEDIA INVOLVED IN INFORMA-
tion transfer failures (less than 3%) involved an TION TRANSFER PROBLEMS
equipment . failure. The rest involved human
errors, or the inability of the human to accom- Number of Percent of
Message medium
plish all of the tasks involved in information trans- citations citations
fer in timely fashion, or a decision by a human
Aural information transfer
that information transfer was not necessary.
Radio 6,834
Interphone 855
Does visual information transfer work Voice 450
better? The data would suggest that it does, Telephone 438
though hundreds of reports involve problems Tape recording 19
with charts and manuals, the reading and setting
of aircraft instruments, the placement, format, 8396 85
and presentation of information on ATC radar- r

scopes, and traffic not sighted. Table 1.2 shows Visual informa tion transfer
the distribution of problems between aural and VideolCRT 587
visual media in all citations in which the appro- Instruments,lights, etc. 416
priate field was coded. Publications 309
Charts 203
Remarkably little of the enormous amount
of information in manuals and charts is wrong. 1,515 15
The problem with printed data is usually that it 10,111 100
is difficult t o find, or t o read, or t o interpret,
especially under poor viewing conditions, in
turbulence, or while under stress or high work-
load. For example, late changes in approach

11
clearance, usually involving a change in landing runway, place the pilot in an unexpected high work-
load situation and make it necessary to find, scrutinize, and memorize new data very rapidly.

Among reports from controllers are many concerned with errors associated with overlying data
blocks, tag drops or swaps, and related problems. Radar often fails to portray primary targets; the
presence and positions of these aircraft thus comprise information not available t o the controller.
Tower controllers report problems in sighting traffic visually at airports and difficulties in estimat-
ing relative positions of several aircraft.

Finally, there are many reports from pilots of conflicts and near midair collisions with traffic
not sighted early enough to permit timely corrective action. Workload is clearly a factor in traffic
detection in terminal areas, where these conflicts are most often reported. These failures may
involve a verbal information transfer problem, a visual perception problem, o r both.

TABLE 1.3.- FAILURE TO RECEIVE INFORMATION Table 1.1 indicates that although the
CORRECTLY bulk of information transfer failures can be
traced to the sender of the message rather
Failure to receive or than to the receiver, failures t o receive the
monitor a radio message 1096 of 5002 citations (22%) information correctly accounted for 11% of
Failure to receive visual the citations. Table 1.3 summarizes an
examination of failures t o receive informa-
1 1 I
tion that was available.

In approximately one-fifth of the cases in which there was ri transfer problem with information
that was made available, it occurred at the receiving end of the transfer link.

DISCUSSION

These data make it quite apparent that verbal communication is an imperfect method of infor-
mation transfer. For that reason, much attention has been given t o alternative means of transferring
the information that is now communicated almost exclusively by voice.

Digital data link, making use of augmented discrete address beacon systems, or transponders,
is entirely feasible using today’s technology and is under test at this time. Information would be
presented on cathode ray tubes in the cockpit, either in alphanumeric or graphic form. The mes-
sages could be composed either manually by controllers or automatically by ATC computers.

Attention has also been directed t o the uplink of digital data regarding pertinent traffic and t o
the graphic display of traffic information within the cockpit. This is not proposed as a means of
collision avoidance per se, but rather as a way of putting the flightcrew “in the big picture” regard-
ing the control of their aircraft by ATC. It is thought that with such information, pilots might well
be able t o assume the responsibility for spacing and perhaps other functions now performed by air
traffic control. It is also believed that such information would permit pilots t o detect errors made
by controllers, something they cannot d o now because they often lack the necessary information.

12
 Collision avoidance would be provided by another set of devices, either on the ground or in
the aircraft; collision avoidance and conflict resolution information would also be presented
visually.

There is no doubt that these devices, if implemented, could handle a considerable amount of
information now transferred aurally, Many of the problems now observed would be substantially
alleviated. One of the most serious - in which a pilot acts on a clearance meant for someone else -
might be eliminated if clearances were seen only by the flightcrew for which they were intended.
What new problems might be created, however, by these solutions of the present problem?

Visual perception of data is by no means free of errors, as evidenced by table 1.3. Both at the
perceptual and cognitive levels, numbers are transposed or dropped; expectation can modify percep-
tion of the printed word as it does auditory information. Careful design of graphic displays can help
t o minimize errors in the interpretation of the information presented, but such errors will continue
t o occur. Visual presentation of data, however, is less transient than auditory presentation. The data
can thus be reread by a recipient.

More important is the question of visual channel capacity. ASRS reports indicate that there
are, in the present aviation system, situations in which pilots find themselves fully occupied or over-
loaded by tasks with a large visual perception component (ref. 5 ) . These situations often involve
descents at high speed into busy terminal areas, using complex procedures. Under these conditions,
which require a very close attention t o visual displajjs within the cockpit (especially if there is
weather requiring concurrent attention t o the radarscope), visual scanning for other aircraft is
markedly reduced.

The effect of adding additional displays, some of which require not merely a casual scan, but
careful attention, may well result in overloading the visual channel unless a way can be found to
relieve the pilots of certain of their present visual tasks. The resolution of this problem is central t o
the mitigation of the information transfer problem in aviation operations.

CONCLUSIONS

Based on reports t o ASRS, it is concluded that information transfer problems are responsible
for many potentially serious human errors in aviation operations. Voice communications, in particu-
lar, are a pervasive problem. Technological solutions exist for many problems related t o information
transfer. These solutions, however, may give rise t o serious new problems unless they are imple-
mented with an understanding of the capabilities and limitations of the humans who operate the
aviation system.

13
I REFERENCES

1. Billings, C. E.: Misunderstanding of Communications between Pilots and Controllers. ASRS Third Quarterly
Report, NASA TM X-3546,1977.

2. Billings, C. E.; Lauber, J. K.; Lyman, E. G.; and Reynard, W.D.: The Uses and Limitations of a Voluntary, Con-
fidential Aviation Safety Reporting System. Proceedings of 26th International Congress of Aviation and
Space Medicine, London, 1978.

3. Billings, C. E.; and O’Hara, D. B.: Human Factors Associated with Runway Incursions. ASRS Eighth Quarterly
Report, NASA TM-78540,1978.

4. Grayson, R. L.: Effects of the Introduction of the Discrete Address Beacon System Data Link on Air/Ground
Information Transfer Problems. NASA Contractor Report (in press).

5. Billings, C. E.: Human Factors Associated with Profile Descents. ASRS Fifth Quarterly Report, NASA
TM-78476,1978.

14
 2. INFORMATION TRANSFER IN THE SURFACE COMPONENT OF THE SYSTEM:

PROBLEMS ASSOCIATED WITH BRIEFING O F RELIEF CONTROLLERS

Ralph L. Grayson

INTRODUCTION

A substantial majority of occurrence reports submitted to the NASA Aviation Safety Report-
: ing System by pilots and air traffic controllers contain evidence of problems in the transfer of infor-
mation among the. participants in the national aviation system. Although failures of information
transfer between giound and aircraft are most common, there are also many failures within the
cockpit and within the ATC system. Because such failures are often the precursors of other failures,
it is important to gain an understanding of what they are and how they occur.

The briefing of a relieving controller by the controller being relieved is a pure information
transfer exercise. There is always the possibility that necessary information will not be transferred
and thus there is a potential for later failures of separation. Since the mode of briefing is invariably
oral. the process is subject t o the many sources of error inherent.in oral communications. The ASRS
has received a small but steady flow of reports concerning problems associated with briefing of
relieving controllers. This discussion summarizes an analysis of such reports; the analysis was under-
taken with the intent of shedding light on the nature of the errors involved and their sources.

OBJECTIVE

The objective of this study was t o characterize the types of human errors reported in connec-
tion with briefing of relief in air traffic control operations, and to examine the factors 'associated
with these errors.

APPROACH

ASRS reports associated with briefing of relief (BOR) were counted to evaluate temporal
trends in their submission. A sample of 5 2 occurrences was chosen for detailed study. These reports
were examined and categorized as to reporter, errors committed, and factors associated with the
errors. An examination of the resulting array of data produced some thoughts about ways of reduc-
ing the rate of errors.

For purposes of this study, a BOR problem was considered t o be a failure t o transfer informa-
tion completely and accurately from an air traffic controller operating a position t o another con-
troller who was t o assume subsequent responsibility for the operation of that position.

15
 RESULTS

Figure 2.1 shows the monthly receipts of ASRS reports since July 1976 that were associated
with briefing of relief. There was a gradual decrease in reports until October 1978, which is the only
month during which n o reports were received. Since that time there has been a slight trend toward a
greater number of reports, but this did not prove to be statistically significant. The increase may be
attributable t o changes made in 1979 in the waiver of disciplinary action provided by FAA to
reporters t o ASRS, which had the effect of fostering multiple reporting of the same occurrence.

The source of each of the reports was

determined. Each report was categorized on
Error committed by the basis of who committed the described
error. The results of these categorizations are
Originator I

of report Relieved Relieving Other '

in table 2.1.
1 Total
controller controller , person
Responsibility for the accuracy of a
Relieved relief briefing is clearly that of the person pro-
controller 18 2 1 21 viding the briefing (here, the relieved control-

1:
Relieving ler). It is not surprising, therefore, that most
controller 13 6 23 BOR .problems were enabled by the relieved
Other person 5 1 1 8 controller. The purpose here is not t o place
blame, but to isolate the factors that caused
Total 36 9 7 52 the BOR to be inadequate.

The BOR problems were next classified according to the information transfer problems
reported. The results are shown in table 2.2.

Factors thought to be important in BOR problems were evaluated; they are summarized in
table 2.3. More than one factor was assigned to certain reports.

TABLE 2.2- CLASSIFICATION TABLE 2.3.- FACTORS ASSOCIATED WITH

OF BOR DEFECTS BOR OCCURRENCES
I I
Category

Absent briefing
I occurrences
Number Of I Factor
I
Number of
occurrences

Nonrecall of pertinent information 14

Incomplete briefing Failure of technique 11
Inaccurate briefing Failure of perception 8
Complacency 5
Total occurrences 52 Distraction 4
Message ambiguity 2
Workload 2
Inattent ion 2

16
 L
0

I
z m 0

HlNOW t13d S l U O d 3 t l d O t139WnN

17
 These factors, alone and in combination, were then studied further with the hope of identify-
ing the root causes of the occurrences and the procedures or practices that could have prevented
them.

DISCUSSION

Although the sample consisted of a relatively small number of reports, the safety significance
of briefing errors is indicated by the fact that in most cases the consequences of such errors were
serious. An example is the following report.

Assuming the final position the controller released aircraft B. When B turned down-
wind, aircraft A departed and was immediately in conflict with B. Aircraft A took
evasive action t o climb above B passing over him by 600 ft. Probable contributing
factor - volume of traffic resulted in an incomplete position briefing.

The following sections describe each type of error and discuss the enabling factors observed in
connection with it. Also covered are a subset of occurrences associated with correct briefings and a
discussion of methods for reducing the occurrence of briefing errors.

Failure to Brief Relieving Controller

The responsibility for providing a relief briefing is defined in FAA Agency Order 71 10.78A,
effective July 1978 and related facility orders, all of which require that a controller being relieved
shall orally brief the relieving controller; the content of the oral briefing is specified. ASRS reports
show no evidence of misunderstanding of this precept.

This deviation was caused by the relieved controller failing t o adequately brief his
relief.
* * *
This situation occurred, I believe, because the air traffic controller was not properly
briefed on the recovery profile being used.

Yet in five cases in this sample, no relief briefing was given. In these five instances, the most
frequent cause was attributed t o heavy traffic workload. The relieved controller was too busy with
current traffic t o provide a briefing.

Traffic built to a point where more controllers were needed. The controller who was
working all positions was too busy t o brief me when I attempted to take radar west
and a final position.

* * *

18
 Supervisor came back to sector with a headset and pulled out my headset plug. I
immediately asked him if he was ready for a mandatory briefing and he told me t o
relieve another sector and that he needed no briefing. I consider this a double stan-
dard since controllers are mandated to give a full briefing before being relieved
from duty.

In the latter case, a supervisor assumed the position, stating to the relieved controller that he
needed no formal briefing since he was already familiar with the traffic situation. Although this may
well have been true, he might still have been unaware of other pertinent information that could
have been transferred in a formal, checklisted briefing.

Incomplete o r Inaccurate Briefings

Twenty-one occurrences involved an incomplete briefing (failure to transfer relevant informa-

tion). Twenty-six involved the transfer of inaccurate information. The factors in these two cate-
gories were examined.

Of the 2 1 incomplete briefings, 9 involved a failure to recall information. This factor was also
found in 7 of 26 inaccurate briefings. Examination of these occurrences indicated that the informa-
tion not recalled was available at the position in about half the cases.

The new controller was told that aircraft A had been stopped at FL330. Tape
recordings indicate that A was never recleared t o FL330.When he reported in he
advised climbing t o FL370. The controller acknowledged but did not verify the
assigned altitude. Later aircraft B requested descent clearance and the controller
noticed the A and B targets merging as A was leaving FL348.

In other cases, the information was not immediately available because of incorrect recall of a
clearance, failure t o mark flight-progress strips during a holding operation, no transponder code or
an incorrect code, and similar irregularities.

We were requested by an inbound sector controller to move 21,000 traffic t o

19,000 and asked what an aircraft at 23,000 was doing there. We were then advised
that the adjacent sector was holding at 20,000 up. I was not aware of the holding
stack and neither was the controller I relieved. He should have been advised of the
stack by the coordinator.
* * *
Aircraft A called stating he was on frequency at 6,000. Assuming it was his initial
call-up I gave him the beacon code on the flight-progress strip. A few seconds later
he acquired about 1 mile behind aircraft B. It was found in the system error investi-
gation that (1) the relieved controller had been in contact with A and amended his
clearance assigning a heading of 090°, (2) had forgotten about A at the time of
relief, (3) the aircraft had not been assigned his correct code which would have
resulted in auto acquisition.

19
 Perceptual errors were uncommon in incomplete briefings but were the next most common
factor in inaccurate briefings. In at least half of the six cases cited in this category, the information
was correct in the full data block or the flight-progress strip, but was not read or was misread during
the briefing. In one case, there was a conflict between the two sources of information and the
wrong one was used.

During the briefing of relief checklist, I was advised by the relieved controller that
B was level at 7,000 due t o A westbound a t 8,000. The adjacent Center called and
advised that the two aircraft passed within 1/4 mile at 7,000. No altitude readouts
were noticed by me. I only had the information from the relieved controller.

Technique failures were cited in one-fifih of the incomplete or inaccurate briefings. In most
cases, the information was available but was not utilized in the briefing. In the cases where it was
not available, this was because of a previous error by the relieved controller.

When I was briefed by the relieved controller, he did not mention anything about
traffic holding at XYZ, nor was anything written about it at the sector. Approach
control later asked what altitude the inbounds were out of if they were descending.
Later approach control advised that the F4’s were clear of XYZ. Investigation
revealed that holding had been coordinated a t XYZ but no written entry was made.
The trainer controller also failed to mark it down and did not brief his relief.
* * *
A was descending to 8,000. The aircraft track, data block, and flight plan were acci-
dentally removed from the radarscope and computer as A left 28,300. One minute
later I was relieved and briefed my replacement on all of the aircraft represented by
data blocks. I did not realize A had been removed. SMT B had cancelled IFR and the
incorrect computer identification number was used.

A few of these occurrences involved system factors. In one case of nonrecall, the flight-
progress strip had not been delivered to the position and the aircraft tag had dropped from the
radarscope; in a second the aircraft was on the wrong beacon code. In a third case, a controller was
being relieved after only 3 min on the sector. Several other human factors were implicated in the
reports, however. Distractions were cited in four instances, workload in two. Complacency or
inattention were thought t o be factors in seven occurrences.

While briefing my relief, I was interrupted by Center several times on the handoff
lines. Due to the interruptions (I felt) I overlooked a point-out for sequence that I
had previously given another controller. My reliever assumed the position and over-
took the sequenced aircraft with another faster aircraft. Computer tapes indicated
that I had taken the handoff via computer and rogered the aircraft checking in on
frequency.

* * *
The radio button was off when aircraft A said he did not want a visual to run-
way 32. I was being relieved by another controller. The two aircraft got extremely

20
 close over the outer marker. Probable cause: either lazy relief procedures or radio
problems.

Errors by Relieving Controller

In nine occurrences, an error was committed by the relieving controller. These reports were
examined t o determine the error patterns.

In one case, the relieving approach controller assumed that an observed target was an arriving
aircraft that had previously been transferred to tower frequency. He then vectored a subsequent
arriving aircraft t o within 200 ft of the preceding arrival aircraft.

Why did it occur? Poor radar due t o weather, weather conditions, misidentified
target, poor cooperation among controllers, all positions not being alert to the
situation.

In two instances, failure t o recall pertinent information during the briefing resulted in a haz-
ardous incident.

We were cleared t o tower frequency just prior t o the outer marker and made another
request for clearance for an approach. No reply, went to tower frequency and were
cleared to land. Requested approval for approach which was granted. Started high
rate of descent, broke out at 800 ft at middle marker too high and too fast. Cleared
back t o Approach Control. Controller apologized for the mixup, claimed he had just
relieved another controller and was told we had been cleared for the approach.
* * *
When I relieved the previous controller, I was told that aircraft A was released t o
Sector 23 and on t o Sector 35. Sector 35 called and asked the status of A. At the
time I did not recall what it was. I later learned that A was in Sector 35 airspace and
n o handoff had been made.

In one instance, the relieving controller assumed that a VFR aircraft receiving Stage 111 radar
service would be assigned 6,500 instead of 7,000 which had been assigned by the previous
controller.

Position relief had occurred just prior to the incident. Possible cause: problems in
briefing of relief. Better briefing would lead to a better awareness of the situation
by the relieving controller.

A controller moved t o a handoff position t o assist a controller who had become very busy.
The coordinator pointed out an aircraft stating “This aircraft is for an ASR,” which was heard as
“This aircraft is VFR.”

I believe that I was not properly briefed and that the radar controller and the coor-
dinator were not communicating properly.

21
 In a similar case the relieving controller received a message correctly but misinterpreted its
meaning.

The relieved controller advised me that aircraft A was “going over the top” of air-
craft B. I believed this to mean an approach t o a 360” overhead. Failure t o under-
stand or my assumption that A was an overhead approach contributed to this inci-
dent as well as failure to inform or make certain that the relieving controller
understood the situation.

Thus a variety of human error factors can be present when the relieving controller makes an
error in the period immediately following position relief.

Control Errors Associated with Relief Briefings

Several reports related t o errors that occurred in association with relief briefings that were not
in themselves faulty. Errors or oversights by the relieved controller prior t o the briefing can result
in a reportable incident after the relieving controller has assumed the operation of a control posi-
tion; other control errors occur shortly after the transfer of responsibility for the position. Both
relieving and relieved controllers reported errors in control during the conduct of the briefing.

Evasive action should have been taken but it was too late when finally noticed by
the controller. He forgot to turn A in on the localizer while he was briefing another
controller who was relieving him.

It is noteworthy that a relief briefing is itself a distraction, albeit a necessary one, involving
increased workload for the controller providing it. Though the relieving controller may receive a
complete and accurate briefing, it also appears that it commonly takes him a few minutes to
become fully comfortable with the traffic situation. During these periods, mistakes can occur if
controllers are not especially alert t o them.

REDUCTION O F ERRORS.IN BRIEFING OF RELIEF

It is selfevident that controllers should and d o strive t o prevent errors in relief briefings, and
the FAA has stressed the importance of this. Certain lessons in the reports studied here may further
assist in reducing these errors.

The three most common factors in these occurrences were nonrecall (failure t o remember per-
tinent information), perceptual errors (failure t o perceive or notice pertinent information), and
technique errors (failure to transfer pertinent information). The fact that pertinent information
may not be immediately available at the time control is transferred points up the importance of
written notes to assist the recall process. “Scratch-pad’’ notations can be helpful t o the briefing con-
troller and more helpful to his relief during his initial minutes on the position. Standardization of
procedures and phraseology in the conduct of briefings can reduce ambiguity and the likelihood of
misinterpretation of the information being transferred.

22
 In cases in which traffic workload tended to prevent or inhibit the briefing process, the basic
problem was that the briefer had to retain his grasp of the changing traffic situation and maintain
radio communications while delivering the briefing. This is especially difficult because most con-
trollers are conditioned, properly, to respond t o any radio or telephone call as a first priority
matter. However, when the volume is such that the briefing is being delayed, this priority can in
some cases be modified judiciously. If no emergency is in progress, most routine radio contacts can
be delayed briefly without causing problems, or a second specialist can monitor the frequency,
accumulating calls on standby while the briefing proceeds. If this or similar techniques cannot be
used, then a better technique would be to delay relief temporarily, allowing the relieving controller
t o monitor the situation until a formal briefing opportunity presents itself. Considerable informa-
tion can be transferred within a few seconds given the undivided attention of a reliever who has
monitored for a while. In any case, every effort should be made t o conduct a formal briefing -
especially under heavy traffic conditions where there is the least margin for error.

ASRS reports indicate a presumption by reporters that the relieved controller is solely respon-
sible for the completeness and accuracy of the information contained in the oral briefing.

I relieved the radar controller and received a briefing on three jets landing in the
XYZ terminal area. Nothing was said in the briefing about aircraft A at 12,000 with
no transponder. The jets had received clearance t o 11,000, 9,000, and 7,000 from
the previous controller. I assumed the position and continued vectoring the jets in
trail and failed t o notice the strip at the very bottom of the board on aircraft A at
12,000.

This indicates a strict interpretation of the language of FAA agency order 7110.78A. It is
obvious that the relieved controller must bear responsibility for any information that is known only
t o him. However, much of the data transferred during the briefing can be verified by a thorough
check of the radar display, contents of full and limited data blocks where available, and flight-
progress strips, which are nearly always posted and available concurrently with radar information.
Corroboration of the data supplied by the relieved controller using all available sources might
improve the quality of the briefing process if this were made a direct responsibility of the relieving
controller.

This is also important because the relieved controller may well be tired after handling heavy
traffic, or bored after a long period of low activity. The relieving controller should be alert t o pos-
sible errors or omissions in the information being given him. Though the responsibility for the
briefing remains with the relieved controller, his counterpart can be of substantial help t o him in
discharging it.

SUMMARY

Failures in the briefing of relieving controllers, as observed in ASRS reports, appear t o be a

continuing problem, though one that has been alleviated to an extent by various efforts over the
past 3 years. Failure t o recall information is an important enabling factor; failure to check the infor-
mation and t o transfer it when available are also major factors. The elements of such failures are

23
always present in this critical task; relieving controllers can be of appreciable assistance in minimiz-
ing errors in the information transfer process.

24
 3. INFORMATION TRANSFER IN THE SURFACE COMPONENT OF THE SYSTEM:

COORDINATION PROBLEMS IN AIR TRAFFIC CONTROL

Ralph L. Grayson

Do not allow an aircraft under your control to enter airspace delegated to another controller
without first completing coordination. - Air Traffic Control Handbook 7110.6SB.

INTRODUCTION

Coordination is a term used widely in the air traffic control (ATC) system to describe control
activities that affect aircraft traversing jurisdictional boundaries of airspace. It is usually used in
reference to the transfer of information between air traffic controllers who control separate seg-
ments of airspace through which controlled aircraft are or will be passing. This information transfer
process is used by the controllers in developing and executing operational plans for controlling the
traffic through their respective airspace segments. Although this coordination is primarily an opera-
tional process, it comprises a significant information transfer element; as a result, a discussion here
of its pertinent issues and problems is most appropriate.

Coordination is the most pervasive, and perhaps the most complex, aspect of the U.S. ATC
system. It is widely recognized as a controller function that is particularly vulnerable to human
error. Furthermore, a coordination failure usually results in the loss of standard separation dis-
tance between aircraft. The Aviation Safety Reporting System (ASRS) data base contains many
items describing occurrences involving coordination failures. This report presents the results of a
study of that segment of ASRS data and the conclusions arising from it.

OBJECTIVE

The objectives of this study were (1) to identify the air traffic control coordination problems
that were reported t o ASRS, (2) to classify them, (3) t o determine the circumstances that appear to
encourage coordination failures, (4)t o examine the human and system factors involved, and ( 5 ) t o
consider possible means of reducing the rate of such failures.

ATC COORDINATION

As is implied by the general definition above, an ATC coordination transaction is a communi-

cation between controllers (not necessarily verbal) the purpose of which is t o arrive at and execute
a plan for handling aircraft as they pass from the jurisdiction of one controller t o that of another.
Coordination is, in effect, a contractual agreement between two controllers arrived at by negotia-
tion and binding them t o perform in accordance with that agreement once it is made. The process
involves the following steps, all of which are composed substantially of information transfer actions.

25
 1 , Communication by one controller to another of a request for Coordination.

2. Responsive message of concurrence or nonconcurrence by the controller receiving the

’request; nonconcurrence will result in a discussion of alternative action.

3. Agreement on a plan of action and who is t o execute it.

.
4. Execution of the plan.

For the purposes of this study these four steps are referred to as initiation (step 1); agreement
(steps 2 and 3); and execution (step 4).

Each of these steps has to be taken with due consideration for time available to complete the
entire process. Allowance must be made for communications with other controllers or flightcrews
and their responses. Furthermore, each step must be completed in a timely manner with an appro-
priate time buffer for unforeseen contingencies such as delay in establishing radio contact. The best
plans are to no avail if the enabling clearances cannot be implemented in sufficient time t o preclude
loss of separation.

Coordination is frequently classified as “interfacility” or “intrafacility” to indicate the number

of facilities involved. This distinction has not been used in this study because the errors that result
in coordination failures appear t o be present in roughly equal numbers, regardless of the physical
locations of the controllers.

Several different media are used for transfemng coordination information. For example, most
interfacility coordination is done by telephone. Also, large sectorized facilities (centers, large
TRACONs) use the telephone for communication between different positions within the facility -
intrafacility coordination. Among smaller facilities, and also within multisector facilities where the
working positions are close together, less formal means often are employed - direct voice, informal
or abbreviated messages, hand signals, and others.

Prior knowledge of an aircraft entering a sector is essential t o the planning of coordination.

This is normally provided by transmitting flight plan information t o affected control sectors by
printing out flight-progress strips before the aircraft enters the control sector. A second procedure
requires either a radar handoff in areas of radar coverage or prior verbal coordination before an air-
craft enters airspace controlled by another controller. The following is from the Air Traffic Con-
trollers Handbook, 7 1 10.65A:

Inter- and intra-facility radar handoffs shall be accomplished in all areas of radar
surveillance. The transfemng controller shall complete a radar handoff or obtain
receiving controller approval before the flight enters the receiving controller’s
airspace.

Routine transfer of flight data is communicated almost entirely by automated equipment,

Flight plans are entered into computers at the en route centers where they are stored until they
become timely; flight-progress strips are then routed t o terminal facilities and appropriate center
sectors. Radar handoffs are automatically initiated by most en route center computer systems. This
precludes oversight by the transferring controller and normally results in handoff to the correct
sector.

Routine transfer of flight data and radar handoffs prior t o transfer of control appear to present
few problems. However, if the originally planned flight must be changed in order t o maintain
standard separation as an aircraft passes from one sector to another, some type of specialized
communication is necessary between the controllers involved. Most coordination failures reported
t o ASRS concern such complex operations and this study concentrated on them.

APPROACH

The first step in the approach was t o retrieve a representative sample of ASRS reports involv-
ing coordination failures. The initial search yielded 1,801 reports out of a total of 9,795 received
during the period from May 1978 through February 1980. A subset of 200 reports of unique
occurrences was generated by selection and screening every ninth report from the original sample;
these would be evenly distributed in time throughout the initial data set. Since the study was
directed to controller-enabled ATC coordination failures, any report in the initial selection that
described a flightcrew-enabled failure was replaced with a report on a controller-enabled problem
received by ASRS at about the same time. This procedure provided an overview of the time period
of the data base in addition t o ensuring that the sample group consisted oivalid, controllerenabled
coordination failures.

Reports were first classified according t o the phase of coordination - initiation, agreement, or
execution - in which the failure occurred. The initiation phase citations consisted of those,
instances in which no effort had been made to begin coordination or in which the process was
interrupted before a request was communicated t o the appropriate controller. The agreement phase
was chosen when initiation had occurred, but no agreement had been reached. The third phase,
execution, was chosen if an agreement had been reached but the controllers responsible for carrying ’ .
out the plan failed to d o so. Errors in execution included failure t o implement the agreement or
doing it in sucli a way that loss (or potential loss) of standard separation resulted.

Further examination revealed the mode of failure that occurred in each case. If appropriate
communications were not made or action was not taken, the failure was considered to be of the
“absent” type. If some communication or action had begun but was not completed, it was labeled
“incomplete.” If the information used in decisionmaking was incorrect, or if implementation was
incorrect, even though the agreement was understood, the failure was labeled “inaccurate.”

Assessment of human factors - errors and the predisposing conditions for those errors - fol-
lowed where these were discernible in the reports. These classifications were based on an inductive
evaluation of the factors stated or strongly implied by the reporter. In some cases, however, the
nature of the error could not be inferred from the reporter’s description of the occurrence; this was
more often true of the predisposing conditions for errors.

Finally, consideration was given t o possible ways of reducing the probability of coordination
failures: in doing so, only present system capabilities were considered.

27
 RESULTS

The results of the analysis of the 200-report sample were distributions of (1) the outcomes of
the coordination failures reported and (2) the human factors (types of errors, predisposing condi-
tions, and environmental factors) related t o the failures. In the analysis, the coordination process
was viewed as the combination of decisions, communications, and actions depicted schematically in
figure 3.1. The three elements of coordination - initiation, agreement, and execution - are shown
as occurring in logical sequence. Success in meeting each main requirement of each element leads to
a successful coordination action whereas failure at any point causes the coordination attempt t o
fail. The diagram further depicts the usual consequences of the coordination failure as observed in
ASRS reports.
'

In the discussions t o follow, the first topic covered is the outcome of coordination failures.
This is followed by an examination of the modes of failure observed in the study set of reports. The
human factors associated with the failures are discussed next, followed by a consideration of various
means of reducing the number of coordination failures.

Outcomes of Coordination Failures

Each of the coordination failures produced, as its direct outcome, a condition wherein an air-
craft that was involved in the coordination action was caused t o proceed along a flightpath leading
t o a potential or actual loss of standard separation. Altitude anomalies predominated; the next most
frequently reported anomaly was lateral displacement from the correct flightpath. In a few cases an
incorrect speed resulted. As indicated in figure 1, when other traffic was involved, these flightpath
anomalies frequently led t o loss of standard separation. In the data set studied there were 134 such
occurrences, ranging in seventy from nonhazardous conflicts t o high-risk situations in which midair
collisions were narrowly avoided. The seventy appeared to be largely a matter of chance not corre-
lated with the kind or cause of the coordination failure involved.

Coordination Failures

Reports were examined' as t o the phase of coordination during which the failure occurred
(initiation, agreement, or execution); these sets were then further subdivided by failure mode.
There were three kinds of failure among the reported occurrences.

1. Incomplete: A phase of coordination was considered incomplete if the first communica-

tions for it had begun but not proceeded to a logical end. This could range from a telephone call not
answered in a timely manner t o a failure to cany out all items of an agreed plan of action,

2. Inaccurate: A phase of coordination was considered to have been performed inaccurately if

the plan agreed t o was based on incorrect information or if an error was made in executing the plan,

3. Absent: A phase of coordination was considered t o have been absent if there was a failure
to initiate coordination, failure to reach an agreement after initiation, or failure t o execute the plan
agreed upon.

28
I IN IT AT ION^

0 correct?

INCORRECT
FL IGHTPATH

DEVIATION

e= Timely?
SYSTEM
ERROR

SYSTEM
ERROR

LTSS

COORDINATION COORD &TION

I L U R E I

Figure 3.1 .- Process of ATC coordination.

29
 I finding is the indication that in 122
I
Phase
reported occurrences controllers did ?.ot
Fatlure 4
Toial
mode Initiation Agreement Execution I
begin coordination, even when there was an
! operational need for it or when a directive
Incomplete 5 10 1 16 in force required it. These data leave little
Inaccurate 2 16 ‘ 40 58 doubt about where the main problem with
Absent 121 2 3 126 ATC coordination lies. However, the data
I also indicate significant difficulty with the
Total 128 1 28 44 200 next most frequently cited failure mode -
I I
errors in executing agreed-upon coordina-

The data in table 3.1 also suggest that failures occur more often when action is incumbent on
individual controllers, as in the initiation and execution phases. The fewest failures occurred during
the agreement phase where, by definition, a dialogue is being conducted - two controllers are
exchanging information to arrive a t an agreement.

Human Factors

The human factors observed in the reported occurrences consist of six kinds of errors and nine
kinds of conditions that were related t o the errors and seem t o have predisposed their commissions.
Table 3.2 lists and defines these human factors.

Each report was classified according t o the enabling error and its predisposing condition. In
some of the reports there was n o indication of what the error was; in others the nature of the pre-
disposing condition leading t o the error could not be discerned. However, most of the reports con-
tain the reporter’s statement about the cause of an occurrence o r a factual description of concurrent
events and traffic that makes it possible t o determine a probable cause.

Coordination errors- Table 3.3 shows the distribution of the types of coordination errors that
were found in the study .data set; the errors are listed by failure modes. (The errors discussed below
are defined in table 3.2.)

Failures of perception, the most frequently occurring errors, resulted in instances in which a
controller based his plans or actions on inaccurate or incomplete information. In coordinating a
planned action, the controller may have failed t o take into account pertinent traffic that was not
, observed or otherwise made known to him, or the execution of a workable plan may have been
~
inaccurate for similar reasons.

In the following example, the reporting controller agreed t o separate aircraft in another con-
troller’s ai-rspace without a complete understanding of the relative flightpaths of the aircraft
involved. Oral inquiry to the initiating controller resulted in further inaccuracy in the position of
an aircraft as perceived by the controller. The end result was a coordination failure due to inaccu-
rate execution.
 TABLE 3.2.- HUMAN FACTORS ASSOCIATED WITH COORDINATION FAILURES

Human errors

Nonrecall Failure to retain information received that bears on present or future traffic
situation

Technique error Use of unauthorized or inccrrectly chosen procedures in the formulation or

execution of a coordination plan

Perceptual error Use of incorrect or incomplete data received from an external source, sensory
acquisition, or analysis as a basis for decisionmaking

Failure t o monitor Failure to check sources of information in a routine and timely manner

Message inaccuracy Contents of a message interpreted differently from originator’s intent

Misidentification Location or other pertinent data applicable to one aircraft utilized for another
aircraft

Predisposing factors

Distraction An influence or activity that diverts an individual’s attention from his normal
responsibility

Excess workload Traffic situation and surrounding circumstances such that the capacity of avail-
able human and system resources t o respond t o requirements is insufficient

Experience/training level Inadequate individual experience or level of training is a factor in the choice of
an inappropriate action or in the improper execution of an appropriate action

Complacency Self-satisfaction which may result in impaired vigdance based on an unjustified

assumption of a satisfactory system state

Airspace configuration Boundaries of controllers’ designated jurisdictional airspace so designed that

they contribute to excessive coordination requirements or t o failure in the
conduct of coordination

Procedural problem A coordination or separation procedure that induces controller error because of
complexity, cumbersomeness, or unfamiliarity due to infrequent usage

Automation mind-set Controllers unfamiliarity with procedures used under systemdegraded condi-
tions when automation features normally in use are unavailable because of
equipment or software failures

Equipment failure Conditions resulting from hardware failures that have an adverse effect on infor-
mation available or on communications facilities

Interpersonal relationship A personality conflict between individuals involved in the coordination process

31
32
 We approved aircraft B to climb to 10,000 after adjacent sector released the air-
space to us without mentioning they had 9,000 traffic, aircraft A, about 23SE. They
did force a data block t o our scope. A few minutes later the location of B was asked
of Approach Control since he still was not on the scope. Aircraft A was still about
15SE. Approach advised that B was 3W of the VOR. We were not concerned since
B was a jet and would easily be gone before A was too close. Moments later a lim-
ited data block was observed 4-5 S of the VOR. Not being sure who it was, I called
Approach to stop B at 8,000 until radar separation could be assured. Approach
advised A was in radar contact and requested altitude t o climb to when they were
clear. 10,000 was approved reference A. The targets were observed t o merge then
auto acquisition occurred on the center discrete code. Aircraft A saw B and asked
what the traffic was that climbed through him. 6

Failures of technique, the next most frequently cited error, resulted from a controller’s choice
of procedural steps or applications that failed t o establish or maintain standard separation. The
individual actions may have been the everyday tools used by the controller, but their use was
inappropriate, untimely, or ineffective. In all these instances, the controller expected that separa-
tion would result even though he had not provided positive control measures that would have
ensured the maintenance of minimum separation standards.

One example is allowing or requiring controllers to separate aircraft that are in the jurisdic-
tional airspace assigned t o another controller with no definite, mutually agreed upon plan for main-
taining separation. In this ASRS report the reporting controller handed off an arriving aircraft t o
approach control at an altitude above other en route traffic which was pointed out t o the approach
controller. Later, another aircraft became a factor and, again, a point-out rather than full-scale
coordination was used t o advise approach control of the new factor.

Aircraft A was cleared by Washington Center from Swann t o Patch with proper
coordination for a hi-TACAN penetration. Washington handed A off t o N.Y. Center
sector 17/18 then handed off A t o Dover Approach at 150 requesting two turns in
the holding pattern. On handoff aircraft B was pointed out t o Dover as traffic for A.
At that time I gave Dover control, for A’s descent, reference B. Minutes later, Dover
called me requesting control from 15,000 t o 14,000 on A. I then gave Dover control
t o 14,000 in the holding pattern and his control for the penetration of A again refer-
ence B. As A began his last turn in the pattern swing, an aircraft C became traffic
for A. At this time I called Dover and pointed out C as additional traffic for A. I
also asked the controller if he was still watching B. He acknowledged. When A hit
the penetration fix both B and C were traffic factors. Again I called Dover and asked
if he was separating A from the observed traffic. He said yes. My radar display
showed less than minimum separation. C did not see any traffic. B saw A.

Although this procedure is not specifically prohibited, its choice in this case can be judged an
inappropriate one since it required the approach controller t o monitor two aircraft that were
neither under his control nor in communication with him and t o separate three aircraft in airspace
under the jurisdiction of another controller. If the reporting controller had retained control of the
arriving aircraft until it was positively separated from other traffic under his control, the handoff
would probably have been routine and the incident might not have occurred,

33
 Technique errors affected the initiation phase of coordination primarily through controller’s
substitution of some other traffic control method for normil coordination. It also affected the
execution phase in much the same way.

There were 18 reports in which controller failure to recall a previously known coordination
factor was cited as the enabling human error. The forgotten factors ranged from the presence of an
aircraft to steps in executing a plan. In the following example a local controller fGrgot that an air-
craft under his control, cleared for a touch-and-go landing, was IFR; as a consequence, there was an
error in executing a routine coordination plan.

Aircraft Apwas on a localizer 27 approach t o satellite airport with a missed approach

left turn to 160°, climb t o 3,000 ft. Tower had other departures and called saying
that B would be ahead of A. Shortly thereafter, Tower called t o ask if we (Approach
Control) had A in radar contact. B was radar-identified and told of traffic less than
1 mile. Aircraft A tagged on the traffic. Approach Control turned the airplanes t o
avoid each other. When A made the touch-and-go, Tower forgot he was IFR and
launched aircraft B.

Although this is an example of a coordination failure in the execution phase, the failures most
frequently caused by nonrecall errors are in failing to initiate coordination when controllers simply
forget about traffic that is present and in need of coordination action.

A message inaccuracy error resulted when the recipient of a message either heard incorrectly or
misinterpreted the contents of the transmitted message. These failures ranged from severe garbling
to use of nonstandard or unusual phraseology. The expectation factor was also prominent in several
cases in which the recipient of the message heard what was expected or wanted rather than what
was actually transmitted.

Misinterpretation of a request may result in approval which can lead to hazardous occurrences
as in this example.

Ground controller requested to taxi aircraft A across runway 36 at taxiway Foxtrot

(intersection of runways 36 and 3) en route to runway 33. Local controller then
cleared aircraft B for takeoff on runway 3. The local controller only heard the word
“across” and misunderstood what the ground controller had said. Aircraft B was
forced to lift off sooner than usual and flew over the tail portion of taxiing air-
craft A .

The rernaining errors were few in number and widely distributed. For example, one error in
aircraft identification was reported. It appears that widespread use of discrete beacon codes and
alphanumerics on the controller radar displays has been effective in providing positive identification
of individual aircraft. Further, the National Beacon Code allocation plan virtually ensures that all
aircraft within the jurisdiction of one automated facility will be assigned a unique transponder code.
Positive identification can then be maintained as aircraft traverse ATC jurisdictional boundaries.

Only five instances were reported in which failure to monitor led to a coordination failure.
I This suggests that maintenance of aircraft identity by the automated system requires less detailed

’ 34
monitoring by the controller. He can resume scanning the display after a diversion or distraction
and there is no necessity to correlate radar targets with the aircraft identification.

Some radar approach-control facilities are not yet equipped with alphanumeric displays.
Others are provided only limited information by the automated system. Failure to monitor under
these conditions can result in coordination errors without regard to the available display, as shown
in the following example.

Both aircraft were on practice instrument approaches to the same runway on oppo-
site direction courses. One was in contact with tower, the other with Approach
Control. They were turned into each other after missed approach. Local procedures
are adequate and clear. This incident occurred because of human inattention and
concentration on the situation.

Predisposiiig coizditioiis- The study data set was analyzed t o determine if a condition was
alleged which may have predisposed or contributed to the enabling error by the individual involved.

In 137 instances no determination could be made from the TABLE 3.4.- PREDISPOSING
infomiation contained in the report. If a reasonable inference CONDITIONS
could be made from the described circumstances, or if the reporter
said specifically that a predisposing factor was present, an appro- Excessive workload 19
priate count was taken. Particularly in reports that involved two or Distraction 18
more controllers, more than one predisposing condition may have Experienceltraining level 17
been present. Complacency 8
Airspace configuration 8
Table 3.4 presents the distribution of predisposing conditions Procedural problem 8
that could be ascertained from the study group of reports. The Automation mind-set 8
relative frequencies of distraction, workload, and complacency are Equipment failure 6
coiisistent with subjective impressions of ASRS researchers about Interpersonal relationship 1
-
the distribution of these predisposing factors throughout the
Total 93
entire ASRS data base. In this case more than one condition could
be associated with a single report (93 citations in 63 reports).

All of the excessive workload citations in table 3.4 were based either on the amount of traffic
in a sector or on the existence of a contingency operation of some sort, such as traffic that had to
divert around a thunderstorm. In this example an arrival controller made a technique error while
dealing with the workload brought on by heavy arrival traffic.

I was departure controller climbing aircraft A to 10,000. Aircraft A had just leveled
when I noticed an arrival tag on a heading that would violate my airspace. I quick-
looked and arrival aircraft B was crossing in front of my departure. I issued traffic,
eleven o’clock, 1 mile out of 10,200. Aircraft A said “We’re looking a t him.” The
arrival function was overloaded and the inbound traffic was being vectored all over
the sky. The arrival controller was working three fixes and flow control was
ineffective.

The disrructioits consisted of extraneous communications within the facility or diverted atten-
tion connected with coinbining or, more often, de-combining positions. In the latter circumstance

35
I there were several complaints about interference on the part of supervisors attempting t o assist in
I
complex situations. The following is an example.

Developmental controller, monitored by supervisory personnel, allowed two aircraft

t o come within 1 mile or less of one another at the same altitude in IFR conditions.
Supervisor took control of frequency to S-turn aircraft A on final approach for
spacing. He became so preoccupied with vectors that he overlooked overflight air-
craft B and aircraft C on final approach converging at 6,000 ft. Developmental
tried t o cut in to ensure descent of aircraft C but was unable to obtain the frequency
and also tried t o tell supervisor to no avail. Developmental tried to alert monitor
controller who was not aware until aircraft had passed.

There were 17 reports about the experience or training level of one of the participants being a
factor; the reported occurrences ranged from failure to recognize the need for coordination t o sys-
tem errors. In some cases, the reporter alleged that supervision of a trainee developmental controller
diverted his attention from either initiation or execution of the coordination plan. In others, a rela-
tively inexperienced controller working without direct supervision committed an error in the execu-
tion of the plan. This report is illustrative.

A handoff from Departure Control was taken by Sector 1. They were unable t o
climb the aircraft immediately so they called Sector 2 with a point-out (should have
been a direct handoff at that altitude). I heard the D2 controller say “radar contact,
my control.” The aircraft was rapidly approaching an approach control boundary.
We then observed the aircraft’s Mode C altitude climbing into aircraft A at 6,000. I
believe the error was caused by a lack of communication and low experience level
among the Sector 1 and 2 controllers, both developmentals. Most controllers with
more experience would have specified the altitude they wanted the aircraft climbed
t o before accepting a handoff or point-out. I would guess the average experience
level in that area is less than 5 years.

In those cases in which controllers made errors ascribed to complacency, the controllers
received information that should have caused them to recognize the traffic situation and then to
take appropriate action; however, in these instances the controllers did not so react. In the follow-
ing example, the controller assumed that the second aircraft was operati,ig VFR since he had no
inbound IFR flight plan. In addition, the local weather should have indicated that VFR flight could
not be conducted at 6,000 ft.

Aircraft A inbound descending t o 6,000 ft 20 miles north of XYZ cleared for a VOR
approach. Aircraft B reported 1 0 south of XYZ in foreign airspace at 6,000 descend-
ing. At this time it was not known that B was IFR. B crossed over the VOR at the
same time as A in IFR conditions. Both aircraft landed without seeing each other,
After both aircraft were on the ground, Center called with inbound on aircraft B
stating that the telephone was out of service for normal communications from the
foreign approach control facility.

Airspace configuration was indicated in eight reports of coordination failure in which aircraft
were under the jurisdiction of different controllers but operating in proximity. This situation can
occur along any jurisdictional boundary and it is particularly notable where two busy airports are
located in the same vicinity. This report is an example.

AR2 had four IFR inbound aircraft, three from the north and east of airport and
SMA A from the south. He decided t o vector SMA A to the west for right downwind
t o runway 13R. A R l was using 13L for VFR arrivals. During coordination with the
adjacent approach control it was understood that they had a departure t o go on
their runway 13L, so aircraft A would have t o be kept close in to the airport.
Actually, the departure from the adjacent airport was using runway 17L but
reporter missed this information. Result was aircraft A passed opposite direction
traffic with 1-1 /2 miles separation.

Reporter suggests that the adjacent approach control should control all IFR and
Stage 111 arrivals since they have more sophisticated radar equipment and to reduce
coordination requirements.

In the study group eight reports indicated that a procedurd problem was a factor. In this
example the controller complains of consistent noncompliance with the procedure set forth in the
interfacility letter of agreement.

SMA A being vectored t o intercept bridge for right traffic entry from the west. I
observed traffic inbound from the south for the bridge. Aircraft A was radar contact
and came on frequency 2 miles west of the bridge heading north for airport. This is
a violation of the Letter of Agreement which requires all aircraft t o be vectored over
the bridge unless otherwise coordinated. Such failures often result in potential con-
flicts with traffic already in the areas.

Automation mind-set is characteristic of some controllers in automated facilities who have

little or n o experience in a nonautomated work environment. Reliance on full system capabilities at
all times becomes the standard or normal operation. If the system is degraded for any reason, either
hardware o r software, some controllers lack confidence in their ability to perform adequately with
less than the complete system. For example, some controllers have asked t o be relieved because
they felt incapable of using the broadband radar system.

This condition appears t o heighten the probability of error, both because of actual skill limita-
tions and because of lack of confidence on the part of the controller.

Eight reports in the study group cited this problem as a factor in a coordination failure.

Equipment problems for which no redundancy is provided can increase the requirement for
coordination t o replace those functions that are normally supported by system equipment. Substi-
tute communication equipment - required for communication between controllers and between
pilots and controllers - may be unfamiliar t o the users or unwieldy and time consuming. Reversion
to broadband radar from computerized displays o r t o a nonradar environment may sharply reduce
the capacity of the controller in terms of traffic volume.

37
 Increased coordination requirements almost inevitably result from loss of major system com-
ponents, and these additional requirements increase both the probability of error and deviations
from normal operating procedures.

Interpersonal relationships characterized by personal conflict between the participants

attempting coordination were cited in only one report. If an individual fails t o cooperate in any
phase of coordination, there can be a failure resulting in hazardous consequences. Failure t o
respond t o a request, failure to agree to an otherwise workable plan, or failure t o execute due to
personality clashes between participants, o r even t o delay any step of the process can create a cli-
mate for failure of coordination.

DISCUSSION

The foregoing examination of coordination failures showed they tend t o be caused by human
errors of a complex and subtle kind. A more detailed look at these errors and the role management
plays in controlling the conditions that are associated with them can suggest some approaches to
reducing the frequency of their occurrence.

Coordination Errors

Table 3.1 shows that the most frequzntly reported coordination failures are “absent initiation”
and “inaccurate execution.” Further, table 3.3 shows that these two failure modes are triggered
almost equally by errors in technique. and perception on the part of the controllers participating in
the coordination transactions. Of the two kinds of errors, mistakes in choice of technique may be
the more insidious because the controller is frequently led into the position of not beginning the
coordination process in time or executing it incorrectly because he is busy applying some different,
ad hoc, method of separation t o the traffic for which he is responsible.

Errors in coordination technique- Several reports of coordination failure portray controllers

using techniques that appear t o have been shortcuts or abbreviations of the full coordination pro-
cedure. Usually such reports cited factors such as workload or inadequate staffing in conjunction
with these events. The reports indicate that the possibility of controller error is greater when short- ’

cuts are used in situations where full coordination would exercise more positive control of the air-
craft involved - especially when the abbreviated technique is not specifically authorized by direc-
tives providing a common understanding t o all parties.

The techniques discussed are point-outs, quick-looking adjacent sectors, use of visual signals in
intrafacility coordination situations, and approvals subject to other traffic. None of these proce-
dures is illegal; they are not even marked departures from routine operating practices in many kinds
of separation situations. However, in the specific kinds of situations covered here it is simply the
case that - as events turned out - the use of normal coordination would almost certainly have been
more successful in avoiding loss of separation.

30
 Many reports indicate that coordination, such as it was, consisted of a point-out only or a
statement t o “miss A” o r a similar abbreviated message. The reporter quoted below states succinctly
the problem associated with this technique error.

Point-out procedure was used. However, 1 d o not believe good judgment was used -
should have been handed off. Point-outs are taking the place of handoffs resulting
in more system errors.

The following two reports depict results of technique substitutions in situations where normal
coordination might have prevented serious conflicts.

Two aircraft were at the same altitude when military A reported a near midair. He
said he needed to climb t o miss GA B. Poor coordination was the cause of the prob-
lem. Some coordination is done by pointing out traffic to another controller. The
other controller could be talking t o someone on land-line giving approval which is
interpreted as approval for the point-out.
* * *
Aircraft A was landing in my sector. After accepting handoff from the adjacent
Approach Control, I asked my data man t o request control. He advised I had control
but t o miss aircraft B over the VOR southbound. I delayed turning A until I was
sure they would be clear of one another. I turned A t o heading 340 and observed
traffic at eleven o’clock 2 miles southbound readout 3,000 ft. Aircraft A advised he
had the traffic and the type was that of aircraft B.

Most of the point-out reports were similar t o the foregoing. Acceptance of a point-out meant
that relevant traffic was observed but not under coordinated control and, in many instances, its
intentions were unknown. In situations where coordination is needed, the point-out technique can
work reliably only where there is a framework of facility directives that defines coordinated opera-
tions utilizing pgint-outs.

Other reports of coordination failures caused by technique errors contained the characteristic
line “approved - traffic is A, B, and C.” The almost inevitable result of this procedure - condi-
tional approvals subject t o separation from other traffic - where no attempt is made to construct
an agreed upon course of action, is that one controller handles traffic in another controller’s
assigned airspace. The following is an extreme example of this kind of situation that was further
complicated by a position relief at a critical moment.

Controller 1 takes Center handoff on aircraft B descending to 7,000. Controller 2

radar-vectored SE bound departure A t o 6,000 and made handoff t o Center. Center
issues climb t o 11,000 with respect to B. Controller 2 clears A t o 11,000 report
leaving 8,000. B calls descending t o 7,000 leaving 15,000. Controller 2 descends B t o
3,000. Relief controller 3 assumes position. Aircraft A reports leaving 9,000. Con-
troller 3 switches A t o Center frequency. Controller 3 checks strip showing 70/30
with no indication that clearance is to 7,000. He now learns that B is descending and
still above 7,000. Controller 3 asked B his altitude, reply was 13,800. At this time
A and B were less than 5 miles apart converging. Controller 3 called Center t o

39
 coordinate altitudes for both aircraft. Center advised they would stop A at 11,000
and to stop B at 12,000. Pilot of B advised he was already of that altitude. Targets
merged.

In trafacility coordination is sometimes accomplished by means of visual signals between con-

trollers. Although no official recognition exists for such procedures, they may be used during
periods of high activity. Problems related t o the accurate understanding of such signals are similar to
some of those encountered in voice communications. Since there are no standardized visual signals,
they are the product of improvisation by the controllers. The meaning of the signals may be appar-
ent when the signal is simply to climb or to descend, but efforts t o convey detailed data, such as
heading or altitude assignments or to provide more than approval of a request as perceived by the
controller being queried, can lead to misunderstanding. Misinterpretation of the intended message
results in actions leading t o a possible hazardous incident.

Another substitute for verbal coordination is the look-and-go concept used in some facilities.
This is the reverse of the point-out case. A controller quick-looks the airspace being controlled by
another, and if he observes n o traffic pertinent t o his plan, he clears the aircraft he is controlling
into the adjacent airspace. This procedure can work effectively provided all pertinent data are avail-
able on the display. Both incidents and serious accidents have resulted because either the displayed
information was incomplete or because the controller failed to observe and perceive correctly the
traffic situation in airspace not under his direct control. This ASRS report describes such an
incident.

Departure control was being operated by a controller who uses “look-and-go” pro-
cedures. As A asked to start descent I approved this and noticed a target with an
‘altitude readout.of 3,500 a t twelve o’clock, 1 mile. A military departure was climb-
ing alongside and through the altitude of my aircraft. This incident could have been
prevented by Departure Control inquiring if I had any aircraft in that area and ask-
ing permission t o enter my airspace rather than the “look-and-go” operation that is
used by some people.

Two accidents exemplify the pitfalls of using the “look-and-go” technique. The first, described
in a National Transportation Safety Board accident report (ref. I), involved a landing cargo wide-
body and a taxiing passenger air carrier that nearly collided after the air carrier had been cleared t o
cross an active runway, The wide-body suffered major damage when it departed the runway t o
avoid collision.

In this case a facility directive authorized the ground controller t o exercise his judgment with-
out conducting coordination with the local controller, provided he observed his radar displays of
the BRITE and the Airport Surface Detection Equipment (ASDE) for appropriate traffic in forma-
tion. The report indicates that the spacing between the two amvals was only 2 miles and that the
standard was 4 miles. The ground controller apparently failed to perceive the radar target of the
wide-body (the second arrival) and assumed that the next arriving aircraft was one displaying a full
data block still well out. Since the local controller did not advise the ground controller of the less-
than-standard spacing, there was a diminished opportunity for the error to be detected. This is a
special case of look-and-go, done in full accord with directives then in force, but it still illustrates
strikingly the lack of redundancy that exists when the look-and-go concept is in use instead of full-
scale coordination.

I 40
 The second accident (ref. 2) demonstrates how primary radar targets present a particular pit-
fall in use of the look-and-go concept. Display of primary radar data may be intermittent or may
provide only weak returns at the time of the quick look. Local procedures normally require that
coordination be effected when an aircraft that is not equipped with a transponder is involved. If
coordination is not accomplished, or if a controller fails t o recall after coordination, an incident or
an accident may result. In this case, a midair collision occurred in which six lives were lost. A break-
down in coordination was involved when look-and-go procedures were in use and one of the aircraft
was not equipped with a transponder. The following exerpts were contained in the report.

While the facility operating procedures were adequate, it was imperative that spe-
cific procedures be adhered to.

When the coordination procedures broke down, there was no redundancy t o provide
an additional safeguard for aircraft.

Visual coordination was an inadequate form of coordination .... In reality, no

coordination took place.

Perreptioil errors- The second major class of coordination errors consisted of failures of con-
trol!er perception. In most cases, the controller failed to observe a potential conflict early enough t o
take appropriate action. In other instances, the implementation of the coordination plan resulted in
conflicts with traffic that was not a part of the coordination agreement. Less frequently, the flight-
path projection of the coordinated traffic was improperly perceived.

I accepted a handoff on B believing he would be stopped at 2,000. A minute later

B reported level at 3,000 only 1 t o 1-1/2 miles from A .

Only rarely could it be inferred from the context of the reports that controllers made these
perceptual errors because they were incompetent to read or interpret the information coming to
them from displays and from radio communications. The most often cited or implied precondition-
ing factor was distraction of various kinds. Controller attention was most often drawn from coor-
dination tasks by the need to respond t o incoming communications from other aircraft or from con-
trollers not involved in the coordination action. This may reflect a possibility that controllers are
conditioned by training and experience to treat such responses as first priority. Another source of
distraction was a variety of events - all quite routine - occurring within the ATC facility: relieving
of positions, decombining positions, supervisory intervention intended to provide assistance, and
(often mentioned) the provision of the on-the-job training.

I was conducting a check ride on a controller. Traffic was moderate to heavy. At a

critical time my attention was diverted by another controller who was very insistent
on talking t o me about a situation earlier in the shift. During this time I felt certain
that the controller had coordinated t o go through another controller’s airspace. He
in fact had not.

41
I The Management Role in Coordination

The crucial role played by ATC management in the coordination process manifested itself in
many of the coordination failure reports. The procedures t o be used are defined in facility direc-
tives and in letters of agreement that management prepares. These are, of course, compromises
because they must best serve all the users yet be compatible with aviation system capabilities. In
developing such procedures, management subdivides and delegates controlled airspace in such a way
that the need for control transfer is minimized, thereby easing the coordination workload as much
as posbible. Finally, management controls, t o a large extent, the immediate working environment in
each facility and so has much t o d o with determining the amount of distraction the controllers must
I - deal with. Thus management’s role is directly involved with all the sources of error noted in the
I coordination failure reports.

Airspace subdivision- Airspace configuration is tailored to provide a smooth flow of traffic

with minimal coordination. One design consideration is that controller workload may be less with a
large segment of airspace containing a relatively large number of aircraft to be worked than with a
smaller segment that requires additional coordination. Each subdivision of airspace creates a require-
ment for additional coordination which may result in an overall increase in workload. The distrac-
tion effect of the added coordination requirement may increase the probability of a hazardous
incident.

Airspace above our sector belongs to another Approach Control from 8,000 to
I
15,000. B appeared t o be in level flight at 7,000 when he should have been at 6,000.
B then continued on into the adjacent approach control airspace. This incident
resulted from B’s penetration of my airspace at the wrong altitude for direction
without my knowledge. My sector only has 7,000. It is extremely complex and there
is no room for error. Too much shelving, too little airspace to maneuver.

Establishment of coordination procedures- ATC management establishes all coordination

procedures. Higher management, regional and Washington headquarters, usually direct general
policy, which is implemented by local directives that prescribe the day-to-day procedures for con-
trollers. For example, agency policy provides for mandatory radar handoffs between controllers if
radar coverage exists. The time, location, and to whom a handoff will be made will be established
by facility management in the form of facility directives, or by interfacility management in the
form of a letter of agreement if more than one facility is involved.

Occasionally ASRS reports indicate that procedures were designed without taking into account
some contingency of traffic configuration. Such reports are concerned most often with allegedly
flawed letters of agreement. Although they sometimes reveal the existence of interfacility prejudices
held by controllers, they also can indicate real hazards in which a faulty procedure creates condi-
I tions that lead to perceptual or technique errors.

Distraction-free environment- It is clearly a responsibility of ATC management to provide

working environments that are as free of distractions as possible. The topic has been discussed and
exemplified earlier in this report. That distraction is an important precursor of human error in the
cockpit was shown by Monan (ref. 3); it is obviously an important factor in air traffic control as
, well.
 Reduction of Coordination Errors

The coordination errors that seem t o be most readily correctable by design or procedural
changes are those that lead t o a controller’s failure t o act in a timely manner, either because he
requires an additional stimulus t o act at the right time or better memory aids than are now available
t o enable him t o keep abreast of a complex and changing traffic situation. There appear t o be two
feasible approaches t o this: more extensive use of existing automation capabilities and further
development of controller memory aids.

Automatioil capabilities- If controllers who are busy with routine matters were to receive
reminders about a coxdination action when time is becoming critical, their failure rates might be
substantially reduced. The present automated systems already have well-developed capability for
doing this but it appears from ASRS reports that some facilities are not equipped or are not using
available capabilities t o the full. The following is a problem that arose at the Boston Center when a
manual handoff was used.

ARTCC initiated a radar handoff after aircraft A was 1 mile inside approach control
airspace. Conclusion - the ARTCC radar controller failed t o handoff A prior to the
aircraft entering rapcon delegated airspace.

Automatic handoff, automatic acquisition, and automatic conflict alert provide reminders to
controllers at different points in the coordination process. The procedures are available, and their
full utilization at all ATC facilities would directly reduce the probability of coordination failures.

The use of automatic hatidoffs by en route facilities would alert controllers when controlled
aircraft are about t o penetrate their airspace. The alert signal is so timed that there is usually suffi-
cient time for remedial action if the handoff controller has not perceived the need for initiating
coordination. Terminal facilities benefit from the use of the automatic handoff by Centers, even
though they may not have the capability themselves. However, routine acceptance of handoffs by
receiving controllers without analysis of depicted data will defeat the effectiveness of the alert
provided.

Auromatic acquisition allows the system to establish a track and display a data block on air-
craft entering radar coverage in accordance with desired parameters. All automated facilities have
this capability. An alert can be provided to the controller when a handoff cannot be made or is
overlooked by the transferring controller. This feature is useful when aircraft depart an area with n o
radar coverage and later enter one with radar coverage. It also allows en route controllers t o see
departing aircraft prior t o exiting terminal airspace, provided they are within en route radar
coverage,

There is general agreement that the conflict alert system has been effective in reducing system
errors, although its primary purpose was t o prevent midair collisions. If a controller receives a 2-min
warning of impending loss of separation, he can (in most cases),maintain separation by prompt
response to the alert. In some cases, no advance warning can be provided. For example, a change of
course can result in an immediate loss of separation which could not have been predicted. In those
cases, the loss of separation has occurred when the conflict alert is issued.

43
 Since ATC radar data is analyzed on a continuous basis, it would be possible t o provide an
earlier alert t o the controller of a potential conflict. In most cases, verbal coordination is initiated
in order to resolve such a conflict. An earlier signal that a conflict was likely to occur might
decrease the number of cases in which action t o coordinate was not initiated, the most common
failure in this sample. False alerts could be minimized by providing discrete sector parameters set
equal to the conflict alert parameter, if desired, or t o any appropriate time value.

Neither this proposal nor the conflict alert system is infallible; the function is only as good as
the data available. Use of Mode€ altitude data as the primary source of altitude information for
conflict prediction would result in improvement and would reduce the likelihood of the system
being defeated by human error.

Memory aids- Many controllers use scratch pads or grease pencil notations to help them
remember important data. Flight-strip notations can be effective reminders of further actions that
may be required. If used more extensively, these and similar devices might help to reduce the num-
ber of recall failures. Because total reliance on recall appears t o be a predisposing condition to
failure, as heretofore discussed, any measure that can assist in minimizing this factor would be
useful.

Operational technique- The need for providing positive control t o the maximum extent in
the procedures of the controllers would seem t o be self-evident. Reliance o n radar separation as the
only tool available is suggested in several ASRS reports. An example is use of vectoring when alti-
tude separation can and should be provided. In airspace designated “positive control area,” con-
trollers often d o not exercise positive control. Unrestricted climb or descent clearances are issued
through other traffic, with the controller anticipating that radar separation can be applied as a last
resort. An impending conflict is observed and there is a hurried effort t o coordinate a change. Some-
times, the time expires before anything can be accomplished, as shown by several examples in this
study report.

Letters of agreement and facility directives often provide for such procedures as route separa-
tion of transitioning aircraft and standard instrument departure routes (SIDS), but the controller
can avoid many worrisome incidents by adopting the position that he will provide positive control
t o the maximum extent achievable, avoiding “betting o n the come,” by utilizing all of the tools
available in the most effective and professional manner.

CONCLUSIONS

Coordination failures occur throughout the air traffic control system. The most frequently
reported failure is failure to initiate action t o establish coordination; the second most frequently
reported is failure to execute the agreed plan. The causes of failure appear to be similar in both
types of errors. The remaining failure modes are reported less frequently; they consist of misunder-
standing the transferred data or the plan agreed upon, or the data used t o develop the plan are
in corre ct .

Perceptual and technique errors appear throughout reports of coordination failure. Memory
aids and reminders can be used to reduce the likelihood of oversight or failure to execute a

44
previously developed plan. Maximum utilization of automation capabilities, where they are avail-
able, could reduce the need for oral coordination and, consequently, reduce the probability of
error.

Errors in coordination often result in hazardous incidents, and they continue to be reported to
ASRS in substantial numbers.

45
 I REFERENCES

I 1 . Aircraft Accident Report AAR-79-11. National Transportation Safety Board, 1979.

I 2. Aircraft Accident Report AAR-78-14. Mational Transportation Safety Board, 1978.

~
3. Monan, W. P.: Distraction - A Human Factor in Air Carrier Hazard Events. ASRS Ninth Quarterly Report,
~ NASA TM-78408,1979.
 N8 1 - 3I16 6
4. INFORMATION TRANSFER BETWEEN AIR TRAFFIC CONTROL AND AIRCRAFT:

COMMUNICATION PROBLEMS IN FLIGHT OPERATIONS

Ralph L. Grayson and Charles E. Billings

INTRODUCTION

This study of problems in communications between flightcrew and air traffic controllers was
prepared as a part of an analysis of information transfer problems in the national aviation system.
It is adapted from reference 1 , an earlier study (by the senior author) of possible effects of the Dis-
crete Address Beacon System (DABS) Data Link on the information transfer problems reported to
the NASA Aviation Safety Reporting System (ASRS); the work reported in reference 1 was per-
formed at the request of Systems Research and Development Service, FAA.

The purpose of this report is t o discuss problems in oral communication between pilots and
controllers. The investigation consisted of review and analysis of pertinent information in the ASRS
data base.

APPROACH

A search was conducted of 6,527 reports submitted to ASRS between May 1, 1978 and
August 3 1 1979. The search technique identified and retrieved reports Concerning:
~

1 . Problems in voice communications between flightcrew and ATC

2 . Problems in conveying infomation in ATIS broadcasts (as specified in the AIM and FAA
ATC Handbook 7 1 10.65A)

3. Problems with information concerning wind shear and minimum safe altitude

In this group of reports, 5,402 information transfer problems fitting one of these criteria were
identified (some reports contained more than one information transfer citation).

The research team studied selected report narratives t o establish the generic types of communi-
cations problems present. A categorization of these types was developed along lines already in use
in classifying ASRS reports. Ten such categories emerged:

1. Misinterpretable - phonetic similarity

2 . Inaccurate - transposition

3 . Other inaccuracies in content

47
 4. Incomplete content

5 . Ambiguous phraseology

6 . Untimely transmission

7. Garbled phraseology

8. Absent - not sent

9. Absent - equipment failure

10. Recipient not monitoring

The retrieved reports were grouped according t o these categories. The groupings were then
further subdivided according to the operational regimes in which the reported incidents occurred
(i.e., terminal operations, en route operations, and various special operations). Selections of these
reports were reviewed t o determine unique characteristics and common features.

RESULTS AND DISCUSSION

The initial pass over the data base, the first step of the search, produced a finding considered
significant t o the central issue of this study: 70% of the reports t o ASRS involve some type of ora1
communication problem related t o the operation of an aircraft. The nature of the problems
reported vaned widely, ranging from failure to originate an appropriate message to failure of the
in tended receiver t o comprehend and confirm the message accurately.

These communications problems were subdivided into the 10 generic types listed previously.
Before taking up these assessments, however, two aspects of communication difficulties require
consideration: the expectation factor and the problem of conveying traffic avoidance information
effective1y .

The Expectation Factor

ASRS reports indicate that many instances of misunderstanding can be attributed t o the
expectation factor; that is, the recipient (or listener) perceives that he heard what he expected t o
hear in the message transmitted. Pilots and controllers alike tend t o hear what they expect t o hear.
Deviations from routine are not noted and the read-back is heard as the transmitted message,
whether correct or incorrect.

Aircraft A was in a block altitude of 12,000-14,000 ft. The instructor pilot and the
student both thought the controller told them to turn left t o a heading of 010" and
descend to and maintain 10,000 ft. At 10,700 ft the controller requested aircraft A's
altitude. The crew responded 10,700 ft. The controller stated the aircraft had been
cleared to 12,000 ft, not 10,000 ft. There are two contributing causes for this

48
 occurrence: 99% of all clearances from that area are t o descend to and maintain
10,000 ft, and as the instructor I was conditioned to descend to 10,000,by many
previous flights. The controller may have said 12,000 ft but I was programmed for
10,000 ft.

Conveying Traffic Avoidance Information Effectively

ASRS reports suggest that the least satisfactory aspect of air/ground information transfer is
the conveyance of traffic advisories and avoidance information. Faults of all kinds are cited, but the
pervasive difficulty that appears to underlie many of these faults is the seeming inconsistency with
which information about traffic is made available.

While descending through 12,200 MSL first officer observed and called traffic at
twelve o’clock level as we were turning through 300”. Turn was continued to
approximately 320” and other aircraft diverged to the left on a southeast heading
with clearance of approximately 1,000 ft laterally. On inquiry, ATC indicated that
the only altitude readout on a target in that area was 6,700 MSL. If our aircraft had
not been turning in on heading approaching VOR, a projected collision course would
have resulted. Situation discussed with ATC supervisor who indicated that a “skin
paint” was later picked up on other aircraft but later lost by adjacent center. Other
aircraft apparently operating without transponder would be primarily cause of this
incident. Contributing would be difficulty in picking up front profile visually at such
closing speeds. Other aircraft made no evasive action and we assume he did not
observe us.

In the present system, air traffic controllers provide traffic advisories as an “additional ser-
vice,” which means that workload permitting, the controller will issue advisories on traffic that he
observes when he is not occupied with higher priority duties. This results in pilots failing to receive
traffic advisories on aircraft that are not readily seen on radar - especially those that are not
transponder-equipped. In addition, it is during periods of high traffic that the workload of the con-
troller is likely t o preclude issuing traffic advisories - precisely when the need is the greatest.

Generic Types of Problems in Oral Communications

Thousands of ASRS reports cite the difficulties in the exchange of information through the
use of oral communications. Some reports concern transfer of information between ground facilities
o r personnel within such facilities. The greater number of reports concern air/ground communica-
tions and a very small number concern air communications alone.

Air/ground communications are conducted by voice radio as they have been for about
50 years. During that time technical advances have improved the quality of voice transmissions and
mitigated atmospheric or induced electronic interference. Remaining technical problems include
blocked transmissions, line-of-sight limitations, and hardware failures that remain undiscovered
until the next occasion for a communication arises. However, the retrieved ASRS reports concern-
ing problems in air/ground communications indicate convincingly that most of such communica-
tions problems involve. human error.

49
 Misinterpretable - phonetic sinzilarity- The “phonetic similarity” category was assigned when
similar-sounding names or numerics appeared t o lead t o confusion either in meaning or in identity
of the intended recipient, thus causing an information transfer failure. A total of 71 reports were
classified in this category. The following narrative is typical.

We were cleared into position on runway 32L for an intersection takeoff. After a
brief hold in position we received what I thought t o be a takeoff clearance. I then
repeated “Roger, ACR 122 cleared for takeoff, straight out departure.” There was
no response from the tower until we were well down the runway approaching V-1
speed. The tower controller then said rapidly, “ACR 122 that clearance was not for
Y O L I , it was for ACR 142.” We heard no other trip respond t o the takeoff clearance
but possibly we responded at the same time as ACR 142 so that tower was unaware
that we had both answered and blocked each other’s response.

Most reported phonetic similarity problems involved execution of clearances by someone other
than the intended receiver.

Iiiaccurate - transposition- In the group of ASRS reports reviewed, there were 85 in which
some part of the message was misunderstood because of a transmitted or recipient-perceived error in
the sequence of numerals within the message. This type of error seems t o occur most often when
ATC gives assigned headings or distances in conjunction with changes in assigned altitudes in the
same clearance. Heading 270 might be heard as a new assigned altitude. The readback then might
not be perceived as incorrect (expectation factor) and an incident might result. One ASRS report
illustrates this problem.

F/O flying, Captain working radio, Center gave clearance t o descend, (either) (1) to
cross 10 DME east at 240 or (2) to cross 24 DME east at 10, F/O set 10,000 ft alti-
tude and 24 DME,and started descent. Leaving 19,200 Center advised we should be
at 240. Captain advised we show 10 at 24 DME, but what altitude did he want at
this time, he then said maintain 180.

Other inaccuracies in content- Other reports cited inaccuracies for reasons other than
phonetic similarity or transposition. There were 792 indicating a message problem of this category.
Generally, they involved messages that were accurately transmitted and received, but they con-
tained, or were based on, erroneous data (formulation errors), or, to a greater degree, they were the
results of errors of judgment in the originator’s decision process. This resulted in the relatively large
number of reports in this category.

Faster aircraft B was overtaking aircraft A so I issued headings that would provide
lateral separation. Later aircraft A requested deviation around weather that I did not
observe on radar. Thinking that a route direct t o XYZ would maintain lateral separa-
tion and provide A with necessary deviation, I issued the clearance. The clearance
brought A back south and since I only had 5 miles in the first place, I immediately
lost separation.

Other reports in this category reflect conflicts in the interpretation of a message between the
sender and receiver where there is n o obvious explanation for the difference in understanding.

50
 Lift off runway 31 climbing t o 5,000 per SID. On initial contact flight was cleared
t o 12,000. Subsequent transmission received and acknowledged to climb to 14,000
and maintain speed less than 250 knots until 13,000 or above. Traffic was observed
at one o’clock on converging course descending. When our flight left 13,000 ft
Departure Control asked our altitude and advised us t o descend to 12,000 and
increase speed, No member of the crew either heard or acknowledged such a
message.

Incomplete content- A reported problem communication was classified as “incomplete” when

the originator failed t o provide all of the information necessary for the recipient to understand it
properly. There were 296 reports classified in this category in the study group.

Between LIT and FAM we were cleared for a Farmington tiansition t o 30 left. T o
the best of both pilots’ recollection, n o statement was made by the controller t o
“expect a profile descent,” when the clearance for a Farmington transition was
given. A flight was in the Farmington area climbing to FL230. Upon hearing aircraft
B talking with Center, we volunteered our altitude as being FL240 and we leveled
off. I was watching B at FL230 and no evasive action was required.

In this example the requirement for profile descent was not effectively transferred, whether
because of input error, failure t o comprehend, or a failure of the voice radio system. These failure
causes are characteristic of the reports in the “incomplete content” category.

A iiibigzrorts phraseology- A reported problem communication was classified as ambiguous or

misleading if the composition, phraseology, o r presentation of the message were such that the
recipient would tend to misinterpret or misunderstand the message when receiving it under normal
conditions. There were 529 reports in which this kind of message problem was cited.

We were being vectored downwind when the controller said to plan on a visual
approach to runway 28. At this time we were at 6,000 t o stay above,departure
traffic. We were assigned heading of 100 and cleared to 4,000. At this point we were
south of runway 28 abeam the airport. Controller said, “The runway is nine o’clock
and 3 miles, can you see the runway? We responded yes. He said, OK, turn t o 360”.
At this point we started our turn and (thinking we were cleared €or a visual
approach) began a descent. He asked our altitude at 3,400. Then he said he had not
yet cleared us below 4,000 but t o stay where we were. Shortly thereafter, he then
cleared us for a visual and changed us t o the tower.

Untimely transmission- Messages were classified as untimely if they originated too late o r too
early t o be useful t o the recipient. There were 7 10 reports that indicated this message problem.

Departure clearance was left turn after takeoff to 120”, climb and maintain 7,000.
We had just cleaned up and finished the climb checklist and at about 4,500 f t
Departure Control gave us VFR traffic at twelve o’clock less than a mile. The
Captain spotted the traffic and pointed it out to the F/O who was flying and nosed
the aircraft over into level flight t o go under aircraft B 50 to 100 ft and slightly
behind him about 100-200 ft. Aircraft B saw us just before we passed under and
behind him - he flinched just enough to slightly raise his left wing. We feel that

51
 radar should have had aircraft B in radar contact at the time we took off and we
should have been advised of the traffic at or before takeoff.

Garbled phraseology- Messages were coded as being garbled if information content was lost or
severely distorted so that the recipient was unable t o understand the intended message. There were
17 1 such reports in the study report group.

Departed on runway 27 with a right turn to 300”. After takeoff the heading was
amended t o 330’ but the transmission did not come through clear t o us and it was
mistaken for 030”. Subsequently we learned that our read-back to the controller was
not received clearly and it was assumed that we had received 330” instead of what
we interpreted t o be 030”. Obviously, too much assumption, probably assisted by
the unusually clear weather. We later learned that our error had brought us in con-
flict with aircraft B that had taken off immediately in front of us. Radar had us
less than a mile from aircraft B when we passed.

Absent - not sent- Problem communications were assigned t o the “absent - not sent” cate-
gory when there was a failure to originate o r transmit a required or appropriate message. In the
study sample, 1,991 reports were classified in this subset. The large number is due to a broad inter-
I pretation that an appropriate message would have broken the chain of events that resulted in a
hazardous occurrence. This could consist of either a point of information or an air traffic control
clearance,

Runway 9 R in use - (heading 120 and told t o expect a nevi heading when in the
air). The aircraft ahead of me was issued right turn t o 240 o r 270. I was left on 120.
This heading aimed me toward aircraft B and I felt very uncomfortable. When Tower
did not give me an immediate turn, I contacted departure radar expecting the other
turn. After radar contact was established, the departure man asked me t o go back t o
Tower and upon returning he (Tower) told me I should have expected the second
turn from him. If Tower had issued “expect further clearance from him,” it would
have made this clear and concise.

Absent - equipment failure- Problem communications were categorized as “absent - equip-

ment failure” when a failure caused a complete loss of the message. The study report group con-
tained 153 of these reports. This finding suggests that the equipment in use is reliable when com-
pared to the human error-initiated problems in message transfer that are reported to ASRS.

Aircraft A and aircraft B were being vectored t o Detroit Metro Airport by Approach
Control. Aircraft C was being vectored to Willow Run airport on the same fre-
quency. The microphone button became stuck on C. As a result the approach con-
troller was unable to communicate with A or B and less than standard separation
occurred. The two aircraft were within approximately 500 ft vertically, the pilot of
B called the tower controller and advised there was a stuck mike on Approach Con-
trol. The tower controller, using the information on his radar display, attempted to
descend and turn B t o avoid the conflict. However, the situation had deteriorated t o
the point that the conflict could not be avoided. Aircraft B apparently took evasive
action.

52
 Loss of communication is an extremely frustrating experience for an air traffic controller. He
is usually helpless t o take action to preclude a hazardous consequence. In the above example, one
pilot had the presence of mind to contact the tower after noting the approach control frequency
was blocked due t o the stuck mike, but it was too late t o avoid loss of separation.

Recipient not monitoring- A problem communication was placed in the “recipient not moni-
toring” category if the recipient failed to maintain listening watch, proper lookout, or failed t o read
available correct information. There were 5 53 reports in this classification.

A conflict occurred between aircraft A on my frequency and aircraft B on Approach

Control frequency. Aircraft A was departing and B was arriving. I was not aware that
the B flight was in the area and the conflict was first noted by an approach control-
ler who must have seen what happened. An investigation followed showing that the
approach controller who had handed off A to me failed t o coordinate all aspects of
control t o me as per letter of agreement between our facilities, and I failed t o catch
the mistake when A came on my frequency. Poor coordination between controllers
and not properly listening to the pilot’s initial contact were contributing factors in
this incident.

A substantial number of reports in this category described the results of traffic advisories not
being issued when the reporter alleged or inferred that the traffic could and should have been seen
on radar. It may be technically correct that a specific target was not seen because of inattention t o
the particular area in question, and, therefore, the area was not being monitored by the radar con-
troller. It appears, however, that many pilots expect traffic advisories at all times if they have been
advised that they are in radar contact.

Another case is the failure of the flightcrew or the controller t o receive a message or initial
broadcast or t o respond to a call when time is critical. In some reported cases the controller became
aware that his original plan for providing separation was not working, and he then attempted t o
correct the situation at the last moment. Lack of instantaneous response gave rise t o the allegation
that a proper listening watch was not being maintained.

Communications Problems Related t o ATIS

The data base was searched for reports that concerned (1) items of information contained in
the present Automated Terminal Information System (ATIS), (2) broadcast problems with ATIS,
and (3) indications of both a communication message problem and a t least one item of ATIS infor-
mation. In addition, reports concerning wind shear and minimum safe altitude warning were iden-
tified. (ATIS items of information included in these broadcasts are specified in the Airman’s Infor-
mation Manual and the FAA Air Traffic Control Handbook 7 110.65A.)

An interesting finding of this search was that relatively few of the message problem reports
were concerned with terminal information services; only 50 such reports were retrieved. It had been
expected that there would be a large number of reports of difficulty in understanding the ATIS
broadcasts. A considerable number of such reports were received in early ASRS operations (1 976
and early 1977), but they decreased to a smaller number during the search period of this study. This
suggests that improvements have been made in response to several FAA directives aimed at poorly

53
 prepared ATIS tapes, use of excessively rapid rates of speech, and technical problems with ATIS
broadcast equipment. It may also suggest that difficulty in understanding ATIS is judged by airmen
t o be a minor matter, one that is easily overcome by repeating the broadcast and therefore not
worthy of an ASRS report. In any case, most terminal-information-related reports described prob-
lems with ATIS that had substantially more serious consequences than having t o listen to a broad-
cast a second o r third time.

The problems present in the terminal information segment of the retrieved reports were classi-
fied as follows:

I 1. Unintelligible transmissions

2. Obsolete approach/runway-in-use information

3 . Noncurrent runway visibility readings

4. Obsolete weather information

I Unintelligible transmission- Reporters have cited difficulty in receiving ATIS broadcasts

because of the rate of speech and the quality of the recording.

On a VFR flight to ICT late in October, I had to listen t o the ATIS seven times to
get active runway, wind, altimeter setting because of the rate the words were
spoken - too fast. As a low-time pilot my workload when landing at an unfamiliar
airport is higher than it should be; I check and double check everything and it is
unsettling t o be unable t o get the information needed. Most places I’ve been that use
ATIS - Chicago, Milwaukee, Madison, Rockford - it seems that making the tape
has become a chore so it reads as fast as possible t o get it over.
* * *
ATIS is supposed to speed up and facilitate arrivals and departures at airports large
enough to warrant its installation. This is a wonderful concept. However, on the
times that I have been into airports that have it (Boise and Portland) the report has
been so distorted as to be all but useless. There is no need for the person recording
the information to speak as fast as he can. I d o not believe that I should have t o
listen to ATIS more than three times at the most t o have all the information
straight. Once should be sufficient. But I have had to listen 5 minutes o r more
before I was able t o clearly understand what it is the man is actually saying. Without
exception, I ,have not been able to clearly understand the content of the ATIS
broadcast the first time simply because the man spoke too fast and, for the lack of a
better word, mumbled as he talked.

Obsolete ATIS dataspproachlrunway in use- A frequent complaint from reporters is a change

of approach o r active runway from the information contained in the ATIS broadcast received by
the reporter.
 Planned approach and landing for runway 3 1 R JFK. Prior t o turning final, runway
was changed t o 22L and aircraft was vectored for ILS or visual approach t o run-
way 22L. Visibility approximately 5 mi smoke and haze. Our aircraft was advised
being vectored for 22L ILS approach t o advise when runway in sight. Throughout
the vector until turning on final we received the 13L ILS identification (IMOH).
Both 22L and 3 I R share a common frequency 110.9. We reported this t o Approach
Control and were advised the ILS was operating normally on 22L. Not until we were
established on visual approach for 22L did the ILS start operating normally with
current identification. At this time Tower advised that 22L was now operating nor-
mally. Not serious in VFR but could be very confusing and possibly cause missed
approach in IFR.

Ru,zway visual range- Rapidly changing runway visual range (RVR) results in both frequency
congestion and cockpit distraction at the most critical time in the execution of an instrument
approach. Since RVR is not transmitted unless the approach is being conducted in near-minimum
weather conditions, it is a critical distraction in which voice transmissions are used in the present
system. In other cases the RVR appears to have been omitted or the reading was not accurate when
it could have been very valuable t o the pilot on approach.

Flight making ILS approach crossing outer marker, Tower reported heavy rain at
airport. Speed and rate along with localizer and glide slope all were normal through-
out approach. Sighted approach lights at 400 ft and began encountering light rain at
300 ft. Runway was in sight and just at touchdown encountered a wall of hard rain
and had no forward visibility. I could see by the center line that we were going off
the left side of the runway. We soon felt our left main gear was in the lights or
possibly off t o the left side of the runway. We continued forward velocity for about
1,000 ft when we again regained forward visibility at which time the captain was
able t o bring the aircraft back over the runway and bring it to a stop.

Obsolete weather informatioil- Instances were reported when ATIS transmitted obsolete
weather information. (These problems are similar t o those reported above.)

Approach was made VFR - on short final encountered rain (which we thought was
light because Tower had not reported any). Rain was heavier than anticipated.
Normal touchdown - wind from left which was not reported blew us from runway
because of hydroplaning. Aircraft came t o a stop just off side of runway 7R.

In another report the wind-shear factor in addition to the obsolescence of the ATIS informa-
tion proved to be a problem.

Several aircraft reported to Tower that there was moderate to severe turbulence on
final approach. Flightcrew monitored Approach Control and Departure Control fre-
quencies while waiting for takeoff. To my knowledge arriving aircraft were not
advised of “wind shear”/turbulence/airspeed excursions. After our takeoff at XX55
I checked arrival ATIS - no mention of approach difficulties - in fact information
was 50 min old. Despite reported hook cloud classically displayed on radar to the
SW, 1 advised local operations of wind problems reported on final, suggesting they

55
 advise pilots in range and that MIA dispatch also be advised - that evening about
six tornadoes hit central Georgia and there was extensive tornado damage to Forest
Park. Arriving A/C seemed t o be left out of the information loop.

Terminal Operations

Examination of communication problems in terminal operations concerned aircraft operating

under ATC control while moving on the surface at the airport, flying within the airport traffic area,
or arriving and departing the terminal area. These problems are discussed below.

Communications at airports without towers were considered t o be advisory only and no

detailed study was made or considered appropriate to this report.

Surface operations- In connection with surface operations, ASRS reports evidence two main
types of communications problems: clearance misinterpretations leading to active runway incur-
sions and failures t o communicate taxi routes to preclude wrong turns and consequent ground
conflicts.

ATIS was received advising departures on runway 4, arrivals t o runway 3 1. Cleared

t o runway 4. We switched to Tower and advised we were ready for takeoff. Tower
said, “Taxi up to but hold short,” which first officer acknowledged, but I thought
Tower said, landing was on runway 31. As I took the runway, I looked t o my left
and noticed an aircraft on about a 3-mile final. The error was caught by me, the
tower, and the first officer at the same instant. Tower advised I was supposed t o
hold short. I immediately cleared the runway well before the traffic landed. Factors
which I think contributed were Tower deviating from the ATIS information and
then not specifically advising us. I have become accustomed t o holding in position
while an aircraft lands on another runway, and we are creatures of habit and I
thought I heard something I obviously didn’t.

Many runway incursion problems appear t o result because a flightcrew acts on a clearance,
onto the runway or for takeoff, intended for another aircraft. This occurs most often because of
phonetic similarity of call signs o r crew predisposition - expectancy.

The taxi problem is most often related to flightcrew unfamiliarity with airport layout, repairs,
and changes; communications problems tend t o be secondary.

Flight operations - airport traffic area- The airport traffic area is the scene of many reported
system irregularities. Prominent among these are traffic conflicts with unknown aircraft or with
aircraft that are not properly in the pattern or on proper approach/departure paths; traffic conflicts
due to sequencing disorders; use of wrong runways; and deviations from intended aircraft trajectory
(course, speed, altitude) during approach or departure. The role of communications problems in
enabling these events is highly vaned but is important in each class.

Conflicts with unknown aircraft: ASRS reports describe conflicts of this type often occurring
when controllers are unaware of the traffic o r are too busy t o issue advisories. These situations fall
under the general headings of absent or untimely communications problems.

56
 While on s heading of 270°, I was instructed t o turn to a southerly heading. I
checked the area off my left wing and saw aircraft B on a converging course t o mine.
His heading was about 300”.I advised Approach Control of the conflicting traffic
and told them I needed a turn to the north to avoid traffic. It was later determined
that aircraft B was not transponder-equipped and not seen by Approach Control
radar. They had turned me south to avoid departing aircraft C.

Sequencing disorders: Examination of ASRS reports shows that sequencing disorders and their
consequent traffic conflicts most frequently are caused by errors in a controller’s planning or judg-
ment of traffic spacing. However, communications problems enter the picture significantly; taking
a sequencing control message by an aircraft other than the intended recipient is a frequent
occurrence.

I was on final for runway 25. On contacting the tower 8-10 miles out I was told t o
reduce to approach speed as they wanted to get a departure out. We were No. 2 fol-
lowing a heavy aircraft B on 2-mile final. After B landed aircraft C was told to taxi
into position and hold. The pilot acknowledged. Aircraft B was told t o turn off at
the high-speed taxiway; B stated he could not make the high-speed turnoff. About
30 sec later Tower called “C hold your position, d o not take off.” This was stated
twice, then the controller asked C if he had started his takeoff roll and C stated that
he had and that he wanted permission t o make a right turn at the first taxiway. At
no time did I hear the tower clear C for takeoff.

Runway assignment errors: Use of the wrong runway for landing or taking off is a frequently
reported airport traffic area problem. In virtually every case the fault is a communications problem.
Most of the problems involve flightcrew misinterpretations of landing or takeoff clearances - some-
times in connection with a last-minute change in the runway assignment.

Visibility restricted but VFR. Aircraft A reported 4-mile final for runway 9, air-
craft B on right base for runway 12, aircraft C was in position for departure at
threshold of runway 9. Aircraft D was holding midfield on taxiway E for departure
on runway 9, aircraft D was instructed t o turn left heading 360 and cleared for
takeoff. Instead of departing runway 9, aircraft D started takeoff roll westbound
on runway 27 toward aircraft C and aircraft A . He was instructed t o abort and
stopped on the runway.

The communications errors leading t o these problems consist of phonetic similarities, trans-
positions, inaccuracies, ambiguities, garbling, and untimeliness.

Deviations from aircraft intended trajectory: ASRS reports record many instances of aircraft
departing from assigned altitudes or headings during approaches or departures in airport traffic area
airspace. Communications problems are an important factor in such trajectory deviation, sharing the
burden of causation with poor flying technique on the part of flightcrews. Most communications
problems involved misunderstood clearance; failures t o issue appropriate clearances and failures to
change frequencies properly also accounted for a significant number of the deviations. It is note-
worthy that altitude deviations predominated in this fault category.

57
 ATC cleared us to descend t o 2,000. At 2,500 we spotted aircraft B 3/4-1 mile
ahead at eleven o'clock at about 3,000. We advised ATC, controller asked o.ur alti-
tude, then stated he had cleared us t o 4,000, and advised us t o maintain visual
separation from B. Then he cleared us to climb t o 4,000. B passed above and behind
our position.

En Route Operations

The en route operations evaluation concerned communications problems that arose when air-
craft under ATC control were cruising en route or transitioning t o or from the cruise condition.
Many of these problems are traceable t o difficulties of ATC coordination within and between con-
trol facilities. Most of the problems that were related t o difficulties in air-to-ground communication
were of three types: altitude deviations, failures of flightcrews to respond effectively t o clearance
amendments for conflict avoidance, and a variety of difficulties related t o weather avoidance.

Aircraft A was cleared to cross 10 n. mi. east of XYZ at FL310, then descend to
FL240 at pilot's discretion 14 n. mi. east of XYZ; A's altitude readout was FL320.
At 10 n. mi. east the pilot was asked his altitude and he reported FL320. He was
issued traffic ten o'clock, 8 miles, flight level 330, northeast bound. I was on a busy
position alone and had conflict alert not come on, I would not have caught this in
time. Pilot called on frequency and apologized for not making his restriction.

Altitude deviations- One of the types of incidents that has the greatest potential for causing
serious accidents is failure that results in an aircraft being at an altitude other than that assigned by
air traffic control. This is especially serious when an altitude restriction has been issued during climb
or descent. Restrictions are issued because of conflicting traffic, and failure to comply will almost
always result in loss of separation.

We requested descent clearance. Center asked whether it was necessary that we begin
descent at that time, and I replied it was, or we would be unable t o get down with-
out circling. Center then gave us a 60" vector turn to the right t o avoid conflicting
traffic 2,000 ft below us. This was misunderstood by the copilot, and, as he had the
conflicting traffic in sight, began descent. I had switched t o monitor ATIS and did
not notice he had begun descent until he had reached FL320. I asked if he had
received further descent clearance. He had not and began climb back t o FL330.
Since we had the traffic in sight safety was not jeopardized, but it could have been
in IFR flight conditions.

Clearance amendment response- Clearances are usually amended because of potential or

present conflicts. A routine clearance may be issued, then a change in the traffic situation makes it
necessary that the clearance be amended. Altitude assignments, heading assignments, and speed
control are sometimes misunderstood by the recipient and the read-back, if given, may not be noted
by the controller. Such clearances are usually time-critical and errors frequently result in loss of
standard separation.

After obtaining position reports from departing aircraft A 3 n. mi. SE XYZ leveling
at 5,000 and amving aircraft B 30 n . mi. E XYZ, A was asked if he would like the

58
 Victor airway. He asked the radial and concurred and was recleared via Victor air-
way to maintain 5,000 and report passing the 246 radial of XXX VOR. B was
recleared via the north Victor airway to cross the 18 DME fix at or below 4,000.
Aircraft A reported crossing the 246 radial and was cleared to climb t o 7,000.
Shortly thereafter, Center reported getting aircraft A on radar and that A and B
were head-on about 8 n. mi. apart on Victor airway north. Clearance was issued t o
A t o stop climb and B‘ to stop descent but the aircraft sighted each other passing
about 1/2 mile apart at near the same altitude.

Weather avoidance- In periods of rapidly moving frontal activity or summer thunderstorms,

weather avoidance becomes a major factor in en route operations. Deviations from planned routes
become commonplace and there are frequent incidents involving uncoordinated penetration of air-
space assigned t o another aircraft o r to a different air traffic controller.

Two northbound aircraft B and C on separate jet routes at FL330 were deviating
from course to circumvent thunderstorm buildups. Aircraft A was a no-radio contact
aircraft in the same area and B and C were provided separation from A. B was
expected t o follow C but he had deviated right of course 5 t o 8 n. mi. Separation
from C was decreasing. Controller attempted vectors to increase separation but pilot
response was slow. Therefore, controller cleared B to FL290 and heading 330.
Again, pilot responded slowly but separation increased.

The Party Line Concept

A popular point of view among pilots is that there is substantial benefit in the “party line”
concept, that is, that monitoring of a communication frequency can provide useful information, for
example, about traffic flow and location of other traffic. Many pilots make extensive use of this
practice, particularly at noncontrolled airports or at lower-activity terminals served by a control
tower.

I was in aircraft A cleared through the airport traffic area at 2,500 MSL. While
passing over the field, I heard the tower clear aircraft B for takeoff on 3 1 L. He was
to climb to 3,000 MSL on a 270” heading. I kept looking, but was unable to see
him. The tower never did advise me he was coming. At about 6 miles west I saw him
as he climbed out from under my left wing. He was traveling extremely fast and
passed about 200 to 300 ft from me.

Some pilots contend that this is the usual means of acquiring a mental picture of the current
traffic situation.

The beneficial use of party line is mentioned incidentally in several ASRS reports that are con-
cerned with some other primary occurrence; this is because almost all reports concern an unsafe
incident or condition. This report is an illustration.

I was descending t o 1,500 MSL in aircraft A and had been told t o enter left down-
wind for runway 05, approaching the airport from the northwest. Aircraft B had
just departed runway 05 and asked for a left turnout t o the northwest. The tower

59
 approved and said, “No reported traffic.” Immediately, I began looking and saw
him approaching me on a collision course. Evasive action was required on my part.
Aircraft B passed my right wing less than a mile. I don’t think he ever saw me,
because the tower had told him there was no traffic NW of the airport.

Party line effectiveness must be evaluated in terms of the number of aircraft on a common
frequency that are pertinent to the traffic situation. In large terminals, approach control is sector-
ized; local control is sometimes divided by the runways in use; ground control is split; and departing
aircraft may still be in terminal airspace after being changed to center frequency. Many military air-
craft are equipped with UHF only. Frequency monitoring may disclose only a fraction of the traffic
that could be involved in an incident.

Errors can result from misufiderstanding an overheard transmission, when a pilot may initiate
an action based on his misperception of the message content. Some ASRS reports concern pilots
acting on a clearance intended for another aircraft. This is an example of such an occurrence.

Aircraft A was told t o taxi into position and hold on runway 35 while aircraft B
was on landing rollout. Aircraft C was told to taxi into position and hold on run-
way 23 behind aircraft D. Aircraft B, landing on runway 33, was holding short of
runway 23. When aircraft B turned off runway 33 and aircraft D rolled past the
intersection of runways 23 and 33, aircraft A was told t o turn right heading 090
after departure, cleared for takeoff. Aircraft C (holding on runway 23) thought the
clearance was for him and started takeoff roll. Aircraft A was rolling also. The local
controller did not hear aircraft C acknowledge for the takeoff clearance nor did he
see aircraft C start takeoff roll until it was too late. Aircraft A and aircraft C missed
by approximately 200 ft at the intersection.

Party line is also capable of making useful information available. The relative position and
flightpath of other aircraft can often be ascertained when they are not in view. The intentions of
the pilot may be overheard and, therefore, they may be taken into account in planning future
courses of action.

But the use of simplex communicatio~isalso poses the problem of misunderstanding by the
intended recipient of a transmission, the problem of blocking of reception by another transmission,
and the possibility of a clearance being accepted and acted upon by other than its intended
recipient.

CONCLUSIONS

It is concluded that ATC-aircraft communications problems are involved in a large fraction of

the occurrences reported to the ASRS. Many or most of these communications problems involve
human errors on the part of the sender or receiver of the message. A small number are associated
with breakdown of communications equipment, frequency saturation, and other system factors.
The air-ground communications link is the cornerstone of the present system for the control of air
traffic, and the problems observed in this study constitute a threat t o the integrity and safety of the
aviation system.

60
 REFERENCE

1 . Crayson, R. L.: Effects of the Introduction of the Discrete Address Beacon System Data Link on Ax/Cround
Information Transfer Problems. NASA Contractor Report 166165, July 1981.

61
 N8 1 3 116”-
5. INFORMATION TRANSFER WITHIN THE COCKPIT: PROBLEMS IN

mTRACOCKPlT COMMUNICATIONS

H. Clayton Foushee and Karen L. Manos

INTRODUCTION

During the last few years there has been a rising concern among those in the aviation system
about incidents and accidents attributable to the improper utilization of resources available to the
human element in the system. This is documented by reports made to the Aviation Safety Report-
ing System (ASRS), over 70% of which cite human errors. According to a study by Billings and
Reynard (ch. I), the most common difficulty is a failure of the information transfer process. These
information transfer problems include messages that are not originated; messages that are inaccu-
rate, incomplete, ambiguous, or garbled; messages that are untimely; messages that are not recieved
or are misunderstood; and less common, messages that are not transferred because of equipment
failure. Although problems occur at all points in the information transfer system, this report deals
with information transfer problems that occur within the cockpit and, more specifically, with com-
munication patterns among cockpit crew members.

Communication patterns among cockpit crewmembers probably play a more significant role
today than ever before. With the large air camers (many employing thousands of pilots) and their
complicated bidding procedures for assigning flight duties, it is possible that pilots may fly together
without having met before. Thus, responsibilities that may be implicitly understood by crews that
fly together frequently have t o be explicitly assigned to members of a newly assembled crew.

There is a growing consensus among human factors specialists, airline training departments,
and social and personality psychologists that communications patterns exert significant influences
on important performance-related factors. At the very least, communication patterns are crucial
determinants of information transfer, but research has shown that they are also related to such fac-
tors as group cohesion (important from a crew coordination standpoint); attitudes toward work;
and complacency. An argument in the crew room prior to departure can affect the interactional
patterns of the flightcrew for the rest of the day. Overbearing captains can severely inhibit informa-
tion transfer from subordinate crewmembers, even in potentially dangerous situations.

In the 1979 crash of a northeastern commuter camer aircraft, these factors appeared to have
played a pivotal role. The accident report (ref. 2) characterized the cockpit situation as follows:
“The captain was a company vice president with over 20,000 flight hours who was known to rarely
acknowledge checklist items or other callouts from any first officer. The first officer . . . had only
been with the company for 2 months. For the first year pilots are on probation, are not represented
by the pilots’ union, and may be terminated with or without cause . . . .” The first officer testified
that the aircraft was well below the glide slope on final approach and that he made all required
callouts t o that effect. The evidence suggests that the captain may have been incapacitated, but the
first officer could not confirm this. The end result was a crash in which the airplane impacted sev-
eral miles short of the runway. The NTSB concluded that “The probability of the first officer recog-
nizing and reacting t o any possible physiologic incapacitation in the captain was remote.’’ .
- - .- _ -

63
Preceding page blank
 As in the following report to ASRS, subordinate crewmembers can become “conditioned”
into this pattern so that they react similarly, even with captains who encourage open channels of
communication :

I was the copilot on a flight from JFK t o BOS.The captain was flying. Departure
turned us over to Center and we were given FL210 which was our flight plan alti-
tude. I noted we had reached FL2 10 and were continuing through it, but was reluc-
tant to say anything. As we climbed through 21,300 ft, I mentioned it t o the cap-
tain, but not forcefully enough, and he did not hear me. I mentioned it again and
pointed t o the altimeter. We were at 21,600 f t when the climb was stopped and we
descended back to 2 1,000. As we started our descent, Center called and told us t o
maintain FL210. The captain said he had misread his altimeter and thought he was
1,000 ft lower than he was. I believe the main factor involved here was my reluc-
tance to correct the captain. This captain is very “approachable” and I had no real
reason to hold back. It is just a bad habit that I think a lot of copilots have of
doublechecking everything before we say anything t o the captain.

It is the central thesis of this paper that the patterns of communication between cockpit crew-
members determine in a large degree the response patterns of the crewmembers.

APPROACH

An opportunity for a more systematic analysis of cockpit communication patterns and perfor-
mance was provided by data acquired in a NASA full-mission simulation study conducted by
Ruffell Smith and associates (ref. 1). In that study, fully qualified B-747 crews flew a simulated,
routine line trip segment (IAD-JFK) followed by a segment (JFK-LHR) in which a mechanical prob-
lem was introduced which necessitated an engine shutdown and diversion from the original flight
plan. The simulation included all normal communications, ATC services, weather, closed runways at
the diversion airport, and later, an inoperative autopilot which further increased pilot workload. As
reported by Ruffell Smith, the scenarios were constructed in such a way that good crew coordina-
tion, cockpit communications, decisionmaking, and planning skills were required, but they were not
complex enough t o preclude an entirely safe operation given proper performance and coordination.

The study reported in reference 1 allowed the examination of flightcrew performance in a very
controlled setting. Errors in performance were carefully monitored and recorded. Eighteen volun-
teer line crews flew the scenario, and marked variations in the behavior of the crews were observed.
Ruffell Smith and his colleagues noted frequent problems in areas related t o communication, deci-
sionmaking, and crew interaction and integration. As a result, it was suggested that one of the major
variables that influenced how effectively and safely a crew handled problems was the identification
and utilization of the various human and material resources available t o them. The presence or
absence of strong leadership seemed to mediate the frequency and seventy of the errors committed
by the flightcrews.

We undertook a study t o examine more closely what role the communication patterns of these
~

flightcrews played in the management of resources and performance. Complete cockpit voice
recordings were available for 12 of the 18 experimental runs. (Audio difficulties prohibited the

64
inclusion of the remaining six). These 12 runs were subjected to a rigorous content-coding tech-
nique in which an attempt was made to classify each statement or phrase into one of 13 categories.
Of the 13 categories, three were combined, leaving 10 usable categories. Obviously, some statements
were nonclassifiable and others were inaudible because of other cockpit noises, simultaneous
speech, etc., and these were not coded. The communication categories, definitions, and frequency
of occurrence of each category in this sample are shown in table 5.1.

TABLE 5.1 .- COMMUNICATIONS CATEGORIES OBTAINED FROM THE RUFFELL-SMITH

COCKPIT VOICE RECORDINGS

Number of Percent of all

Category De finit ion
citations communications"

Crewmember observation Remarks made by crewmembers aimed at 2316 44

orienting others to some aspect of fight
status (e.g., references to instruments, navi-
gation, etc.)

Commands Specific assignments of responsibility, usually 648 12

by the captain, in the form of declarative
statements

Inquiries Requests for information regarding some 1308 25

aspect of flight status, progress, etc.

Response uncertainty Statements indicating uncertainty or lack of 40 .8

information with which to respond to a
command, inquiry, or observation

Agreement Statements indicating agreement with'an 85 1.6

observation, command, crew response, etc.

Acknowledgments Recognitions of a command, inquiry, etc. 587 11

Tension release Laughter or humorous remarks 128 2.5

Frustration or anger Derision for an incorrect response or anger in 51 1

response to some occurrence

Embarrassment Any response apologizing for an incorrect 26 .5

response, etc.

Pushes Repetitions of already administered com- 27 .5

mands or inquiries

'Total of 5,216 communications.

65
 RESULTS AND DISCUSSION

The number of occurrences in each category was tabulated and statistically analyzed. Overall,
there was a tendency for crews who did not perform as well t o communicate less, suggesting that as
expected, poor crew coordination tends t o result in more marginal performance.

A number of interesting relationships emerged between the various categories of communica-

tions and the types of errors made in the simulation. A negative correlation’ ( r = -0.51) between
crewmember observations and systems operational errors was obtained. This relationship appears
quite logical. When more information regarding flight status was transferred?there were fewer errors
related t o system operation (e.g., mishandling of engines, hydraulic and fuel systems, misreading
and missetting of instruments, and failure t o use ice protection). The relationship should serve as
important evidence in support of the concept of cross-checking and redundancy among the cockpit
crew.

Similarly, there was a strong negative relationship ( r = -0.61) between systems operational
errors and acknowledgments. When crews frequently acknowledged commands, inquiries, and
observations, these kinds o f errors were less apparent. It would appear t h a t acknowledgments serve
an important function of validating that a certain piece of information has been transferred. These
kinds of communications also serve as reinforcements t o the input of other crewmembers. Frequent
acknowledgments were also associated with a lower incidence of tactical decision errors ( e g ,
amount of fuel dumped, flap settings, and braking). Most significant, however, is the fact that
acknowledgments were strongly negatively associated with total errors (r = -0.68). “Low error”
groups tended to acknowledge communications more often, overall.

Commands were associated with a lower incidence of flying errors ( r = -0.64) (engine handling,
neglect of speed limits, altitude errors, and the lack of formal transfer of control of the aircraft
between captain and first officer). Often communications of this type seem to ensure that cockpit
duties are properly delegated, but it is also suggested that too many communications of this type
can have negative consequences. The use of commands appears to be one of the significant variables
of interpersonal style. An identical piece of information can be related t o other crewmembers in
one of several different ways. For instance, a communication such as, “Check the plates for the
profile descent,’’ which would constitute a command, could also be relayed, “I think we should
check the plates for the profile descent,” an observation; or “Why don’t we check the plates for the
profile descent?”, an inquiry. Although it is rare that a single communication will alter or affect the
cockpit atmosphere, repeated commands from the captain, for instance, might prove intimidating
to subordinate crewmembers. First officers have often reported that the styles of some captains can
inhibit them from offering information even in potentially dangerous situations. Although the use
of command communications is both necessary and beneficial in many situations, it is possible that
other types of communications give the recipients of information a greater sense of responsibility.

Also interesting was the finding of a negative relationship between agreement and total errors
(r = -0.61). “Low error” crews tended to exhibit more agreement; however, it is not possible to

‘The correlation coefficient, “r,” varies from -1 to + l ; it is a statistical measure of the strength of association
between two variables. A high correlation coefficient, either positive or negative, indicates a considerable degree of
correspondence between the variables. A coefficient near zero indicates very little relationship between the variables.

66
determine whether the lack of agreement in “high error” crews was caused by the errors being com-
mitted or whether this lack of agreement tended to cause certain errors.

Additionally, there was some evidence of higher rates of response uncertainty ( r = 0.68):
frustration/anger (r = 0.53),and embarrassment (r = 0.53) in “high error” crews, although again it is
difficult t o determine whether the communication patterns gave rise t o the errors or vice versa.

In short. within-cockpit communications patterns seem t o be significantly related t o perfor-

mance. The lack of appropriate communications between cockpit crewmembers is an obvious prob-
lem, but it is also apparent that the style of communication plays an important role. In an effort t o
obtain further documentation of problems of this nature, a search of the ASRS data base was
undertaken. Eightyeight reports were positively identifiable as instances of communications prob-
lems within the cockpit. It should be noted that the frequency of incidents reported t o ASRS where
within-cockpit communication problems were partially causal is probably higher. However, the
narrative descriptions are sometimes not specific enough with respect t o antecedent conditions t o
allow classification. Moreover, it appears that pilots are reluctant to report instances of interper-
sonal conflicts in the cockpit.

Nonetheless, many pilots did report communication proble’ms with other crewmembers as
causal factors. Of these reported within-cockpit communications difficulties, 35% cited problems
with crew coordination dealing with poor understanding and division of responsibilities. Often the
lack of appropriate acknowledgments and cross-checking was a factor as in the following example:

On takeoff from Atlanta, our flight used an improper heading for climb-out . . . .
During this period, immediately after liftoff and while engaged in the climb-out and
weather analyzing, I said “075, the heading, right?” He (the first officer) looked at
me quizzically and said what I understood as, “Yeah, OK.” We continued t o climb
and were told t o contact departure . . . they told us to turn immediately to 110 . . .
then asked, “What heading were you cleared to?” I said, “Tell them 075’, that’s
what we read back wasn’t it?” The copilot did not answer me, so I looked at him
and he again had that odd look. I repeated louder, “Tell them 075 . , .” At n o time
were we aware of the serious problem with the other aircraft that unbeknownst t o
us had taken off on runway 8 . . . I believe it was solely due t o poor cockpit com-
munication . . . . I thought 075’ was the correct departure heading, and to confirm
it, I asked the copilot. But my question came at a time when we were very busy
. . . he thought I was asking his evaluation of a direction. Coincidentally, that, t o
him, was a good direction and he answered in the affirmative. I took his answer as
a concurrence of my question of proper takeoff heading. . . .

As in the Ruffell Smith study, the lack of appropriate acknowledgments and cross-checking is
a frequently occumng problem. It was previously noted that acknowledgments probably serve the
purpose of verifying that information has been transferred and that it is correct. Thus, the use of
these kinds of communications is important for monitoring and cockpit crew redundancy purposes.
This function is clearly illustrated by the air carrier captain who submitted this report to ASRS:

Departed SFO on Porte One departure cleared. Misread clearance and began turn at
2500 and about 4 miles instead of at least 2500 and 6 DME miles. WX was VFR and
terrain clearance was no problem. Turned left to 180’ and Departure Control

67
 requested 1 turn back t o 280'. I did so immediately and shortly after back t o 180'.
Did not observe any traffic. No evasive action. Read clearance incorrectly and
copilot had his head up and locked. He should have caught this if he had been mon-
itoring this departure. Airline lacks adequate training program. Cockpit standardiza-
tion is a joke.

A total lack of communication between cockpit crewmembers was cited as a factor in 12% of
the reports, thus lending credence to the positive relationship between communication and good
performance obtained from the Ruffell Smith data. The following typifies the potential severity of
such occurrences:

I was pilot in command of airlines flight . . . Chicago t o El Paso. The checklist was
completed and I lined up for takeoff. (As throttles were advanced, the takeoff
warning horn sounded.) Following checks . . . the warning horn sounded two more
times and then stopped. My hand was in the process of transitioning back to the
thrust levers with the intent of aborting the takeoff . . , . I was at 60-70 knots
when the horn stopped. I elected t o continue the takeoff since in my judgment all
the parameters for a safe takeoff were accomplished. What I did not know, however,
was that the warning horn circuit breaker was pulled by the engineer without his
advising me o f such action . . . . He preempted my actions by pulling the circuit
breaker, an action I never intended to make , . . not advising the crew . . . resulted in
a false impression of a safe configuration for flight.

There were numerous examples in this category of cockpit crewmembers not communicating
regarding errors even when they had access to the correct information. Misunderstood clearances
(examples of ground-to-air communication problems) posed frequent problems; and, as in the
following example, these often evolved into within-cockpit communication difficulties:
~

After takeoff from LGA runway 4 on a heading of 055, Departure Control cleared
my flight to 12,000 ft on a heading of 340", but which I misunderstood and thought
I heard him say 240" heading. After completing the turn to 240", he called my flight
t o ask my altitude and heading. My first officer returned the call with 6,000 and
240" heading. The departure controller then said we were supposed t o be on a 340"
heading and t o turn t o 340'. I started my turn to 340" and as I was rolling through
270°, he gave me a heading of 2 10". I then told Departure Control I understood my
heading was originally 240°,but he came right back and said he had given me 340'
heading. I had misunderstood him along with my second officer, who said t o me he
had understood the heading was 240". This mistake on our part was confirmed by
the first officer who said he repeated back to Departure 340' and had set his head-
ing bug on 340". My heading bug was on 240". I think the contributing factor to
this incident was the first officer not saying anything to me about our heading when
I rolled past 340" in m y turn to 240". This lack of attention to our flight progress
and his not bringing.it t o m y attention was why the 240" heading occurred. It is my
company's procedure for required callouts; with callouts for deviations from head-
ing, altitude, etc., when they occur at any time during any phase of flight. My first
officer did not make this callout as I rolled in my turn through the 340' heading he
knew we were supposed t o be on and which I had misunderstood to be 240" instead
of 340".

68
 Over 15% of the reports dealt with information which was believed by one o r more crewmem-
bers t o have been transferred, but for one reason or another (e.g., interference or inadequacy of the
message) was not:

I was monitoring an autopilot single autoland approach on flight . . . t o runway 22L

at EWR in visual conditions when a GPWS “pull-up” warning occurred . . . the entire
crew’s attention was directed to confirming configuration position, speed, and sink
rate, all of which were normal. I commanded, and simultaneously with the copilot,
selected 50” flaps, considering that the landing flaps not selected mode was possibly
the reason for what I considered t o be “probably” a false warning, due t o a possible
failed flap position switch. The FE said something which 1 could not clearly grasp
while the “pull-up” was sounding. As the flaps extended to SO”, the GPWS warning
silenced. A normal auto landing occurred. Sounds simple and somewhat everyday -
until during taxiing I was informed that the FE had “inhibited” the GPWS without
being commanded t o . . . . He stated he asked me if I wanted it cancelled, but I
didn’t reply so he assumed I did.

In a number (1 0%) of instances, communications between cockpit crewmembers were defi-

cient as a result of overconfidence and complacency - assumptions that everybody understood
what was taking place, when in reality they did not. In about 5% of the reports a lack of confidence
in subordinate crewmembers caused captains to become overloaded by trying t o d o too much. This
was seen in the Ruffell Smith study and is a frequent contributor to the hesitancy of subordinate
crewmembers to question the actions of captains.

On arrival at DCA landed runway 33 and made a left turn on taxiway K which leads
t o the gates after being cleared to the gate by Ground Control we were in a position
that happens very often. We were blocking the taxiway from Page to the active run-
ways with aircraft waiting for us to clear. There was another aircraft parked t o our
right and an aircraft parked to our left which was waiting to push back, and the air-
craft agent was standing behind the right wing watching as we started between the
two aircraft. After checking the clearance on the left wing and getting a clear signal
from the agent, I leaned forward and t o the right t o check the clearance on the right
wing. This was done because I had a new copilot and did not trust his judgment. As
I turned back t o check the left wing, the agent was giving a hold signal. At this time
the left wing tip was just under the tail cone of the center engine. The indicator for
the nav. light (a small piece of plexiglass that stands about 3 inches above the wing)
hit the tail cone breaking the top of the plexiglass off.

Another common problem ( 1 6%) was interference with pertinent cockpit communications by
extraneous conversation between crewmembers or between crewmembers and flight attendants.
This often reflected a cockpit atmosphere that had become too relaxed, as in the following:

While I was in the passenger cabin waiting t o go to the bathroom (this took over
15 min because of crowded conditions and bathroom hog), the copilot failed t o
turn down 5114 toward DEN. (We were over ONL bound for LAX). When I
returned t o the cockpit, we were 86 miles from BFF. About that time DEN center
asked us our routing t o LAX. 1 told him we should have been on J114 and asked for
a vector to DEN. We were about 70 miles off course at the time. I have never been

69
 that far off course before. The copilot was telling war stories (he was a fighter pilot
in Vier Nam) t o the engineer and wasn’t payitig any utterition to his job. I told him
his behavior was inexcusable, but I’m afraid the man didn’t really understand how
serious this could have been. This was a BOS-LAX nonstop of about 6 : 0 5 and I
usually have to leave the cockpit a t least once.

Finally, several reports dealt with role and personality conflicts creating a state of affairs on
the flight deck such that communications between crewmembers have completely deteriorated :

I was the first officer on airlines flight into Chicago O’Hare. The captain was tlying,
we were on approach to 4R getting radar vectors and moving along at 250 knots. On
our approach, Approach Control told us to slow to 180 knots. I acknowledged and
waited for the captain t o slow down. He did nothing, so I figured he didn’t hear the
clearance. So I repeated “Approach said slow t o 180,” and his reply was something
to the effect of “I’ll d o what I want. ” I told him at least twice more and received tlie
same kind of answer. Approach Control asked us why we had not slowed yet. I told
them “We were doing the best job we could” and their reply was “You almost hit
another aircraft.” They then asked us to turn east. I told them we would rather not
because of the weather and we were given present heading and maintained 3,000 ft.
The captain descended to 3,000 f t and kept going to 2.300 f t even though I told I i i r n
our altitude was 3,000 f t . His comment was “You just look out tlie damn window. ’’
Finally we were cleared for the approach.

In this report, the first officer was apparently providing information in accordance with his
responsibilities, but the information was unheeded. One wonders what occurs in this particular
captain’s cockpit with a less assertive first officer.

SUMMARY

It is apparent from the data that cockpit communications patterns are related to flightcrew
performance. It would be a mistake, however, to infer from these data that more communication
among flightcrew members necessarily translates into better performance. The type and quality of
communications are the important factors - not the absolute frequency. These observations are
reinforced by a number of reports from ASRS in which it is implied that within-cockpit communi-
cations difficulties are significant factors in the information transfer problem.
 REFERENCES

1 . Ruffell Snlith, H . P.: A Simulator Study of the Interaction of Pilot Workload with Errors, Vigilance and Deci-
sions. NASA TM-78482,1979.

2. Arcraft Accident Report No. AAR-80-1. National Transportation Safety Board, 1980.

71
 ’ ??$I-31168
6. INFORMATION TRANSFER DURING CONTINGENCY OPERATIONS:

EMERGENCY AIR-GROUND COMMUNICATIONS

Richard F. Porter

INTRODUCTION

Although radio communication is always crucial to the safe and efficient operation of the
nation’s aviation system, the necessity for the effective transfer of information is never more clearly
obvious than when an aircraft is in distress.

The reports in the data base of the Aviation Safety Reporting System (ASRS) provide a unique
resource for identifying deficiencies that have occurred in communications in actual emergencies,
and for a pragmatic evaluation of the causes and consequences of information transfer dysfunctions.

This report presents the results of a study of safety-related problems that have occurred as a
consequence of communications problems in emergency situations. All information was obtained
from a review of pertinent reports contained in the ASRS data base.

OBJECTIVE

The objectives of this study task were: (1) t o describe the safety-related problems occurring as
a consequence of information transfer deficiencies that arise when air/ground communications are
(or should be) used as a resource in in-flight emergency,situations, and (2) t o define the system
factors, the human errors, and the associated causes of these problems.

SCOPE

This study is based on a search of 13,000 ASRS reports submitted between May 1 , 1978 and
September 1 7 , 1980. For any particular incident to be pertinent to the study, three conditions were
necessary. The qualifying elements were:

1. An emergency situation

2 . An information transfer problem arising in radio communication associated with the

emergency

3. An ensuing safety-related problem

To obtain a more appropriate sample of incidents, the study was not confined to formalIy
declared emergencies. Instead, an “emergency situation’’ was defined as an unforeseen combination

73
Preceding page blank
of circumstances that calls for immediate action to avoid disaster. This interpretation is based on a
standard dictionary definition and is compatible with the statement in the Air Traffic Control
Handbook (FAA, 71 10.65A, ll 1550) - “. . . an emergency includes any situation which places an
aircraft in danger; i.e., uncertainty, alert, being lost, or in distress.”

APPROACH

The approach to the study consisted of the development and implementation of a search
strategy for the extraction of pertinent cases from the ASRS data base, followed by an analysis of
relevant information from those cases.

Each report in the ASRS data base

TABLE 6.1 .- DESCRIPTORS SELECTED FOR contains key descriptive phrases in five
FURTHER SEARCH fields. These are the enabling factors,
associated factors, recovery factors, sup-
Situation Descriptor plementary factors, and descriptors.
Low fuel In-flight engine shutdown
The initial screening consisted of a
In-flight engine shutdown
search for all reports that contained the
Engine
Engine problem word “emergency” in any phrase in any of
Aircraft engine-out performance the five fields. Of the 176 reports pro-
Aircraft equipment-problem/engine duced, 45 were judged to be either dupli-
Power plant problem cate reports of other incidents or were
Engine control problem otherwise deemed t o be unsuitable for this
Aircraft equipment-operating- study. This initial screening therefore
problem/engine yielded 13 1 discrete emergency situation
In-flight fire reports for later analysis.
Weather IMC in VFR flight
Weather in terminal area
The next step was t o expand the
Weather forecast study sample by examining the 131 reports
Weather report to identify other descriptive phrases with a
Weather avoidance relatively high frequency of occurrence, the
VFR on top hypothesis being that key words that
En route weather appear to be correlated with emergencies
would be likely to be useful in finding
Landing Precautionary landing other pertinent study cases, even though
Return/land the word “emergency” did not occur in a
Off airport landing descriptor.
Pressurization Aircraft equipment-problem
Table 6.1 lists the key descriptors that
pressurization
were selected. In addition, the vocabulary
Command authority Command authority
of descriptors used in the ASRS was exam-
ined t o find any further descriptors that
Disorientation Disorientation could be (intuitively) associated with emer-
Spatial disorientation gency situations and that appear a reason-
able number of times in the data base.

74
Three additional descriptors were selected : blind broadcast, sensory illusion, and loss of aircraft
control.

Excluding the 176 reports previously examined, it was found that at least one of the
26 selected descriptors appeared in 850 reports in the data base. From these, 37 were judged t o
satisfy our definition of an emergency situation.

Combining the 131 cases from the first search with the 37 additional cases gave a total of
168 emergency incidents,

The final phase of the data search consisted of segregating those cases that satisfied the two
remaining study criteria from those that did not. Each of the 168 emergency cases was carefully
examined t o identify ( 1) any information transfer problem arising from air/ground communication
(or lack of it) associated with the emergency; and (2) any safety-related problem that may have
followed such an information transfer deficiency. Only those cases that exhibited both of these
features were retained.

Fifty-two instances were found in which the study criteria were satisfied. The analysis of the
52 pertinent cases consisted of seeking common factors t o permit the classification of all cases with
regard to the nature and effect of the information transfer problem.

RESULTS

For an incident to be germane to this study, three elements were essential:

1. An emergency situation

2. An information transfer problem or dysfunction associated with the emergency situation

3 . An ensuing safety-related problem

In this section, the results of the ASRS data base search are summarized with respect to all
three of these elements. The intent here is t o document the search results with as little interpreta-
tion as possible. The discussion and interpretation is presented in the following sections.

Emergency Situations

The selection process yjelded 168 distinct cases in which an emergency situation existed. The
total population of emergency cases is broken down in table 6.2 by generic type and by the opera-
tional category.

For the purposes at hand, “propulsion problem” includes any difficulty with an engine, pro-
peller, or fuel system; “VFR pilot/weather” includes all situations in which a noninstrument-rated
pilot is in instrument meteorological conditions or is in imminent danger of being forced into such
conditions; “aircraft systems” includes problems with any aircraft subsystem other than the

75
 TABLE 6.2.- TOTAL POPULATION OF EMERGENCY SITUATIONS

Number of cases
Emergency type ~

Total ?ercent
General number if total
Air carrier Military
aviation
~~

Propulsion problem 12 1 22 35 20.8

VFR pilot/weather 0 0 34 34 20.2
Aircraft systems 16 3 8 27 16.1
Low Fuel 8 6 12 26 15.5.
Severe weather avoidance 7 0 1 8 4.8
Lost 0 0 7 7 4.2
Miscellaneous 10 4 6 20 11.9

Subtotals 53 14 90 157 93.5

Unspecified .-. __- -_- 11 6.5

Total 168 100 .o

propulsion system; and “severe weather avoidance*’ includes threats from thunderstorm activity or
severe turbulence. The “unspecified” type of emergency is one in which the reporter describes a
problem associated with an emergency condition but does not explain the nature of the emergency
itself. The remaining type designations are selfexplanatory.

All but 7 of the 168 emergency situation reports contained an explicit or implied reference to
radio communication. Regarding the 7 that did not, it cannot be concluded that radio was not used
because the nature of the problem was such that the immediate value of such communication would
have been limited, and reference may have been omitted by the reporter. Typically, these cases
involved off-airport landings by light aircraft operating VFR and were caused by sudden engine
stoppage.

Information Transfer Problems

In 5 2 of the 168 emergency cases there were subsequent communications problems. These
5 2 cases were all pertinent t o the study and constitute the total data set of the investigation.

Table 6.3 lists the number of emergencies in the final data set, using the same classifications of
emergency type and operational category as were used in table 6.2.

The nature of the information transfer problem in each case is listed in table 6.4. The first
column in the table gives the source, o r instigator, of the problem as perceived by the reporter in
each case.

Only the primary problem in each case is listed in table 6.4, but it should be noted that more
than one problem is evident in some cases. For example, Case No. 2 is primarily a language problem,
but is complicated by a lack of ATC coordination. To quote from the narrative:

76
 TABLE 6.3.- NUMBER OF EMERGENCIES WITH INFORMA-
TION TRANSFER DEFICIENCY AND SUBSEQUENT
SAFETY PROBLEM
I
I
Number of cases
Emergency type
IAir carrier Mihtary
General
aviation
Total
cases

Propulsion problem
VFR pilotlweather
Aircraft systems
2 1 ;
0
2 1
6
4
5
9
4
8
Low fuel 0 4 2 6
Severe weather avoidance 1 0 1 2
Lost 0 0 6 6
Miscellaneous 1 4 1 6
Subtotals 6 10 25 41
Unspecified -- 11

1
I
Total 52

TABLE 6.4.- INFORMATION TRANSFER

PROBLEM

1 Source 1 Problem
1 Number
of cases I
ATC Lack of coordination 11
Controller inattention
(distraction by emergency) 8
Lack of perception of emergency 7
Faulty technique 4
Frequency congestion 1
Auditory interference in tower I

Aircrew Language difficulty

Reluctance to declare 5
Failure to use all resources 1
Pilot communication errors 5

I
I
Other 1
I
Radio/radar limitations
Equipment failure
I I

77
 While on duty as an air traffic controller at the DAB Tracon, an alarmed pilot began
screaming Mayday, obviously wanting assistance on frequency 12 1.5. For approxi-
mately 10 min, between transmissions to other aircraft I tried t o radar identify the
aircraft. He had a very poor (self-admitted) understanding of the English language.
His number was never found out until he landed. Due t o the position of the aircraft,
ORL Approach was able to assist aircraft along with myself. Neither I nor ORL
Approach knew of each others’ intentions which added t o confusion . . . .

(Note: The case numbers referred t o in this chapter are described in an appendix t o the Battelle
Columbus Laboratories Report (ref, 1) from which this chapter was derived. Copies of the original
report are available upon request.)

Although most of the problem types given in table 6.4 are self-explanatory, the “controller
inattention” and “lack of perception of emergency” may require clarification.

The “controller inattention” in all cases pertains t o a problem caused by the distraction of the
controller by an emergency. This type of problem appears as a communication deficiency with one
o r more aircraft other than the one actually involved in the emergency, whereas all other categories
describe a problem between the subject aircraft itself and a ground facility. An example is described
in the narrative of Case No. 23, reported by the pilot of a fourengine air-carrier aircraft:

Aircraft A was cleared t o land . . . runway 1L by Tower. After touchdown, we

observed a light (white) aircraft B, in position for takeoff at about midrunway. Prior
t o A’s landing, B had been cleared into position and forgotten by tower controller
due t o a landing emergency situation on runway 1R . . . . Aircraft was stopped with-
out incident.

The “lack of perception of emergency” category appears in some cases as an attitudinal prob-
lem (as viewed by the reporter), or as a simple lack of understanding of the problem by the con-
troller. In either event, this study considers these cases as situations in which the flightcrew has
apparently not been able to communicate the urgency of the situation to the controller. For what-
ever reason, the ground controller in these cases does not seem to share the aircrew’s concerns.

An example from the narratives may clarify this categorization. From Case No. 26, reported
by an air-camer pilot on an IFR approach:

About 5 miles from the marker we were told that the 22 ILS was now in use and t o
turn left to 300 for vectors, We did this and were faced with a 15-mile line of
thunderstorms on the radar with heavy contour. We stayed on that heading but told
the approach controller that we had t o deviate soon. He said other aircraft were
going through the area with no problems (nobody crashed). An air carrier heavy was
asked how the ride was and he said it wasn’t too bad, however he would not recom-
mend sending anybody else through that area. I . . . elected to d o a 180 and get out
of the area. Approach was upset t o say the least . . . and from then on very uncoop-
erative . , . . He said . . . that if I could not comply with a request to fly into a cell
area there was nothing further he could do for me.

78
 The listings of table 6.4 may seem to place an imbalance of blame on the controllers. In this
regard, it must be pointed out that not only were 68% of the emergency cases handled without
incident, but 12 reporters (pilots) explicitly praised the assistance received from the air traffic
control system.

Ensuing Safety-Related Problems

In each of the 52 cases of table 6.4, the information transfer problem had an adverse influence
on the resolution of the emergency situation or created a separate problem.

The postemergency problems created by the information transfer deficiency can be placed in
two general categories: those in which the problem is essentially a continuation or worsening of the
original emergency, and those representing a new and different hazard. The 5 2 pertinent cases are
almost equally divided between those two categories, with 24 of the former and 28 of the latter. A
further breakdown of the nature of the postemergency problem is given in table 6.5.

TABLE 6.5 .- CLASSIFICATION OF ENSUING

SAFETY-RELATED PROBLEMS

Number
General type Nature of problem
of cases

Continuation or Delay in landing

worsening of Extension of emergency
existing emergency Landing without
emergency equipment
Unnecessary increase in
workload

New/different Traffic conflicts ( W A C /

hazardous LTSSlrunway conflicts)
situation Involving emergency
aircraft 4
Not involving
emergency aircraft 10
Operation without appropriate
clearance 8 .
Noninstrument pilot in IMC 2
Hazard of delay in terminating
emergency state 2
Hazard of collision with other
aircraft or ground 2
I

79
 DISCUSSION

In a search of over 13,000 reports in the ASRS data base, 168 emergency situations were
uncovered. Of these, over two thirds were resolved with no information transfer problem in
evidence.

The remaining 52 cases, which comprise the data set for this study, were categorized into 12
different types of information transfer problems and 9 distinct subsequent safety-related problems.
Although any extensive quantitative breakdown of the data must be interpreted with caution
because of the small number of cases, the data exhibit interesting features which are examined in
this section.

I
Correlation of Communication Problem with Type of Emergency

Comparing the data in tables 6.2 and 6.3, it appears that some types of emergency situations
rarely are associated with communication difficulties, while the opposite is true for others.
Table 6.6 summarizes the frequency of occurrence by type of emergency.

-
Number of Percentage with
Number
Emergency type information information
of cases
transfer problems transfer problems

Propulsion problem 35 9 26
VFR pdot/weather 34 4 12
Aircraft systems 27 8 30
Low fuel 26 6 27
Severe weather avoidance 8 2 25
Lost 7 6 86
Mjscellaneous 20 6 26
Unspecified 11 11 100
-

For five of the seven causal factors other than “unspecified,” the percentage of emergency
cases that produced communication problems is remarkably constant, ranging only from 25% to
30%. The remaining two causal factors, however, show a wide variation. For the “VFR pilot/
weather” grouping, the frequency of occurrence is very low (4 out of 34), whereas the “lost” group-
,ing shows six out of seven cases pertinent to. the study. Both of these apparently anomalous groups
are populated entirely by general aviation pilots.

Half of the six pertinent cases in the “lost” group are identified as student pilots; but two of
these correctly used their radio t o make known their problem and the third was handicapped by a
language limitation and rather cavalier earlier treatment by a controller. Two of the other cases also
mention a language problem, and in the remaining case the.aircraft was not within direct range of
any ground radio facility. In short, the high frequency of occurrence within this group in the study
set may be coincidental and insignificant because of the small number of cases involved.

80
 Conversely, the infrequent appearance of communications problems within the “VFR pilot/
weather” group is particularly difficult t o explain. Again, as a group, these pilots appear t o use their
radio properly and effectively once they are enmeshed in the emergency condition. At the same
time, the evidence suggests that this type of problem is handled particularly well by ATC.

Most Frequent Safety-Related Problems

From table 6.5, there are 14 instances of traffic conflicts as a consequence of emergency-
related communication problems. Furthermore, 10 of these conflicts do not involve the aircraft
that was the subject of the emergency. If we add t o these the 8 cases of operation without appro-
priate clearance and the 2 instances of collision hazards brought about by ambiguous or unsafe ATC
instructions, the direct or implied hazard of collision appears in 24 of the 52 cases.

The next most frequent problem is a delay in the landing of the emergency aircraft, appearing
in nine cases, and the emergency was prolonged through a delay in rendering assistance in eight
cases. In five cases, the subject aircraft landed without the alerting of emergency equipment.

From table 6.5, it may be inferred that the degree of success in resolving the emergency
depends on whether one adopts the point of view of the emergency aircraft itself or the other
airspace users. In all cases, the emergency was handled satisfactorily in that n o accidents occurred.
On the other hand, as pointed out above, 46% of the study cases (or 14% of all emergencies) suggest
an actual or potential hazard t o other aircraft in the system.

Correlation of Cause and Effect

Table 6.7 is an illustration of the apparent degree of correlation between the information
transfer deficiency (along the top) and the ensuing safety-related problem (along the side). The
number in each intersection is the number of cases in which the particular cause and effect have
been found.

The most frequently observed cause and effect combination is seen t o be a traffic conflict
caused by the distraction of a controller by the emergency. Interestingly, none of the seven cases of
this combination involved a traffic conflict with the emergency aircraft itself.

Other relatively frequent combinations are traffic conflicts caused by a lack of ATC coordina-
tion (see ch. 3) and delays in landing caused by the pilot’s reluctance t o declare an emergency
condition.

Because of the relatively large number of cause-andeffect categories and the relatively small
number of study cases, any further attempt t o find significant correlations in table 6.7 is not
justified.

81
 N

I
N - I
 CONCLUSIONS

An extensive statistical correlation of cause and effect among reported information transfer
problem cases cannot be made because of the relatively small number of pertinent cases and the
broad variety of communication problems and ensuing hazardous situations. Nevertheless, from a
review of the 168 emergency incidents in the ASRS data base, the following conclusions may be
drawn :

1. The most common safety-related problem is a traffic conflict not involving the emergency
aircraft itself. Traffic conflicts, direct or implied, are a hazard in 46% of the communication prob-
lem cases and in 14% of all emergency situations.

2. The most common infomation transfer problem arising from an emergency is a lack of
interfacility coordination within ATC. The second leading problem is controller inattention caused
by distraction resulting from an emergency.

3 . Emergency situations in which noninstmment-rated pilots are trapped in instrument

meteorological conditions lead t o subsequent communications problems much less frequently than
other emergency situations. It may be inferred that controllers, for whatever reason, are particularly
sensitive and adept with regard to these emergencies.

4 . There is no evidence in these data that pilots, even with minimum experience levels, are not
familiar with proper emergency radio procedures.

83
 REFERENCE

1. Porter, Richard F.: Information Transfer Problems in Emergency Airaround Communications. Battelle Colum-
bus Laboratories, ASRS Office, Mountain View, California, October 31, 1980.

84
 NEI 1 - 3 I169
7. THE INFORMATION TRANSFER PROBLEM: SUMMARY AND COMMENTS

C. E. Billings and E. S.Cheaney

INTRODUCTION

The United States national aviation system is indisputably the world’s most efficient and
effective. It moves huge numbers of people and large volumes of cargo safely and routinely in good
weather and bad. Its aircraft are the backbone of the world’s civil air fleets. Despite ever-increasing
pressure from environmental interests, population growth, and fuel-cost increases, the system has
thus far kept pace with demand.

It may seem perverse, then, in the foregoing chapters, t o have dwelt upon the system’s prob-
lems rather than its achievements; to have focused on its failings rather than on its enormous
success. But the system has been so successful over the last five decades precisely because of its
tolerance of and its constructive attitude toward criticism, whether from within or without. It was
in order t o solicit critical comment that the FAA’s Aviation Safety Reporting Program was estab-
lished in 1975; its mission, and that of the corollary NASA Aviation Safety Reporting System, is to
highlight deficiencies and discrepancies in the national aviation system. This study is offered as a
constructive attempt to illuminate a problem that bears directly upon system safety.

In this concluding chapter, we broaden our focus of consideration from the specifics of the
foregoing chapters to the aviation system aS a whole - and the information transfer problems that
are found within it. An attempt is made to characterize these problems independent of the settings
in which they occur, and in so doing, t o suggest possible intervention strategies for consideration by
the designers, managers, and operators of the national aviation system.

SUMMARY OF FINDINGS

Information transfer problems are perceived by experienced analysts to exist in a substantial

majority of all reports submitted t o the Aviation Safety Reporting System. The absolute incidence
of such problems cannot be deduced from these data, but during the period of study (May 1978 to
July 1980), on average, over 4,800 such problems were reported per year.

Over one third of these problems involve the absence of information transfer in situations in
which, in the opinion of the analysts, the transfer of the information could have prevented a poten-
tially hazardous occurrence. In another third, information transfer took place, but it was adjudged
incomplete or inaccurate, leading in many cases t o incorrect actions in flying or controlling aircraft.
One eighth of the reports involved information transfer that was correct but untimely (usually too
late to be of assistance) in forestalling a potentially hazardous chain’of events. In one tenth of the
reports, the information was transferred but was not perceived or was misperceived by the intended
recipients. The remainder of the reports involved equipment problems and a variety of miscel-
laneous specific conditions.

85
 SUMMARY OF ENABLING FACTORS

The focus of these studies was not primarily upon whether information transfer problems
exist, but upon the factors that appear t o be responsible for their existence. It is not possible from
retrospective data to state whether such factors cause the problem under examination, but it is
possible to state with confidence that certain factors are frequently found in association with infor-
mation transfer problems and that they may be causative. Several facets of the information transfer
phenomenon were examined in an effort to find factors in common. Such common factors were, in
fact, observed.

The human behavioral attributes found frequently in association with information transfer
problems, in rough order of frequency, were:

1. Distraction

2. Forgetting

3 . Failure t o monitor

4. Nonstandard o r ad hoc procedures or phraseology

5 . Complacency

When present, these attributes were associated with failures at all points in the information transfer
chain.

In addition to these human factors, certain system factors were also found o r imputed t o be
associated with information transfer failures. These factors included:

1. Nonavailability of traffic information

2. Degraded information

3. Ambiguous or (rarely) absent procedural guidance

4. Environmental factors (noise, confusion)

5 . High workload

6 . Equipment failure

The first three of these, of course, are in themselves information transfer problems. The fourth was
associated with difficulty in performing the tasks required. The fifth was associated with task
demands that could not be met by the worker. The last, though present in significant numbers, was
the least frequent system factor reported.

86
 Many other factors were observed in specific contexts; they are discussed in preceding chap-
ters. The factors listed above, however, appeared to be of general importance. Possible reasons for
this are set forth in the remainder of this chapter.

IMPLICATIONS OF THE FINDINGS

Although discovery of enabling factors is helpful in understanding why information transfer

problems exist, it is desirable t o examine in more detail when, how, and why these factors appar-
ently disrupt information flow. The more one knows about the dynamics of these unwanted occur-
rences, the more likely it is that strategies can be devised to deal with them.

The Pivotal Role of the Air Traffic Controller

Figure 7.1 is a schematic representation of the primary operational information transfer inter-
faces in today’s aviation system. (This schema does not depict intracockpit communications, or
secondary air-ground communications. The interfaces shown are those that bear directly on aircraft
management in the airspace system.) Because of the central role of the air traffic controller, any
study of infomation transfer problems must of necessity be largely a study of controller communi-
cation behavior. Can we gain any insights into such behavior from these data?

AIR TRAFFIC
MANAGEMENT

OTHER
SECTORS -
-m ATC t- OTHER
CONTROLLER --W CONTROLLERS

It
FLIGHTCREW

Figure 7.1 .- The role of the controller in information transfer.

Many of the reports analyzed here described occurrences in which ASRS analysts believed
that communications that did not take place would have averted the hazard; about the same num-
ber of reports described a variety of problems associated with information transfer transactions that
did take place. The reasons for these problems may well be different but they are equally important
t o those who may wish to intervene effectively.

Information Transfer Problems Affecting the Controller

The data indicate that there are several kinds of information transfer failures that affect the
controller’s ability t o perform his job and, specifically, t o ensure that separation is maintained.

87
 Failure to perceive- The controller may be unaware of an aircraft of concern because the air-
craft is not shown on his radarscope (most often because it is a primary target) and the pilot has
not communicated with ATC. Cunningham (ref. 1 ) has discussed the problem of unknown aircraft
in highdensity terminal airspace ; in previous studies, we have pointed out some behavioral aspects
of this problem (ref. 2). Local controllers may be unable t o perceive an aircraft in their airport
traffic area because of restricted visibility.

In both of these situations, the problem is simply that a pilot, for whatever reasons, has not
drawn ATC’s attention to his presence, either by use of a transponder in areas of radar coverage, or
by voice communication. ASRS data make it plain that this class of information transfer failure
seriously affects the functional effectiveness of the air traffic control system. It is involved in a
substantial fraction of the large number of conflicts reported.

The controller may also fail to perceive a visible aircraft, or a conflict, because he is distracted
by other operational o r nonoperational concerns. Absent or incomplete information transfer trans-
~

actions that occurred earlier (relief briefing or coordination) may fail t o bring t o a controller’s
attention an aircraft of potential concern to him. Finally, the controller may perceive an aircraft,
then forget about it because it was not of concern at the time it was perceived. These human fac-
tors - distraction, failures to monitor, and forgetting - are described over and over again in reports
from con trollers.

Mispercep tion- There are several factors associated with misperceptions and misunderstand-
ings by a controller. He may quick-look an adjacent sector and fail t o notice all pertinent traffic.
Tag swaps, scope clutter, and other hardware o r software anomalies may gwe him a false picture of
his traffic. Transponders without altitude reporting (Mode A), inaccurate briefing or coordination
messages or communications may leave him with a correct perception of an aircraft’s present posi-
tion but an incorrect perception of the aircraft’s flightpath; this may lead him t o assume the air-
craft’s future course of action. Expectation plays a part in the formation of these assumptions. If
the aircraft thereafter heads for an incorrect runway, as an example, the controller may or may not
notice the deviation.

I
The Controller Role in the Genesis of Information Transfer Problems

It goes without saying that the controller’s decisions (and therefore his subsequent instruc-
tions) can be n o better than his information. If the information is absent, inadequate, or incorrect,
his decisions will be incorrect, though they may or may not cause a subsequent hazard. In many
cases, however, the information reaching the controller may be correct; thereafter, he can become
involved in the creation of a subsequent problem by an error. There are three types mentioned
prominently in these data.

No information trunsfer- A substantial fraction of ASRS pilot reports of conflicts castigates a

controller for not providing a timely traffic advisory. The credibility of controllers is not enhanced
by the frequency with which, in response to a question from the flightcrew, they report that “It
just appeared” at the six o’clock position. Many pilots realize that the provision of traffic advisories
is an “additional service”; few have much sympathy for that dictum if they have just had a near
midair collision. Pilots usually have n o way of determining whether traffic was not announced to
them because the controller did not know about it, because he was busy with another task, or
because he failed to perceive the conflict.

Untimely information transfer- Many of these reports also involved traffic advisories. Again,
the pilot was unable to tell whether the target had just become visible, or whether it had been
visible and the controller failed t o notice it. The tenor of reports from controllers suggests that the
former is more often the case, though a certain proportion of reports from controllers describes
situations in which they perceive a potential future conflict, defer action in order t o get better
(more) information, then become distracted by other more immediate tasks and forget the pre-
viously noted target. As noted in chapter 3, forgetting is also a problem with respect t o coordina-
tion transactions. Other occurrences falling in this category are instructions given at operationally
untimely moments, such as a revised clearance at the time of touchdown or early in rollout, or a
new instruction just after liftoff.

Incomplete information trunsfer- It is our impression that incomplete messages are usually a
result of pacing stress, less commonly a result of partially blocked transmissions. The latter failures,
however, are sometimes serious, especially when they involve clearances. It must be recognized that
an apparently incomplete message may result either from an omission by its originator or from
incomplete perception by a receiver. The message may, or may not, be recognized by either party
as incomplete.

Inaccurate information framfer- Again, misperception by either the originator or receiver of a

message may be responsible for inaccuracy. Chapter 4 points out that this category of message prob-
lem either could involve a formulation e n o r or could be a correctly formulated statement based on
an inaccurate decision. The effect on the receiver is the same, but the problems may be quite
different. The ASRS data indicate that there are several classes of events in the latter category. They
are summarized in figure 7.2. Incorrect decisions may proceed from either correct o r incorrect data;
this schema characterizes only the decision process.

The various decision options for each perception (whether correct or incorrect) are indicated
by the X’s in each row of the figure. These can be roughly bounded by curves.’ The shaded area
represents the realm of perceptions for which more than one decision strategy is available. One may
operate extremely conservatively, near the upper curve, or one may opt for a more venturesome
decision strategy, near the lower curve.

Study of the ASRS data suggests that the shaded portion of the figure contains the circum-
stances of many of the inaccurate decisions mentioned by controllers in aviation safety reports. If
one excludes reports in which the conflict was misperceived, this part of the figure characterizes a
good deal of the data.

The Role of the Pilot

The role of the pilot receiving ATC services in information management is somewhat different
from that of the controller in the present aviation system. His task, except in an emergency, is to
receive advisory information, accept instructions, and t o act upon them. He provides an element of
redundancy by reading back clearances, announcing altitude on initial callup, etc., but otherwise he
provides little information unless it is asked for.

89
 ~

DECISION OPTION

DELAY
PERCEPTION TAKE ACTION TAKE PRE- TAKE PRE- TAKE
NO TO GET MORE CAUTIONARY CAUTIONARY REMEDIAL
KTION INFOR- ACTION LATER ACTION NOW ACTION NOW c

MATION

NO CONFLICT

POSSI BLE
FUTURE
CONFLICT

PROBABLE
FUTURE
CONFLICT

LIKELY
CON FLlCT

IMMINENT CON-
FLICT, PROBABLE
LOSS OF
SEPARATION
LESS THAN
STANDARD
SEPARATION

CRITICAL
CONFLICT
~~

Figure 7 -2.- Controller decision processes.

As was indicated above, the pilot could usefully d o more, especially when flying under visual
flight rules. The provision by the ATC System of separation t o participating aircraft depends on
ATC’s knowledge of all traffic that could come into conflict with system participants. The acquisi-
tion by ATC of this information depends either on information supplied by Mode C transponders
or on the same information provided orally by pilots. The provision of position, altitude, and inten-
tions by VFR pilots (regardless of whether they intend to utilize ATC services) could do much to
improve the quality of the services provided by controllers, especially in terminal areas.

Pilots often lack information with which t o evaluate controller decisions and instructions.
They are therefore unable t o question such decisions intelligently (even if the decisions are wrong)
unless the decisions are obviously inappropriate. This places a heavy burden on the controller, who
in this respect is unprotected -by the redundancy so carefully designed into most aspects of the

90
aviation system. It is to the credit of controllers that they so frequently detect errors made by
pilots; it is equally to the credit of pilots that they frequently detect controller errors, notwith-
standing their lack of access t o much information available in the ATC system.

The Role of Automation

Studies by ASRS investigators and others (ref. 3) have made it clear that automation, whether
in the cockpit or in ATC, is associated with costs as well as benefits with respect t o human perfor-
mance. ASRS reports describe a spectrum of behaviors, from unwarranted suspicion of, to over-
reliance upon, automated devices.

Automation is used somewhat differently in the cockpit and in today's air traffic control sys-
.
tem. The transfer of much of the routine information upon which t h e air traffic management sys-
tem relies is performed automatically; most routine handoffs are likewise automatic. More recently,
automatic alerting and warning systems (conflict alert, minimum safe altitude warning) have been
implemented in system software. In controller reports as well as in those from pilots, one finds a
broad range of views on automation, from suspicion based on warnings that seemed inappropriate
(or warnings that did not occur when they were needed) t o overdependence on automated
functions.

The most critical reports, however, are those describing failures of the automated system to
transfer or supply information. These reports suggest that at least some controllers using automated
equipment have come, over time, to rely heavily upon its continued full functioning (see ch. 4).
They are seriously taken aback when it fails t o perform to their full expectations. We have indicated
above that ASRS reports give evidence of a tendency t o delay taking positive action to separate air-
craft when a controller perceives that action may or may not be needed (see fig. 7.2). The particular
problem with delayed decisionmaking behavior, as adduced from ASRS reports and experimental
evidence (Bisseret, personal communication, 1981) is that it assumes that there will be a continuing
supply of information necessary for a later decision; it also assumes that the controller will remem-
bcr t o return to and resolve the problem. Such behavior is clearly motivated by overreliance upon
an extremely reliable system; it is nonetheless a potential point of weakness in the system.

Behavioral Impediments t o Information Transfer

The study of Foushee and Manos (ch. 5 ) suggests that certain categories of behavior act as
impediments to information flow in the cockpit. More important, their study shows that there is a
significant association between information transfer and flightcrew performance. To summarize
their observations, it was found that numbers of commands, acknowledgments, observations, and
comments indicative of agreement were all positively associated with performance. Communications
indicating uncertainty, frustration or anger, and embarrassment were negatively associated with
performance.

These findings are consonant with the central thesis of this report: that inadequate informa-
tion transfer is associated with errors in aviation operations. The findings are important because
they also indicate that there are behaviors that interfere with information transfer, just as there are

91
 system factors (e.g., frequency saturation, high workload, and inadequately presented data) that
interfere with the free flow of information.

Enhancement of Information Transfer

I
Many steps have been taken in recent years t o enhance the flow of information in the aviation
system; others, especially the implementation of data link, are planned for the near future, but it
will be some time before these new developments begin to have an appreciable effect.

I These and previous studies lead us to conclude that there is a real and present need for better
information transfer between general aviation aircraft flying in or at the periphery of radar-
controlled terminal areas and the air traffic system. We believe that the most effective way of
accomplishing this information transfer is by use of Mode C transponders, although an alternative
method is by voice communication between pilots and controllers. The problem with the latter
method is that it does not provide the controller with a continuously visible target of adequate
quality. We believe that many general aviation pilots are unaware of the degree of protection this
step can provide them in terms of separation from participating traffic, and that the advantages t o
them of Mode C transponders have not been made sufficiently clear.

We conclude that there is a need for wider and more disciplined use of “scratch-pad’’ notes
and other memory aids by air traffic controllers. We urge that designers of the future ATC system
give thought to the desirability of incorporating memory aids into the control and display inter-
faces of that system.

We conclude that distraction is a serious problem that inhibits effective performance, both in
the cockpit and in air traffic control. We note that FAA has introduced distraction into its general
aviation flight tests, and we suggest that it might be helpful t o consider ways of helping controllers
deal with both operational and nonoperational (e.g., noise) distractions as well.

We conclude that failure to decide upon and relay information concerning precautionary
action when potential conflicts are first noticed is one of the factors contributing t o controller
errors reported to ASRS. This behavior contributes to failures to ensure that proper separation is
maintained. The use of memory aids can help to keep such conflicts from being forgotten, but the
mental set of the controller with respect to the resolution of potential conflicts is probably equally
important. We suggest that further study of decisionmaking delays may be usefuI in clarifying the
I reasons for this phenomenon.

Complacency is not widespread in reports of information transfer failures, but it can be a

serious problem when it is present. We urge controllers and pilots alike to be aware of its insidious
nature and of the marked decrements in perceptivity and thinking that it can cause. The system
does not always work, and the only protection against potential disaster when it fails is alertness
and forethought regarding alternative ways of accomplishing the objective.

We conclude that there is insufficient awareness of the pervasive nature o f the information
transfer problem in its various manifestations, and that this lack of awareness may be in part
responsible for nonstandard or inadequate communications practices on the part of both controllers

92
and pilots. We believe that wider dissemination of information concerning this problem may help
to alleviate it to some extent. This report, it is hoped, may assist in resolving that most critical
information transfer problem.

Ames Research Center

National Aeronautics and Space Administration
Moffett Field, California 94035, March, 1981

93
 REFERENCES

1. Cunningham, F. L.: The Midair Collision Potential. Proceedings of the 24th Annual Meeting of the Air Traffic
Control Association, Atlantic City, N.J., 1980.

2 . Billings, C. E.; Crayson, R. L.; Hecht, A. W.;and Curry, R. E.: A Study of Near Midair Collisions in US.Termi-
nal Air Space. ASRS Quarterly Report No. 11, NASA TM-81225,1980.

3. Wiener, E. L.; and Curry, R. E.: Flight Deck Automation: Promises and Problems. NASA TM-81206, 1980.

94