Appendix A—Overview Part 4 of 4

Decision Workshop ment for greater speed escaping fires. The session begins with four presenta- More than 30 people are entrapped each tions that outline the psychological and Improving Wildland Firefighter Perform- year. Our crews are not as proficient at organizational aspects of wildland ance Under Stressful, Risky Conditions: escape, fire shelter deployment, and firefighting. These talks will set the tone Toward Better Decisions on the Fireline decisionmaking under stressful, risky for the session as we focus on the and More Resilient Organizations conditions as they could or should be. individual firefighter, fire crews, and Partly, this reflects attitudes, and partly organizational structures on the fireline. June 12–16, 1995 it is a lack of knowledge. We hope the unique mix of professionals Village Red Lion Inn will create a synergism that leads to 100 Madison Street This workshop will explore firefighter meaningful change and a safer firefight- Missoula, Montana psychology, interactions among fire- ing environment. fighters and among fire crews, and better ways to organize. To do this, we have Overview— brought together experts in psychology, sociology, organizations, fire safety, and Ted Putnam It has become increasingly clear since wildland firefighting. We will be looking Workshop Organizer 1990 that wildland firefighters are exper- at the current situation on the fireline and iencing collapses in decisionmaking ways to make the often-dangerous job Dave Thomas and organizational structure. Wildland of wildland firefighting safer. Finally, we Workshop Moderator fire agencies have lost 23 people since will develop a series of recommenda- 1990 who might have survived had they tions for implementing the changes Jerry Meyer simply dropped their tools and equip- needed to improve firefighter safety. Workshop Facilitator

23 Appendix B—Agenda Part 4 of 4

0900–1000 1100–1200 MONDAY, JUNE 12 Recognition Primed Decision South Canyon Revisited: Lessons 0800–1200 Strategies from High Reliability Gary Klein, Ph.D., Klein Associates Organizations Karl Weick, Ph.D., University of Keynote Presentations Studies of firefighters show that they Michigan (open to the public) rely primarily on Recognition-Primed Decision (RPD) strategies as opposed There is an emerging body of work that 0800–0900 to sifting through alternatives and has begun to describe how Behavioral Bases of Accidents and comparing how they rate on different organizations that face the possibility of Incidents: Identifying the Common evaluation dimensions. The RPD model catastrophic error every day, cope with Element in Accidents and Incidents explains how people can make this prospect. These organizations, Curt Braun, Ph.D., University of Idaho decisions under conditions of time referred to as high reliability pressure and ambiguity and shifting organizations, include nuclear power Human behavior plays the largest role conditions. Specifically, the model plants, air traffic control systems, in firefighter safety. Faced with the explains how experienced aircraft carriers, flight crews, and known and essentially constant risks of decisionmakers can generate a chemical plants. Several issues that are a wildland fire, human behavior is the reasonable course of action without discussed in studies of these only factor that can greatly increase or having to contrast alternatives, and how organizations are similar to issues that decrease the risk of injury. Despite its they can evaluate a course of action have surfaced in discussions of the large role in safety, however, behavior without comparing it to alternatives. South Canyon Fire incident on July 6, is frequently overlooked during accident 1994. The purpose of this presentation investigations. Endeavoring to discover We can also use the RPD model to will be to discuss some of these the cause of a workplace injury or understand some of the errors that can similarities with special attention being fatality, investigators often focus on the arise in naturalistic settings. These focused on issues of communication, special environmental circumstances primarily stem from inadequate group structure, stress, mindsets, and not on the behaviors that preceded experience bases. In turn, these error leadership, and sense making. that accident. This overemphasis on types suggest some strategies for circumstances fails to consider the fact decision-centered training. that the vast majority of accidents result 1200–1300: Lunch not from the environment but from Closed Workshop Session Begins known risky behavior that is part of 1000–1100 (limited to invited participants) common work practices. The “Cultural Inertia” Impacts of Team Decisionmaking 1300–1700: Focus will be on the To address safety challenges, specific David Hart, TIG, Inc. dynamics of individual decisionmaking consideration must be given to the and individual firefighter experience element common to all accidents: Cultural attitudes permeate the human behavior. A behaviorally based decisionmaking of teams working within safety program can reduce the risks the organization. “Anytime, Anywhere,” 1830–: No-host bar and dinner (if there associated with wildland fire “Can Do,” and “Make It Happen” are is sufficient interest) suppression by: (1) identifying examples of adopted cultural attitudes antecedent behaviors that lead to that have both assisted and (in some accidents; (2) determining the cases) inhibited crew effectiveness. TUESDAY, JUNE 13 frequency of these behaviors; (3) This discussion investigates the 0800–1200: Focus on interaction evaluating the training programs and impacts of cultural-based attitudes as between crew members management systems that either barriers to individual and crew directly or indirectly support the decisionmaking processes. It includes behaviors; and (4) developing a training discussion of lessons learned from 1200–1300: Lunch and management remediation program other high-risk/high-threat aimed at changing behavior. environments in establishing a non- 1300–1700: Focus will be on attribution/non-retribution environment, interactions between crews and and overturning cultural attitude organizational structure barriers within the individual and the organization as a whole.

24 Part 4 of 4

ride (round-trip). We will return to the WEDNESDAY, JUNE 14 marina around 1730-1800 and will be FRIDAY, JUNE 16 Trip To Mann Gulch back in Missoula by about 2000. Each 0800–1700: Workshop volunteers participant should bring a sack lunch finish write-ups on findings and and a canteen of water as well as recommendations 0545: Assemble at Village Red Lion Inn snacks for the return trip. In addition, and board bus bring clothing appropriate for the weather, as well as other items desired 0600: Depart Missoula such as cameras, daypacks, and so forth. Suggested Reading: 0600–0800: Workshop discussions enroute We plan to have two guides Young Men and Fire. Norman Maclean. knowledgeable about Mann Gulch 1992. University of Chicago Press. 0830–1800: Mann Gulch guided tour available to retrace the events. An EMT Chicago, IL. with a first aid kit and radio will also On August 5, 1949, a wildfire overran accompany us. Fireline: Summer Battles of the West. 16 firefighters at Mann Gulch. There Michael Thoele. 1995. Fulcrum were only three survivors. Significant The walk up Mann Gulch is about 1-1/2 Publishing, Golden, CO. controversy has surrounded this fire, miles over grassy, rocky ground. Sturdy including firefighter decisions and work or hiking boots (well broken in) Mann Gulch Fire: A Race That Couldn’t actions as well as the ensuing are strongly recommended. The slope be Won. Richard Rothermel. 1993. INT- entrapment investigation. On July 6, is steep, but people of varied fitness GTR -299. U.S. Department of Agricul- 1994, a wildfire overran 49 firefighters levels have tackled it successfully ture, Forest Service, Intermountain at South Canyon. There were 35 walking at their own pace. Snakes and Research Station, Ogden, UT. survivors and 14 fatalities. Many footing are the only other hazards, investigators believe the two events are though snakes are rare. The trip will be connected and ask how much have we canceled if rain or strong winds are learned in the intervening 45 years. A forecast. More information will be *Mann Gulch Fire: A Race That Couldn’t trip to Mann Gulch has been planned to provided at the workshop. Be Won. Richard C. Rothermel. explore that connection and foster further workshop dialogue in a “real life” *Available on the worldwide web at: setting. 1800–2000: Return trip to Missoula http://www.xmission.com/~int/pubs.html

Mann Gulch is about 150 miles east of For a printed copy, contact: Missoula and 25 miles north of Helena, THURSDAY, JUNE 15 Montana. A bus will leave the Village 0800–1200: Focus on future Tom Cook, Program Assistant Red Lion Inn parking lot promptly at research, investigations, and training Fire Behavior Project 0600. It is about a 2-1/2 hour drive to Intermountain Fire Sciences Laboratory the Gates of the Mountain Marina and 5775 Highway 10 West about a 30-minute boat ride down the 1200–1300: Lunch Missoula, MT 59802 Missouri River to Mann Gulch. We are Phone: (406) 329-4820 paying for the bus, but each participant 1300–1700: Workshop Fax: (406) 329-4825 will need to pay about $10 for the boat recommendations DG: T.Cook:S22L01A

25 Appendix C—Participants Part 4 of 4

Participants Management Review Team following National Forests. Because of his the South Canyon Fire and serves on extensive wildland fire suppression Dave Aldrich—branch chief for the Federal Wildland Fire Policy and experience throughout the U.S., Paul ground operations safety, Forest Service Program Review. He has an undergrad- has been active in fire suppression/fire Washington Office, Fire and Aviation uate degree in political science from behavior course development and Management. Dave began his Forest Grinnell College and a master’s degree presentation. He has an undergraduate Service career as a seasonal employee in public policy from the University of degree in mathematics from Colorado in 1958 on the Powell District in R-1. He Michigan. State University and is pursuing gradu- has worked in R-1 and R-3 in fire man- ate studies in fire ecology and effects at agement jobs as well as the National Colorado State University. Office and at the Intermountain Fire Jon Driessen, Ph.D.—is a professor Sciences Laboratory in Missoula. He has of sociology at the University of Mon- been a fire behavior analyst on national tana. He also holds a faculty affiliate Dave O. Hart—has extensive experi- fire teams and has been involved with appointment at the Missoula Technology ence as in instructor and facilitator in the several national fire training courses. and Development Center. He specializes delivery of crew resource management Dave chairs the NWCG Safety and in the sociology of work and for the past training. He is a former Air Force B-52 Health Working Team. He has a BS in 12 years has studied the culture of work instructor navigator, and instructor at forestry from the University of Montana. in Forest Service field crews. His latest the Air Force’s Undergraduate Navigator project for the Forest Service is a 48- Training. He served as lead facilitator minute video, Making a Crew. Jon has for Hernandez Engineering, the crew Bill Bradshaw—works for the Forest a doctorate in sociology from the resource management contractor for the Service Washington Office in fiscal and University of Colorado. Air Force Air Mobility and Air Combat accounting, specializing in incident Commands. He is an Air Force Reserve administration and claims. Bill has been C-130 navigator for the 731st Airlift active with decision analysis projects in Kelly Esterbrook—is currently a Squadron, which provides wildland the past and is currently involved with smokejumper squadleader with the firefighting support through the MAFFS efforts to enhance wildland firefighter Forest Service, in Redmond, OR. Kelly (Modular Airborne Fire Fighting System) safety through improved attitudes, started her Forest Service career on the program. Dave cofounded TIG Inc., leadership, and responsibility. Rogue River National Forest in 1978. where he works as a training consultant She spent 2 years on Rogue River and facilitator. He is currently responsible engine crews and 2 years as a Rogue for assessment, design, development, Curt Braun, Ph.D.—is an assistant River Hotshot crewmember. She then and delivery tasks associated with the professor of psychology at the University spent four seasons on the Deschutes new Army National Guard Special of Idaho. As an ex-firefighter for the National Forest as an engine foreman Forces Decision Training Program. He Sawtooth National Forest, he evaluates and one season with the Redmond received his bachelor’s degree in aircraft fire suppression from the firefighter’s Hotshots. She began jumping in 1986. maintenance engineering from Parks perspective. He recently coauthored She was detailed to the Union Hotshots College and is pursuing a master’s Human Decisionmaking in the Fire in 1992 as superintendent. In 1994 she degree in aerospace studies from Environment, which will appear in an completed Technical Fire Management Embry-Riddle Aeronautical University. upcoming special issue of Fire Manage- Training. ment Notes. He holds a Ph.D. in human factors psychology with an emphasis on Jerry Jeffries—has spent his entire human performance from the University Paul Gleason—is currently fire ecolo- Forest Service career in fire and safety. of Central Florida. gist for the Roosevelt and Arapaho He recently was named safety project National Forests in Northern Colorado. leader at the Missoula Technology and His emphasis is the restoration and use Development Center. From 1990 to Jim Douglas—is Director of the Office of fire as a natural process to achieve 1995 he served as safety and health of Hazard and Fire Programs Coordina- land management goals in the central manager for aviation and fire manage- tion for the U.S. Department of the Rocky Mountain ecosystems. From 1991 ment in R-1. Before that time, he was Interior. Before coming to that post, he to 1994, he was fire management officer for many years safety manager on the served as the Interior Department’s fire for the Estes-Poudre and Redfeather Bitterroot National Forest. He has held a program coordinator. His career with Ranger Districts in Northern Colorado. variety of fire positions during his career, Interior began in 1979 in the Office of Prior to that time Paul spent 23 fire including interregional hotshot crew Policy Analysis. He was also in the seasons with the Interagency Hotshot supervisor, division supervisor, line boss, Department’s Office of Budget for 7 Crew programs on the Angeles, Mt. air attack boss, and air support group years. He was on the Interagency Hood, Okanogan, Pike and San Isabel supervisor. In 1992 he received the 26 Part 4 of 4

Government Employees Insurance Co. officer on the Ochoco National Forest. He Robert J. Martin—is the Forest (GEICO) public service award for fire holds a bachelor’s degree in forestry/fire Service national aviation safety and prevention and safety from a group of management from the University of training manager at the National Intera- over 200 nominees worldwide. Montana. gency Fire Center, Boise, ID. Bob’s aviation experience covers the fields of maintenance, accident investigation, Jim Kautz—is a videographer/photog- Mark Linane—is the Los Padres Hot- piloting, and program management. For rapher at the Missoula Technology and shot superintendent on the Los Padres the past 30 years he has been employed Development Center. He began his fire- National Forest in Region 5. The crew in the military, commercial, and public fighting career at Darby Ranger District is located at the Santa Barbara Ranger sectors of aviation. His Forest Service and was a smokejumper for 3 years in District office compound 10 miles north career began in 1977 in R-3. Since that R-1. For the past few years one of Jim’s of Santa Barbara, CA. Mark has 30 time, he has served at national fire primary responsibilities has been to years of wildland fire experience, the last center and R-6, Portland, OR. During provide photo documentation as part of 23 as superintendent of the Los Padres 1987-1988, he worked with U.S. wildland fire entrapment investigations. Hotshots. He is considered a leading Customs air interdiction program and Jim holds a degree in film and television spokesperson for the hotshot commun- returned to the Forest Service in 1989. from Montana State University. ity. He has been involved with safety and Bob received his BA in aviation manage- training issues for years, most recently ment from Boise State University and working on the revision of the Strike his MA in aviation management from Gary A. Klein, Ph.D.—is chairman Team/Task Force Leader training Embry-Riddle Aeronautical University. and chief scientist of Klein Associates. course. He has performed research on natural- istic decisionmaking in a wide variety of Jerry Meyer—has worked for the task domains and settings, and has Lark S. McDonald—has performed Forest Service since 1971 in a number developed significant new models of assessment, development, and design of capacities, primarily in timber man- proficient decisionmaking. His research work in human factors training programs agement. He has also worked as a interests include the study of individual for a wide variety of aviation-based firefighter, wilderness guard, writer/editor, and team decisionmaking under condi- applications, including aeronautical and historian. His most advanced red- tions of stress, time pressure, and decisionmaking and cockpit manage- card qualification is division/group uncertainty. He has furthered the devel- ment for civilian pilots. He has served supervisor, but he most often takes field opment and application of a decision- as designer and developer for crew observer assignments. Jerry will facili- centered approach to system design and resource management programs for the tate the workshop discussions. He holds training programs. He has also studied Navy T-45, Air Force T-1, and commer- a BA in history/political science from the applications of case-based reasoning for cial MD-80 for McDonnell Douglas Corp. University of Montana. domains such as the cost/benefit evalu- He has worked as a development ation of training devices and developing program manager for United Airlines and marketing projections for new products. Martin Marietta, and as the lead instruc- Dave Pierce—is currently the smoke- He holds a doctorate in experimental tional designer for Hernandez jumper project leader at the Missoula psychology from the University of Engineering, the crew resource manage- Technology and Development Center, Pittsburgh. ment contractor for the Air Force Air a position he has held since 1980. His Mobility and Air Combat Commands. His Forest Service career began in 1964 as recent work has included assessment a “smokechaser” on Red River Ranger Buck Latapie—is currently the fire and adaptation of CRM training for use District in Idaho. From 1965 through training and safety officer for R-6. He with Air Force test pilots and their 1968, he was a smokejumper in R-6 and has served continuously on incident ground-based engineer and logistic R-1. Between 1969 and 1971, he worked management teams since 1983 in plans, counterparts. In a further extension of in private industry as a commercial pilot. operations, and as an incident com- moving CRM-type training into high-risk, In 1971 he returned to firefighting as a mander. Early in his Forest Service high-stress environments, he recently smokejumper with the BLM. With 30 career he served as a hotshot foreman cofounded TIG Inc. with David Hart, years of experience working with both and engine foreman. In 1978 he was which currently provides decisional train- the Forest Service and BLM smokejump- hotshot superintendent on the Bitterroot ing and leadership programs for teams ing programs, Dave has accumulated National Forest. He later worked as a with the National Guard Special Forces. some “street smarts” about initial attack silviculturist on the Fremont National He received his education in aviation firefighters. During his years with the Forest, district fire management officer management and psychology from BLM Alaska smokejumping program, he on the Deschutes National Forest, and Metropolitan State College, Denver, CO. was responsible for organizing smoke- as a forest aviation and fire management jumper crews for safety and effective- 27 Part 4 of 4 ness. At MTDC, he has finished several present position in the Washington David A. Thomas—is fire manage- projects related to safety in smoke- Office. He has taught several classes on ment officer on the Superior Ranger jumping/aviation operations where the various aspects of fire and risk manage- District, Lolo National Forest. He started objective was to develop materials ment at the University of Idaho and at his Forest Service career as an emer- designed to change institutional attitudes. the district, forest, regional, and national gency firefighter in 1967. He has been levels of the Forest Service. He is the a member of fire crews on the Kootenai unit leader for the Risk Management and and Clearwater National Forests. Later, Ted Putnam—is a fire and safety Decision Analysis unit of the National he was helicopter foreman of an 18- equipment specialist at the Missoula Interagency Prescribed Fire Behavior person crew. Dave has been a member Technology and Development Center. Analyst course taught at the National of numerous type I and type II incident He started working for the Forest Advanced Resource Technology Center management teams. He was a fire Service in 1963 and spent 3 years on in Marana, AZ. The Interagency Man- behavior analyst on the 1988 fires in district fire crews, 8 years as an R-1 agement Review Team for the South Yellowstone National Park. As a smokejumper, and 3 years as a super- Canyon Fire asked Jim to lead a team to prescribed fire manager, Dave has devel- visory smokejumper. In 1976 he came to develop a report on the collection, distri- oped and implemented many prescribed MTDC. He is responsible for developing bution, and utility of live fuel moisture burns ignited for various silvicultural and firefighter’s protective clothing and fire information. Jim has a BS in mathe- fuels management objectives. Dave shelters, including training materials. He matics from Auburn University, and an holds a BA in geography from the is a member of two National Fire Protec- MS in forest resources and a Ph.D. in University of Montana. tion Association standards-setting forestry, wildlife, and range sciences committees for protective clothing and from the University of Idaho. His Ph.D. equipment. Ted holds a Ph.D. in experi- work concentrated on the application of Karl E. Weick, Ph.D.—is the Rensis mental psychology from the University artificial intelligence, decision science, Likert Collegiate Professor of Organiza- of Montana. and cognitive psychology to fire tional Behavior and Psychology at the management. University of Michigan. He is also the former editor of Administrative Science Jim Saveland—is a fire ecologist for Quarterly, the leading research journal the Forest Fire and Atmospheric Lyle Shook—is currently safety and in the field of organizational studies. He Sciences Research Staff, Washington, health manager for R-5. He has 21 years studies such topics as how people make D.C. He began his Forest Service career of experience in Forest Service wildland sense of confusing events, the social in 1978 on a district fire crew in Elk City, fire operations in Regions 3, 5, and 6. psychology of improvisation, high relia- ID. Jim spent 4 years as a smokejumper His experience ranges from hotshot and bility systems, the effects of stress on in R-6 and R-1. In 1984 he became fire helitack crews to acting Regional fire thinking and imagination, indeterminacy management officer on the Moose Creek coordinator. In the incident management in social systems, social commitment, Ranger District. In the incident command system he is a type I operations chief, small wins as the embodiment of system, Jim was a division/group super- plans chief, and safety officer. He has wisdom, and linkages between theory visor and a fire behavior analyst. In 1988 been a type II incident commander for 3 and practice. His writing about these he transferred to the Southern Fire Lab years. He has been in his current topics is collected in four books, one of in Macon, GA. Jim became project position since 1988. which—The Social Psychology of leader in 1991. In 1994, he moved to his Organizing—is cited as furnishing

28 Part 4 of 4 significant background for Peter’s and sociologist he spent 1 year conducting Tim Crawford, Gates of the Mountains Waterman’s In Search of Excellence. baseline social data collection and social Marina, by Helena, MT Karl has consulted with a variety of impact analysis for landscape analyses organizations, including Corning Glass, and environmental impact statements. Mary Jo Lommen, Mann Gulch EMT, Narco, Cole Products, Dalton Foundries, He has been a college instructor for 5 Superior Ranger District, Lolo National Southland Corp., Motorola, Texas Instru- years. Patrick has a Ph.D. in sociology Forest, Superior, MT ments, Lockheed, the National Science and an MBA from Boston College. He Foundation, the National Institute of also holds a BS in psychology from the Dave Turner, Mann Gulch interpreter, Education, and the National Institute of University of Oregon. Helena Ranger District, Helena National Mental Health. He has a Ph.D. in Forest, Helena, MT psychology from Ohio State University. Steve Wolf—is a research associate The following people were invited to the at Klein Associates. He has played a key workshop but were unable to attend: Pat Wilson—is manager of the technical role on projects concerned with Grangeville smokejumper base, a posi- expert knowledge and decision support. Paul Broyles, training and safety tion he has held since 1987. He started He was the project leader on a recently specialist, National Park Service, his firefighting career in 1974 on an completed effort sponsored by the Navy NIFC, Boise, ID engine crew with the Idaho Department to develop a decision support system for Jim Cook, superintendent, Arrowhead of Lands. He became engine crew crew members in a combat information Hotshots and training specialist, foreman in 1976. In 1978 he was an center. He heads a related effort to National Park Service, NIFC, Boise, ID assistant foreman of the now-defunct examine potential training applications. Mary Jo Lavin, director, Fire and Coeur d’Alene Hotshots. The next His current projects include a review of Aviation Management, USDA Forest season he joined the St. Joe Hotshots National Fire Academy curriculum Service, Washington, D.C. as a lead sawyer. He began smokejump- designed to enhance rapid decisionmak- Holly Maloney, squadleader, Lolo ing in 1980 in Missoula, transferring to ing on the fireground. He has been a Hotshots, Lolo National Forest, Grangeville in 1981. He became a squad member of a technical team studying Missoula, MT leader in 1983. He served for 2 years on helicopter pilot safety, allocation methods Stan Palmer, safety and health group the forest safety committee and currently used by fire direction officers, and review leader, Bureau of Land Management, is a member of the National Aerial of human-computer interface designs for NIFC, Boise, ID Delivered Firefighter Study, and a group a surveillance platform developed jointly Gina Papke, superintendent, Zig Zag that is rewriting the Smokejumper and by the Army and Air Force. He holds a Hotshots, Mount Hood National Forest, Paracargo Handbook. BS in psychology from Wright State Zig Zag, OR University, Dayton, OH. Bill Russell, acting director, Aviation and Fire Management, Region 3, Patrick Withen—a smokejumper Special thanks to these people for their Albuquerque, NM based in McCall, ID, he is assistant assistance during the workshop: Jerry Williams, director, fire operations, professor of sociology at Centenary Fire and Aviation Management, USDA College, Shreveport, LA. His fire experi- Laurel Chambers, workshop note- Forest Service, Washington, D.C. ence includes 14 seasons as a smoke- taker, Superior Ranger District, Lolo jumper, 1 season on a hotshot crew, and National Forest, Superior, MT 2 seasons on a helitack crew. As a forest

29 Appendix D—Keynote Presentations Part 4 of 4

Addressing the snag will eventually fall, but not when. spectrum of situations where the same Common Behavioral If the environmental hazard remains behavior can also result in an accident. essentially constant, only one Element in Accidents component is left to vary: the actions of Krause and Russell (1994) suggest that and Incidents the human. accidents result, not from unique Curt C. Braun, Department of Psychology, circumstances or behaviors, but from University of Idaho The level of risk created by the snag can the intentional display of risky behaviors be mitigated or exacerbated by the that occur with such regularity that they Virtually every college student has behavior of the individual. Injury and have become common practice. These faced the philosophical question, “If a loss are more likely when the individual authors contend that an accident tree falls in the woods and no one is fails to attend to the known risks. When represents an unexpected result of an there to hear it, does it make a sound?” the individual is struck by the falling unsafe act that has become part of the The answer of course is no; the falling snag, the proximate cause is apparent, working culture. Despite the best efforts tree does not make a sound. While inattentiveness. It is not apparent, to mandate safety, risky behaviors many people struggle with this answer, however, that this was an isolated case increasingly become acceptable it is important to remember that the of inattentiveness. This inattentiveness practice each time they are performed answer relies, not on the physics might represent a general pattern of without negative consequences. The associated with a falling tree, but rather behavior that places the individual at process is similar to that seen in on the definition of sound. Sound is a risk in a variety of situations. To individuals who interact with hazardous subjective sensation created when the adequately respond to the accident, products. Safety researchers have ear is stimulated by changes in the consideration must be given to both the found an inverse relationship between surrounding air pressure. Given this proximate cause and the behavioral safety behavior and familiarity definition, a tree falling in the woods pattern. Unfortunately, traditional safety (Goldhaber & deTurck, 1988). The makes no sound when an ear is not programs have placed far more empha- probability that an individual will comply present. A comparable safety question sis on the former than on the latter. with safety guidelines decreases as might be, “If there is a snag in the woods familiarity with the product increases. and there is no one there, does it pose a risk?” Again, the answer would be no. Wildland firefighters are not immune to As with the sound example, the answer Human Behavior and this process. In response to the South centers not on the physics of a falling Accidents Canyon fire of 1994, Rhoades (1994) tree, but rather on the definition of risk, writes, “And sometimes, even often, the a chance of loss or injury to a human. Few will argue that most accidents and risks we take in doing our jobs include In the absence of a human, a falling mishaps are directly related to unsafe violating the 10 Standard Fire Fighting snag creates no threat of injury or loss. behaviors. A review of the national air Orders or ignoring the 18 Situations Although this relationship appears traffic control system revealed that 90% that Shout Watch Out.” He further obvious, it is important to realize that of the committed errors could be directly writes, “Nonetheless, very seldom does there are two components to this linked to human inattentiveness, poor our inability to comply with the orders question: the snag, and the presence or judgment, or poor communications cause us to abandon our tasks...” absence of the human. Both play a role (Danaher, 1980). Mansdorf (1993) lists Rhoades’ statements reflect the fact that in creating a risky situation. nine different causes of accidents and it is possible to violate standard safety attributes all of them to human error in practices without the worry of negative If an individual is injured by a falling the form of inadequate training, consequences. More importantly, how- snag, clearly both had to be present. supervision, and management. Given ever, Rhoades’ comments suggest that This situation can easily be represented this consensus, the solution is simple; the violations have occurred with such by the following model: change the behavior where the great regularity that they have become accidents occur. Despite the intuitive accepted practice in wildfire suppression. Environmental Hazard (Snag) + appeal of this approach, efforts to Human = Accident increase safety in this manner often fail to produce the anticipated reductions in The role of the snag and the individual accidents. These failures occur because Accident Prevention From in this situation are significantly different. traditional safety programs generally a Behavioral Perspective The fact that the snag will eventually focus on the unique circumstances and fall is well known and in contrast to the risks that, like the snag, remain relatively An effective prevention program begins actions of the human, represents a constant. Moreover, these programs by understanding that accidents often relative constant. We know that the often do not consider the broad reflect the unfortunate outcome of

30 Part 4 of 4 hazardous acts that have become misses can provide better insight into the-job-training (OJT) is a widely used common practices and that these employee safety. Mansdorf (1993) technique but it suffers from many practices frequently span a multitude of reports that for every serious industrial shortcomings. Trainers are frequently different job tasks. To be effective, a accident there are approximately 10 unaware of instructional techniques, safety program must: 1) identify the ante- minor accidents, 30 property damage training occurs only when time is made cedent behaviors that result in accidents accidents, and 600 near-miss accidents. available, the situation typically dictates and near-miss incidents; 2) determine what skills are learned, and trainees how frequently these behaviors occur; The overarching motivation driving a often take a passive role merely watching 3) evaluate training and management reporting system should be the acquisi- and not demonstrating behavior programs; 4) provide consistent and tion of reliable and valid data. To facili- (Gordon, 1994). Managers and super- active feedback and reinforcement, and tate this process, the reporting system visors must assess the extent to which 5) develop remediation plans. must encourage reporting from all levels training relies on OJT and develop of the work force. Moreover, individuals specific programs to maximize its Identifying Antecedent Behaviors. should be instructed as to their reporting usefulness. Traditional accident investigations tend responsibilities. With regard to the to be very myopic, focusing only on the logistics of the system, every reasonable After training, are the behaviors circumstances immediately involved in effort should be taken to reduce the cost practiced? Just as firefighters exercise the accident. The purpose of an investi- of complying with reporting requirements. to maintain a level of physical fitness, gation is to identify the accident’s cause These efforts might include simplifying skills learned in training must be with the aim of creating new procedures, reporting forms, the use on-site or tele- practiced to ensure competency. In a equipment, and standards to eliminate phone based interviewers to whom recent article on decisionmaking in the or at least minimize the risk (Mansdorf, unsafe acts can be reported, the use of fire environment Braun and Latapie 1993). This investigative approach, anonymous data collection systems, the (1995) noted that training should include however, must go beyond the traditional creation of safety surveys, the use of the rehearsal of behaviors that are microscopic analysis to identify behav- trained field observers, or the use of needed in stressful conditions. Safety iors that are common in a variety of automated data collection systems. critical behaviors must be practiced until accidents. To facilitate the identification Such reporting programs might also they become automatic. Finally, what of these behaviors an investigation team guarantee immunity from disciplinary is the perceived priority of safety? Do should be composed of individuals from actions for individuals who report. supervisors and managers expect safe all levels of the work force (Krause & behaviors? Are firefighters asked to Russell, 1994; Mansdorf, 1993). Evaluating Training and Management. work in high-risk conditions that are Moreover, efforts should be taken to There are a variety of questions that outside of safe parameters? Is there an reconstruct the accident with the aim of must be asked when evaluating training established code of conduct that speci- identifying the underlying behavioral and management. Are instances of the fies the safe behaviors an individual is patterns that might have precipitated it. desired behavior demonstrated during expected to display? Finally, is there Once identified the investigation needs training? For example, fire shelter an accountability system to which all to assess the extent to which these training has traditionally placed more firefighters are held? The answers to behaviors have been present in other emphasis on getting into the shelter than these and other questions provide an incidents or accidents. Finally, the on other factors such as situational indication of the priority safety is given. investigation must assess the degree to awareness, site evaluation, ground which the actions reflect the acceptance preparation, and contingencies all of Feedback and Reinforcement. The of hazardous and risky behavior as which are essential to a successful concepts of training and reinforcement common practice. shelter deployment. Are employees are closely related. At its most basic trained in the selection of the appropriate level, training serves to educate an indi- Assessing the Frequency. To assess behavior? Invariably more than one vidual about the various reinforcement the frequency of unsafe acts, a system option is available for each situation. In contingencies (Anderson, 1995). That for reporting accidents, and near-miss a situation where a burnover is inevit- is, during training an individual learns accidents must be created. Near-misses able, a firefighter can deploy a fire the actions and behaviors that will be play an important role in assessing the shelter or attempt to escape. Factors reinforced when training is complete. frequency of risky acts. From the that influence this decisionmaking pro- After training is complete, are the trained behavioral perspective, near-misses cess must be considered in advance. behaviors expected and reinforced? represent accidents without the conse- Training should include techniques and Moreover, have the trained behaviors quences (Krause & Russell, 1994). procedures used to evaluate the various been directly or indirectly extinguished Moreover, given that unsafe behaviors options. Is there a system to continue by example or directive? For example, infrequently result in accidents, near- training apart from the classroom? On- are firefighters more often reinforced for

31 Part 4 of 4 taking risks than for demonstrating Remediation Plans. Shortcomings in References good judgment? training, supervision, or management should not be viewed in isolation but as Anderson, J. R. (1995). Learning and While it is important to assess if trained representative of a company-wide memory. New York: Wiley & Sons. behaviors have been reinforced, it is pattern of behaviors. Efforts to remedi- just as important to determine if unsafe ate these shortcomings must endeavor Braun, C. C. & Latapie, B. (1995). behaviors have been inappropriately to address both the specific behaviors Human decisionmaking in the fire reinforced by environmental events. and the broader culture. Each plan environment. Fire Management Notes, Although the ultimate goal of firefighting should identify short-term and long-term 55, 17-20. is fire suppression, a suppressed fire is objectives and the criteria against which not an appropriate reinforcer for firefight- the plan will be evaluated. Danaher, J. W. (1980). Human error in ing behavior. This unsuitability stems ATC system operations. Human from the fact that all fires eventually go Factors, 22, 535-545. out independent of the actions taken by firefighters. This inevitability makes fire Conclusions Goldhaber, G. M., & deTurck, M. A. suppression an indiscriminate reinforcer. (1988). Effects of consumers’ familiarity That is, fire suppression could reinforce Programs aimed at enhancing safety by with a product on attention to and both safe and unsafe behaviors. Some addressing the proximate cause of an compliance with warnings. Journal of would agree that factors such as accident only consider a small portion of Products Liability, 11, 29-37. weather often play a larger role in sup- the safety picture. Merely addressing pression than firefighters, but still argue the proximate cause fails to consider that Gordon, S. E. (1994). Systematic that firefighters should be reinforced by the system either directly or indirectly training program design: Maximizing the fact that the size of the fire has been trains, reinforces, and even expects effectiveness and minimizing liability. limited. There might be some truth in employees to demonstrate hazardous New Jersey: Prentice Hall. this statement, however, it is not behavior. An effective safety program completely verifiable because firefighters must consider both the proximate cause Krause, T. R., & Russell, L. R. (1994). often take advantage of areas where the and the working environment that The behavior-based approach to fire would stop on its own (e.g., natural promotes hazardous behavior. The proactive accident investigation. fuel breaks). program must identify unsafe behaviors Professional Safety, 39, 22-26. and assess their prevalence. It must Care should be taken in determining evaluate training to ensure that individ- Mansdorf, S. Z. (1993). Complete the types of reinforcement and feedback uals not only gain the necessary skills manual of industrial safety. New Jersey: individuals obtain from the environment. but are provided with opportunities to Prentice Hall. The containment and suppression of exercise and practice those skills. The fires, the saving of structures and safety program must survey supervisors Rhoades, Q. (1994, August 26). Storm resources, and other similar events and managers to determine if skills King: Effective fire fighting calls for make poor reinforcers because they are learned in training are actively reinforced, bending the rules sometimes. The indiscriminate and because they target and finally, it must make recommenda- Missoulian, pp. A5. the outcome of behavior and not the tions that affect behaviors and the behavior itself. Efforts must be made to system that supports them. reinforce the safe behaviors independent of the outcomes.

32 Part 4 of 4

Naturalistic Decision emerged called Naturalistic Decision The first study I performed to generate Making and Wildland Making (see Table 1). This field models and training recommendations emerged because governmental for decision making under pressure and Firefighting* sponsors such as NASA, FAA, the certainty was a study for the Army. The Gary Klein, Ph.D., Klein Associates Inc. military, and others, realized that they Army Research Institute wanted some 582 E. Dayton-Yellow Springs Road had spent a lot of money and built data on decision making in real, stressful Fairborn, OH 45324 decision models that did not work in the situations, and I thought that urban (513) 873-8166 field. They wanted to get away from firefighters would be a good example of August 8, 1995 building analytical models which didn’t people who had become experts at work when they were brought into making such decisions. We studied The Recognition-Primed Decision action. Naturalistic Decision Making commanders who had about 20 years Model describes what people actually uses expert decision makers, and tries of experience, and studied the most do when they make difficult decisions. to find out what they actually go through difficult cases they had. Of the cases we This has many implications for training in their decision making process. studied, there was an average of five and helping people make decisions changes in the fire and in the way it had under stressful situations. It can also Instead of restricting decision making to to be handled. About 80 percent of the help explain the factors behind bad the “moment of choice,” experts are decisions were made in less than a decisions. asked about planning, situational minute. As we started the study, we awareness, and problem solving to find found that each expert firefighter told us The standard method of decision making out how these all fit together. This model that they had never made any decisions. is the rational choice model. Under this is used to understand how people face They explained that they simply followed model, the decision maker generates a decisions in shifting and unclear procedures. But as we listened, we range of options and a set of criteria for situations and under high stakes. Team realized that in each case, there was evaluating each option, assigns weights interactions and organizational one option which they thought of quickly. to the criteria, rates each option, and constraints with high stakes are also They evaluated that one option, and if it calculates which option is best. This is used as factors. For years, seemed viable, they went ahead with it. a general, comprehensive, and researchers had been simply asking quantitative model which can be college sophomores what they would do We began to wonder how they came up applied reliably to many situations. given a set of options, and a clear goal. with that first option and how they were Unfortunately, this model is impractical. For Naturalistic Decisionmaking able to evaluate one option without People making decisions under time research, experts are asked to size up others for comparison. The strategy pressure, such as fire fighters, don’t actual situations, using all cues and used by the firefighters is the basis for have the time or information to generate constraints to set goals and make the Recognition-Primed Decision (RPD) options and the criteria to rate each decisions. Model (see Figure 1). The first level option.

The rational choice model is also too Table 1—Features of Naturalistic Decision Making research. general. It fits each situation vaguely, but no situation exactly. The worst news Positive Features Contrasts is that in studies in which people have been asked to follow the rational choice • Studies people with expertise • Studies novices model exactly, the decisions they come • Tries to describe • Tries to evaluate up with have been worse than decisions • Takes a broad focus • Takes a narrow focus they make when they simply use their • Task context: field settings • Task context: laboratory settings own experience base. This model is of Time pressure Ample time little value to training because it does Shifting conditions Stable conditions not apply to most naturalistic settings or Unclear goals Stated goals to how people actually make decisions Degraded information Precise information Subtle cues and patterns Clear inputs when faced with complex situations Team interactions Individual tasks under time pressure. Decision aids Organizational constraints Individual tasks which have been produced to assist with High stakes Low stakes the application of the rational choice • Focus on cognitive processes • Focus on analytical strategies model have been largely ineffective. • Relies on Cognitive Task Analysis • Relies on performance measures Because of these drawbacks, a field

*Presented at the U.S. Forest Service Conference, June 12, 1995. 33 Part 4 of 4

Level 1 Level 2 Level 3 SIMPLE MATCH DIAGNOSE THE SITUATION EVALUATE COURSE OF ACTION

Experience the Situation Experience the Situation Experience the Situation in a Changing Context➛ in a Changing Context in a Changing Context more data

➛

➛

➛

Diagnose no➛ ➛ Perceived as typical [Feature Matching]➛ Is Situation Typical? Perceived as typical ➛ [Prototype or Analog] [Story Building] [Prototype or Analog] [Prototype or Analog] ➛inference

Clarify ➛ yes

➛

➛ Anomaly ➛

Recognition has four byproducts➛ Recognition has four byproducts Recognition has four byproducts

Expectancies Relevant Cues Expectancies Relevant Cues Expectancies Relevant Cues

Plausible Goals Typical Action Plausible Goals Typical Action Plausible Goals Action 1...n ➛ ➛

Evaluate Action (n) ➛ [Mental Simulation]

yes, ➛ but➛ no

Modify Will it work? ➛ ➛

➛ yes

Implement Course of Action Implement Course of Action Implement Course of Action

Figure 1—Recognition-Primed Decision model. consists of a simple match, where told a story of a fairly conventional fire. diagnose the event, and when evidence decision makers experience a situation He described parking the truck, getting doesn’t fit the story, the firefighter has and match it to a typical situation with out his hoses, and going into the house. to come up with a new scenario which which they already have experience. I asked him why he went into the house fits the new evidence. There is still no Because of this, they know what to instead of simply working from outside, comparing of options. expect. They know what’s going to as I would have been tempted to do. happen, they know what the relevant He explained that he obviously had to On the third level, decision makers cues are, what the plausible goals are, go in because if he attached it from the evaluate the course of action they have and a typical action. They are able to do outside, he would just spread it deeper chosen. Originally, we weren’t sure how all of this because of their experience inside the house. He took into account people could evaluate single options if base. Experience buys them the ability the nature of the fire, the distance of they had no other options to compare it to size up a situation and know what is the house from other buildings, and the to. As we looked through the materials going on and how to react. That’s what structure of the house. But, even while we were getting, we found that a decision researchers weren’t learning he was attending to these conditions, he decision maker would evaluate an option when they studied college sophomores never saw himself as making a decision. by playing it out in his/her head. If it who didn’t have an experience base. He never experienced that there was worked, they would do it, if it didn’t, they another option. He immediately saw would modify it, and if modifications An example of the first level of the RPD what needed to be done and did it. failed, they would throw it out. In the model is a firefighter I interviewed early incidents we studied, commanders in the process. He explained to me that The second level of the model includes simply generated each option and then he never made decisions. After trying diagnosing the situation. On this level, evaluated it for viability. Usually the to press him on the issue, I asked him expectancies are violated. The first option an experienced firefighter to describe the last fire he was in. He firefighter is trying to build a story to generated was a viable option, but they

34 Part 4 of 4 also understand that they should simply do more than point out the mistakes information. The third factor was a de be satisfying, not optimizing. They will which were made in an exercise. It minimus explanation. In this situation, not necessarily pick the best option. would be an attempt to show decision the decision maker misinterprets the They will pick the first one which is makers what went wrong with their size situation, all the information is available, possible and involves minimal risk. up, and why. Another method would but the decision maker finds ways to The first viable option is chosen and include cognitive modeling and showing explain each clue away, and persists in improved upon, if necessary. It is not expert/novice contrasts. This would be the mistaken belief. compared with all other options to see done by allowing novice decision makers which one will be best. As soon as it is to watch experts. Novice decision The problem of lack of experience has deemed viable, it is chosen and applied. makers would also benefit by learning many effects (see Figure 2). about common decision failures. On- Inexperienced decision makers lack the Naturalistic Decision Making has the-Job Training should be emphasized understanding of situations to be able implications for training. Decision rather than simply assuming that once to see problems and judge the urgency training needs to teach people to deal the traditional training is finished, of a situation, and properly judge the with ambiguous, confusing situations, decision makers are ready to begin to feasibility of a course of action. These with time stress and conflicting function proficiently. Test and are skills which could be developed to information. Situation awareness, evaluation techniques and training improve decision making. pattern matching, cue learning, and device specification could also be typical cases and anomalies can be improved. All of these might have an The field of Naturalistic Decision Making taught by giving people a bigger effect on the ability of firefighters to research is more appropriate than experience base. Training could teach deal with stressful situations. traditional decisionmaking models for decision makers how to construct understanding how crisis managers, effective mental models and time Why is it that people do make bad such as firefighters, handle difficult horizons and how to manage under decisions? I looked through a database conditions such as time pressure and conditions of uncertainty and time of decisions to identify reasons behind uncertainty. We have broadened our pressure. bad decisions. We came up with 25 focus from the moment of choice, to decisions which were labeled as poor. take into account situation awareness, Methods for providing better training Of those, three main reasons for bad planning, and problem solving. By so include changes in such things as ways decisions emerged. By far, the most doing, we have gained a stronger of designing training scenarios. Another prominent reason was lack of vantage point for understanding errors strategy is to provide cognitive feedback experience. A smaller number of poor and for designing training interventions. within After-Action Reviews. This would decisions were due to a lack of timely

Lack of experience

• Poor sense of typicality • Poor ability to see patterns and make discriminations • No basis for expectancies • Missing causal knowledge

➛ ➛

➛

Insensitive to the early signs of Misrepresent the situation Unaware of weaknesses in a a problem course of action • De minimus explanations • Anomalies are not recognized • Alternate SAs are not considered • Urgency is misjudged • Difficulty in handling complex: • Multiple interactive cues • Ill-structured problems

Figure 2—NDM factors that————➛ poor decision outcomes.

35 Part 4 of 4

Cultural Attitudes and All these factors affect how well the others a lifetime. Another way of looking Change in High-Stress, decision making machine works. If you at it is to say that your attitudes come think back, you’ve probably encountered from your values and goals (remember High-Speed Teams most (if not all) of these factors during those DM factors). So the attitudes you David O. Hart, TID, Inc. fire fighting operations. use as firefighters come from your training and experience as firefighters.

What is Decision Making? DM and Attitudes What Do We Do with As we saw in the other presentations, In this discussion, the factors we’ll be Attitudes there are a variety of ways to model concerned with are those that relate to decision making. The importance here and affect cultural attitudes. In general, Attitudes help us make sense out of our is that it can be modeled, described, and attitudes that enhance the DM process surroundings and allow us to build and examined. By examining decision are seen as positive, and those that act maintain our Situation Awareness (SA). making as a system, we can learn how as barriers to effective DM as negative. How? By providing each of us a set of attitudes, individual and cultural, affect Many attitudes have both positive and rules and guidelines we use to gather the quality of our decisions. negative effects. All this may seem and process information. Therefore, intuitively obvious to even the most attitudes aid in our decision making by There are as many decision making casual observer, but it is important to framing and shaping the information we definitions as there are models. For this establish a common ground before we use to make our decisions. You could discussion we’ll need to have a common delve too deeply into this subject. In almost say that attitudes are imbedded reference to work from when talking the spirit of “crawl, walk, run” we’ll need in every aspect of decision making. about decision making. Also, because to first understand how attitudes affect Good, bad, or indifferent, attitudes we are talking about organization and the individual before we can understand affect the quality of our decisions. team decision making, we’ll focus the the impacts of cultural attitudes on an following definitions in that direction. A organization. On a team, the synergy that develops definition of decision making to keep in can compensate for attitudinal failures mind during this discussion is: or barriers in one of its members. Effective teams recognize attitude The process “of reaching a decision Attitudes and the Individual problems and find ways to work around undertaken by interdependent individuals the “attitudinal outages”. A good to achieve a common goal. What Before we go too much further, we’ll example of this is the issue of women distinguishes team decision making is need another definition. This time we’ll as crewmembers in combat aircraft. the existence of more than one informa- be defining attitudes. Many male aircrew have a real tion source and task perspective that “attitude” about women in the cockpit. must be combined to reach a decision.” The American Heritage Dictionary of Probable fallout from this barrier is the English Language defines attitudes reduced communication, increased as: “a state of mind or a feeling; stress, conflict, with a resulting loss of disposition.” A longer definition is: “An efficiency and effectiveness. A good Decision Making Factors enduring organizational, motivational, team will recognize the barrier and emotional cognitive process with respect react by: Close examination of this definition to some aspect of the individual’s world. reveals many important aspects of the Attitudes and beliefs imbued with • Increasing communication to and decision making process in high-stress emotional and motivational properties.” around the affected people, environments. These include, but are Another shorter definition, is: “Affect for • Closely evaluating the information not limited to: or against a psychological object.” sent by the affected parties to weed out any attitude biases, • Ill-structured problems They all say the same thing—an attitude • Finding ways to reduce stress • Uncertain, dynamic environments is how you feel about something. Now (knowing military crewmembers, • Shifting, ill-defined, or competing goals that we know what attitudes are, let’s humor would be a likely choice), • Action-feedback loops see where they come from. • Defusing any conflict before it • Time stress engages the entire crew. • High stakes Generally, your experience forms, has • Multiple players an effect on, or shapes your attitudes. • Organizational goals and norms Some attitudes may last only minutes, 36 Part 4 of 4

We’ve looked at the what, how, and It is important to note that despite these We see this in the report where the why questions regarding attitudes and undesirable results, critiquing and “can-do” attitude is attributed with the the individual, and even looked briefly correcting the failure is difficult because compromise of the 10 Standard at a possible individual attitude outage you can’t “see” the cause. Firefighting Orders (SFOs) and 18 scenario and the team’s possible Watch Out Situations (WOSs). response. Now let’s turn our focus to teams. When there is a disconnect between Where Attitudes Come From training and experience, a barrier to effective decision making exists. This We’ve already determined that an disconnect causes a gap between the Attitudes and the Team individual’s attitudes come from his or individual and resulting team perception her values and goals. The same holds of reality and actual reality. This Cultural attitudes—what are they, and true for any organization. The cultural example is more ambiguous than the why are they different? As to what they attitudes grow out of the organization’s previous two, but when seen in an actual are, our definition is still valid, but with values and goals. The source for these example, it leaps right out at you. The this added: the attitude is shared by attitudes can be either internal or SCF report found that “some firefighters every member of the organization. external to the organization. failed to recognize the capability and Organizations and teams use attitudes limitations of the fire shelters and for the same purpose as individuals, to Internal sources are the easiest to deployment sites.” And “some build and maintain their knowledge of identify. Policy statements, directives, questioned the value of the fire shelters the environment. The big difference is and even official memos are examples under any conditions and may not have that the synergistic effect of teams of how organizational goals and values been carrying shelters.” It is apparent magnifies and multiplies the effect of manifest themselves. that the training received was not attitudes. supported or validated by the Looking to the South Canyon Fire (SCF) experience of the cited firefighters. The multiplication and magnification cuts incident, the Grand Junction District This kind of gap between perception both ways. Positive attitudes provide a Management Team directive that all fires and reality can, and has produced, uniform strength and negative attitudes, be “initial attacked and suppressed as deadly results. uniform weaknesses. An example of a soon as possible” is an example of policy positive effect is providing baseline working as a cultural attitude. What you The final internal example is the attitude goals, values, and priorities (once gain from this attitude is a concrete or sense of being part of a larger again, remember the DM factors), to direction for the firefighting teams. The “family.” This is most often seen as an establish a cohesive team more easily goals of their decisions are elitist attitude. In this case we use elitist and quickly. Failures are much more unambiguous. On the flip-side, this to mean special, different, or set apart. insidious. attitude can become a decisional one- It is often expressed with the statements way road. It doesn’t provide a way out “we watch out for our own,” or “we take When an attitude fails (e.g., is no longer of a fire that cannot be suppressed. care of our own.” This increased valid) or is working against a team, it Also the added emphasis on mission awareness of your teammembers becomes an attitudinal “blind spot.” accomplishment can come into direct translates into an increased safety Because everyone in the team and/or conflict with existing safety attitudes. awareness. Carried to an extreme, it organization possesses the attitude, no can result in a lack of leadership. The one can perceive that there is a The “can do” attitude identified in the B-52 bomber crash at Fairchild AFB in problem—there is nothing to compare it SCF investigation report is common to Spokane was allowed to happen against. For example, the team has an many high stress, high speed teams. It because the commanders at the base attitude barrier that inhibits helps build team cohesion, which is failed to ground the pilot for flying the communication. By reducing the amount important to the team for synthesizing aircraft outside its operational limits of information flow, and possibly, information and integrating the because, he was “one of our own,” and information quality, there can be a individual perceptions of the situation for fear of “ruining his career.” substantial loss of synergy, into a common perception. But taken cohesiveness, leadership, recognition, too far, this attitude can have lethal For external sources of organizational awareness, and communication. All consequences. By going above and attitudes, we’ll look at two particular to these elements, working at full capacity, beyond to complete the job, mission firefighting, and one common to the are crucial to effective decision making. success is prioritized ahead of safety. entire federal government.

37 Part 4 of 4

Pressure from the public and media Attitudes, Training, and This last attitude is a training/experience generates the attitude that fires with the Experience trap stemming from the fire training most public attention should be attacked attitude and the fire shelter attitude. first. Normally, being responsive to the Attitudes, training, and experience have needs of your customer is seen as a a deeply interrelated relationship. What impact did these attitudes have positive goal and attitude. But by Cultural attitudes affect the emphasis of on the incident? First, we need to allowing people outside the training, and experience shapes and recognize that safety and operational organization to control priorities, you modifies our attitudes. When experience effectiveness are opposite sides of the end up with shifting, ill-defined, or and training validate each other, there same coin. The first Standard Fire competing goals (sound familiar?). is usually a positive attitude effect. Order supports this. At South Canyon, When they don’t support each other, the additional emphasis suppression The harsh spotlight of the news media there’s usually a negative attitude effect. received was both caused by and can have a similar effect. An resulted in the erosion of safety margins. organization is usually highlighted Start with the training attitude that by Each time the firefighters “got away with” because of some failure or near-failure. emphasizing fire behavior, fuels, pushing into their safety margins to The organization usually responds by weather, and tactics, entrapments will be suppress a fire, it reinforced the attitude reacting with abrupt changes in goals avoided. Add to that the historically low that they could do the job with a smaller and values, then attitudes, then frequency of losses, an experience margin for error. The fact that some of decisions. In the case of the SCF, the based invulnerability attitude (i.e. “it the firefighters were uncomfortable with reaction was increased emphasis on won’t happen to me”) can develop. The the situation at South Canyon safety, but unless the spotlight is on overall experience, expertise, and demonstrates that Grand Junction’s something that needs to be changed, success of firefighters fosters the attitude suppression directive was causing some the resulting changes may not be for that they can handle any fire (i.e. elitist, shifting and competing goals. This the good of the organization. can do, or 10:00 fire), which in turn feeds erosion of the safety attitude coupled the training and experience attitude “why with SA and communication breakdowns The last external example is one that should we over-learn emergency critically compromised the team and everyone connected with the federal procedures (fire shelter use and bailing individual decision making ability. government, most state governments, out of a situation). From this vantage Among the elements that led to this and some corporations have felt: “do point, it would appear that these attitudes breakdown are physical and mental more with less”. In a perfect world this are leading firefighters to lean on luck fatigue, recognition gaps, weather would allow organizations to get the and circumstance to keep them safe. information not communicated or used, most from their resources. safety concerns not communicated, Unfortunately, we don’t live in a perfect The combination of low frequency of concerns about who was in charge world. In reality, this attitude is a time losses (experience), and highly (leadership) and the numerous bomb just waiting to go off. experienced teams (experience) compromises of the SFOs and WOSs. conspire to subvert important safety When the blow-up occurred, these “Do more with less” pushes people and procedures and attitudes (training). came together with deadly results. The equipment to perform beyond their attitudes also blocked the last escape capabilities, usually by sacrificing the path—dropping tools and packs, bugging normally accepted margins of safety. It out, and using shelters. usually takes a catastrophe many times Attitude Impacts on SCF worse than the SCF for the federal After situations like these many leadership, from Congress on down Cultural attitudes played a significant questions are raised. Some that need through each agency involved in the roll at South Canyon. Some of the to be answered in order to affect any concerned operation, (in this case cultural attitudes that were carried into kind of change are: wildfire fighting) to realize that you do the fire were: less with less. Adopting a “do less with • Do tactical teams know to increase less” attitude would mean letting some • “All fires will be initial attacked and meaningful communication during a fires burn themselves out when they suppressed as soon as possible.” crisis? Also, do they know how to don’t directly threaten the local populace. • “Highest priority fires are ones that communicate effectively? Unfortunately decisions like these threaten life, residences, structures, • What about pre-planning for crisis usually come at an immeasurable cost. and utilities.” situations? • “We can handle the fire.” • Do tactical teams get the best informa- • “Can do” tion before and during a fire crisis? • “It won’t happen to me.”

38 Part 4 of 4

Changing Cultural Attitudes For the individual, making the change Where does it start? Programs which can be as simple as changing the work to improve team attitudes and Before we look at examples of how these behavior associated with the attitude. effectiveness usually consist of a changes are affected, let’s look at why This can happen very quickly, but may number of inter-connected training that change is made. not have a lasting effect. As soon as modules. the need for the change has passed, Why do cultural attitudes change? the individual is likely to revert to old Initial “awareness” training is designed Because it is recognized that the long behavior patterns and start the cycle all to introduce the program and set the term goals of the organization are not over again. Actually changing the stage for the training to follow. It is being met. attitude is more difficult than changing usually directed at all organizational the behavior. It takes more time, but members who are targeted for change. How do you recognize that an attitude has a more permanent effect. For an is no longer valid? Since you have no organization, the time and effort is A leadership/management “staff” course “attitude out” light, you usually know by greatly magnified. for the senior management is also unwanted results produced by practicing conducted in the initial phases. These the behavior associated with the attitude. Commitment, or lack thereof, will either programs provide management The feedback from the environment may make or break this type of program. personnel the essentials to fulfill their be obvious or subtle. Because of the role in the change process. They need blind spot effect talked about earlier, it What needs to be changed? Initially to “walk the talk” if they expect the rest is harder for teams to find the offending a survey of the organization should be of the organization to do the same. attitude than for individuals. conducted to determine the attitudes and Organizations, being larger and more values regarding team effectiveness. Baseline training is the longest and complex than their component teams, Areas that are typically covered in this most in-depth phase. It provides the find it more difficult digging out an type of survey are leadership, background, vocabulary, skills, and invalid attitude. communication, recognition and feedback the teams need to affect this management of stress, needs for change. Why is it harder for an organization to achievement, and job satisfaction. For change an attitude? You have many accurate data to be gathered the need Instructors and Evaluators play a special more people needing to change and for anonymity is essential. In addition a role and therefore need special training. change is naturally difficult for people. cross-section of the entire organization, This type of training is focused on Because they’re doing something new top to bottom, left to right, needs to be observing, instructing, and evaluating and different, it take times and effort to sampled to prevent inaccurate, the new attitude. make it stick. Let’s look at a typical misleading, and skewed data. This process by which organizations can information is then used as a benchmark Finally, continuation training provides change attitudes. Then we’ll look to to measure the change against, and to ongoing reinforcement of the concepts commercial and military aviation as help determine the types of tools to learned in the baseline training. For the examples of organizations that have necessary to make the change. best results, it should be practiced in an undertaken this kind of change. environment as close to actual as How does it happen? Using the data possible. Preliminary Requirements. Before from the survey, a program of change is the change process can be started, the developed. Usually this takes the form As with support, training must also run organization, in particular the senior of training or organizational interventions. from the top down. No one is exempt leadership and managers, needs to The program is usually developed by or from training, no matter what their recognize that their greatest contribution in conjunction with professionals standing in the organization. Each to this sort of change is providing a involved in this arena. Credibility of the phase builds on the previous. This supportive environment that will foster developers, program, and delivery continuity is necessary so that previous the growth of the change effort. personnel is critical to the program’s training isn’t invalidated by the next success. This is the first step in assuring phase. The training that is the most Patience, perseverance, and the buy-in of the front-line teams. important is usually the most neglected. commitment from the leadership and managers is absolutely necessary. Finally, programs should be designed Instructor/evaluator and continuation Recognizing that this sort of change to fit seamlessly into the culture. It training are probably the two most critical happens one person at a time and that can’t be seen as one time fix or just modules for assuring long-term success. it will be slow and sometimes difficult, another training requirement. To The instructors and evaluators must they will be supporting the change and change the culture, it must be part of embrace the change and its concepts their own role in the effort. the culture. and procedures, or the training will be 39 Part 4 of 4 useless. Lack of buy-in from the We’ll look now at commercial and mili- answer—but how do you keep it from instructors and evaluators can result in tary aviation to see what brought them happening? This answer took the form of training invalidating training, and evalu- there and what they’ve done and gained. Cockpit Resource Management (CRM). ation invalidating or ignoring training. A program for change was initiated at a Continuation training, on the other hand, number of airlines. It probably looked keeps the ball rolling. Remember this In the Beginning... like the program we just outlined. What is a long term program, not a quick fix they found was that certain elements in Band-Aid. These concepts and skills The 1970s saw a number of air carrier the human equation needed change. need to be revisited not just annually, crashes. The fact that aircraft crash They were, and are, communication, but at every training opportunity if it is wasn’t new, but the reasons for stress management, leadership, going to be a permanent part of the crashing were. More and more decision making, and attitudes. culture. As with anything new, practice, accidents were being attributed to These programs are designed to make practice, practice makes perfect. One “human” or “pilot error.” Highly the pilots and flight engineers more final, important point regarding experienced, trained, and motivated effective and efficient flight crews. continuation training—Keep it Fresh!! (sound familiar?) crews were allowing Nothing will kill a program faster than aircraft to crash. Most notable is the As the programs became more and tired, overused training material. As Portland DC-8 crash where the crew more a part of the airline culture, the new information becomes available it flew the aircraft out of gas while benefits of this type of training was should be integrated into the program. troubleshooting a gear problem on a seen in other areas within the clear night within sight of their community. They also started seeing Looking back...we see that this is just destination. Another is the L-1011 that some return on their investment. a sample of what a program for cultural slowly descended into Florida change could look like. A real program everglades as the crew tried to decide A notable (but not isolated) case is the is much more complex, but then again, what was wrong with a 68¢ lightbulb. Sioux City DC-10 crash. Enroute to their real change is much more challenging. The crew was focused on the lightbulb destination, the #2 engine, the one in the and no one was minding the store: why? vertical stabilizer, disintegrated. Pieces Other organizations have undertaken to of the engine cut through the hydraulic change attitudes within their culture. The “why?” questions were asked by lines for the primary, secondary, and Most notable is the aviation community. the airlines also. Human error was the standby systems. Without hydraulic

40 Part 4 of 4 power, the pilots were unable to control for them to see the benefits first. MAC Last Words any of the flight control surfaces. By all “spun” the airline programs to better fit rights, the aircraft should have crashed, their environment. The military was Changing a cultural attitude can be a killing everyone aboard. That’s what interested in the effectiveness and daunting process. But in this the engineers at the airline and aircraft efficiency issues, but were more environment, as in some of the others manufacturer said. But Capt. Al Haynes interested in CRM’s major by-product: we’ve talked about, ignoring an attitude attributes his and the passengers’ SAFETY. In an environment where that is in conflict with the organization’s survival directly to CRM. The open, your enemy is actively trying to reduce goals and values is not just inconvenient, continuous communication, creative you to an aluminum rain shower, a it’s downright lethal. synergy, and their recognizing and using program that keeps you from doing all available resources are principles at your enemies’ job is always attractive! By believing that what you’re doing is the heart of CRM, and were the ones important, you will be able to make the used successfully by the flight crew. Today, CRM is an inseparable part of changes in your culture. These changes the airline culture. Human factors will have far reaching benefits for the Increased focus on and awareness of related accident rates are down, individual and the organization in safety, effective and efficient flight operations incidents are down, safety is up and so decision making, and operational helped to broaden the scope of the is efficiency. The program is working. effectiveness. program. The first to be brought into the fold were the cabin crew, hence the As for the military, the change is still —TIG, Inc. is a consulting company in Aurora, name change to Crew Resource taking root. Military CRM hasn’t reached Colorado. We specialize in the delivery and development of Crew Resource Management Management. Then it spread to the the stage the airlines have, but then as (CRM) and human factors training. TIG, Inc. is maintenance organizations. we have said, these things take time. It currently providing services to the Army Guard has also moved out of the aircraft arena. Special Forces and The USAF Reserves flying In the early to mid-1980’s military Other military units are seeing the and maintenance organizations. aviation became aware of the benefits benefits of CRM. Maintenance, test of CRM. The USAF Military Airlift engineers and pilots, and special forces Command (MAC) was the first to come units are just a few that have embraced on board. Their operations were the the concepts of CRM. closest to the airlines, so it was natural

41 Part 4 of 4

South Canyon Revisited: that they were jumping on a fire that not, the departing Incident Commander would likely be out by 10:00 the next says in his statement, “I knew (ia.) that Lessons from High morning. After landing, all some of Mackey would look (sic) at fire from the Reliability them knew was that Dodge had air before they jumped and that he Organizations1 scouted the fire on the South slope with would make a decision on what to do Karl E. Weick, University of Michigan Harrison, had used the phrase “death with it after we left. I did not feel that trap” to describe what he found, and smokejumpers needed additional had ordered the second-in-command guidance” (Report, 1994, p. A 5-9). In this paper I want to explore the idea William Hellman to march the crew down Mackey got off to a bad start, and the that organizing to prevent wildland fire the North slope toward the Missouri quality of the briefings didn’t improve disasters such as the South Canyon Fire River. Dodge didn’t say whether this much from then on. For example, the on July 6, 1994 in which 14 people lost tactic was to escape the death trap or Prineville Hot Shots were not told how their lives, is an ongoing struggle for to position the crew to fight the fire, or Gambel Oak burns when it is dry, nor alertness. My intention is to look more simply to get closer to the river. When were they told that in previous days, closely at that struggle. I want to do 4 the fire spotted to the North side of the fires had made spectacular runs things. First, I want to discuss 4 pieces gulch, Dodge turned the crew around through this material in Colorado. of my earlier analysis of the Mann Gulch and angled them up toward the ridge, fire that seem relevant to South Canyon. and soon ordered them to drop their Why so much casualness? One In particular, I want to discuss briefings, tools, and then to enter an escape fire, possibility is that everyone seriously leadership, tools, and wisdom. all without verbalizing his reasons underestimated how much continuing (Dodge, 1949, p. 121). Since the crew effort and shared information it takes to Second, I want to discuss organizational did not know each other well, since build coordination and hold it together, issues at South Canyon that are less Dodge knew only 3 of them, since especially during transitions from an visible in Mann Gulch. These include several were on their first jump, and initial attack to an extended initial attack, discrepancies, levels of experience, the since Dodge himself was rusty on from one level of complexity to another will to communicate, and Watch Outs leading a crew (Maclean, 1992, p. 41), it level, and from one organization to involving management. Third, I want to was imperative to build some common another. The investigation team, on p. 6, touch on solutions. And I want to understanding and common action into states the following: “as is typical in conclude by discussing some questions this assortment of strangers. That extended attack situations, firefighting about South Canyon that continue to didn’t happen. groups arrived on the fire at intervals haunt me. from dispersed locations and blended But neither did it happen 45 years later into the existing organization.” The key at South Canyon. The South Canyon word there is “blended.” Blending accident investigation team allocated sounds like something that occurs Similarities Between Mann almost a full page (Report of the South automatically not something that people Gulch and South Canyon Canyon Fire Investigation Team, 1994, work at. Many would say it’s especially p. 26: hereafter referred to simply as hard to blend into an “existing Briefings. The struggle for alertness at Report, 1994) of their report to “Safety organization” if that organization itself is Mann Gulch was undermined by many briefings” as a “significant contribution” invisible, as was the case for some of the same things that undermined it at to the 14 deaths. The hand-off of the people at South Canyon. Some people South Canyon, one of which is briefings. fire the evening of July 5 from the BLM trying to blend did not know who the Briefings are an attempt to give people crew to the smokejumpers and Jumper- Incident Commander was, or which in a crew a common framework in in-charge Mackey is a good example of radio traffic had the force of authority, advance including assumptions about how not to brief people. The hand-off is or what the suppression strategy was what they may face, how it will develop, by radio rather than face to face, is made since it seemed counter-intuitive. and how the crew will function and after the BLM crew who know the update its understanding of what is terrain and foliage has left the scene, The questions that need to be pursued going on. and the jumpers inherit a handline which are, why does briefing continue to be is partially constructed but already lost treated casually and what does better At Mann Gulch, the crew of 14 by the time they collect their gear and briefing sound like? Back in 1949, essentially proceeded without much of are ready to extend it. Without checking during the investigation of Mann Gulch, a briefing. They basically knew only whether the assumption is correct or Henry Thol’s father understood the

1Paper presented at Decision Workshop on Improving Wildland Firefighter Performance Under Stressful, Risky Conditions: Toward Better Decisions on the Fireline and More Resilient Organizations, Missoula, Montana, June 12-16, 1995. I acknowledge with appreciation the generous assistance of Ted Putnam in the development of these arguments, as well as the inputs from other participants in the decision workshop. I also thank David Thomas and Paul Gleason for their continuing encouragement. Please direct any correspondence concerning this manuscript to Dr. Karl E. Weick, School of Business Administration, University of Michigan, Ann Arbor, MI 48109-1234. Telephone: (313) 763-1339. 42 Part 4 of 4 essentials of a briefing even if much of raced uphill (Sallee, 1949, p. 76) and trained to wear steel-toed shoes at all his emotional testimony (“I owe this to reportedly said “to hell with that, I’m times and often refuse to take them off my boy”, p. 201) was tough to follow. getting out of here,” when Dodge ordered when they are ordered to abandon a “Usually the foreman he always looked people to jump into his escape fire. sinking ship. Fighter pilots report being out for all, to take care of anything that reluctant to eject from the “warm womb” happened. We always looked out for At Mann Gulch people were torn and “cocoon” of oxygen in a cockpit that that before he put the men on the fire between 2 conflicting influences. But, is out of control into a far more harsh line. He had something to fall back on the same thing happened at South environment. It is just as hard to drop . . . let’s go in there boys, the wind isn’t Canyon. Haugh and Erickson both shoes or an aircraft as it is to drop a blowing now. We’ll go in there. But yelled at the retreating Hotshots to drop pulaski and a pack. watch out, the wind can change any their tools (Report, 1994, p. 16) and run moment” (Thol, 1949, p. 200). More for the ridge while Thrash, who was at At Mann Gulch, Dodge told his crew to recently, researchers have studied the head of the line of jumpers and “drop all heavy tools” 200 yards after effective cockpit crews in aircraft and hotshots stopped and began to deploy they turned upslope. According to have found that better briefing leads to his fire shelter as did smokejumper Sallee (1949, pp. 75-76) and Rumsey better performance. This is relevant Roth. Hipke and Blecha said in (1949, p. 103) people either threw away because in cockpits, as well as on fire essence, to hell with that, I’m getting everything or nothing. Dodge in his lines, people often work with strangers. out of here and continued to run. testimony said he “didn’t know until In particular, effective leaders establish later that they had discarded shovels and reaffirm norms of conduct for This similarity may be merely a coinci- and pulaskis” (1949, p. 118). Sallee behavior in the group, and insist that dence. It may be more significant. It reported that with the fire racing at them, people keep each other informed on seems worth exploring, however, smokechaser Harrison was sitting what they were doing and the reasons because it adds uncertainty to a situation resting “and he still had his pack on” for their actions and the situational model that already has lots of puzzles. Uncer- (Sallee, 1949, p. 88). that gave rise to those reasons and tainty about leaders puts increased actions. Almost no one at Mann Gulch demands on crews, dispatchers, and This same pattern was repeated at or South Canyon heard someone say, pilots at a time when they are close to South Canyon. Some of the smoke- overload. Uncertainty pulls groups apart jumpers who deployed their shelters 1) Here’s what I think we face: which, makes them more susceptible to above the lunch spot, did drop their 2) Here’s what I think we should do; panic (Weick, 1993, pp. 637-638). And tools. But in doing so, they were struck 3) Here’s why; uncertainty in the face of unclear leader- by the enormous symbolic significance 4) Here’s what we should keep our ship often cuts off the flow of information of what they were doing. One observed eye on; because people don’t know who to send that putting down a saw was like 5) Now, talk to me. it to and responsibility keeps shifting at running up a white flag (Rhoades will. As we will see later, uncertainties statement); another (Petrilli), that the Leadership. But Mann Gulch and about leadership were not confined to “Pucker factor” went up a notch (Report, South Canyon are similar not only in South Canyon. They extended up 1994, p. A5-69). their casual briefings. There was through the organization and this sets uncertainty about leadership in both the tone for actions reflected throughout What about those who didn’t drop their cases. At Mann Gulch, leadership the organization. tools? If dropping your tools signifies moved uneasily among Navon, you’re in deep trouble, keeping them Hellman & Dodge. At South Canyon, it Tools. A small, but powerful similarity may help you feel you’re safe. To hold moved uneasily among Blanco, between Mann Gulch and South onto your tools is to stay in control, to Mackey, Longanecker, Shephard, Canyon is that, in both cases, when remain a firefighter rather than a victim, among others. At Mann Gulch, as at people were fleeing the blowup and to appear calm. I’m still in it. This is not South Canyon, crew members were not were told to drop their tools so they just an issue of symbolism since tools closely acquainted with their foremen could move faster, some resisted. are needed to scrape an area clear due to continual rotation of people Several calculations suggest that this before deploying a fire shelter. But the among crews and assignments. (Fite, resistance may have cost them their reluctance to drop tools may come from 1949, p. 28). Dodge knew only 3 lives (Report, 1994, p. A3-5). They other sources such as economics, people in his crew, Hellman, McVey, would have been able to move 15-20% habits, avoidance of failure, predictions and Thol (Dodge, 1949, p. 125). faster (Putnam, 1994) without their of fire behavior, and social dynamics. Hellman, who was better acquainted packs and tools. Firefighters are not the Equipment is expensive and jumpers, at with the men (Dodge, 1949, p. 125) only people who are reluctant to drop least, are told repeatedly and early in was near the front of the line as they their tools. Naval seamen on ships are their training to carry out everything that

43 Part 4 of 4 is dropped to them. Habits built up behavior could have created the know is only part of what could be during training are much more likely to impression that the crew was at the flank known, and therefore, you need to stay involve moving with tools in hand, rather than the head of the fire which alert. You need to avoid excess rather than moving and discarding tools. meant there was no need to drop tools. confidence that you know everything and People have no idea what it feels like to excess caution that you know nothing, if run and discard tools or even how to do Finally, people may hold onto their tools you want to stay flexible. it. Rhoades in his statement mentions as a simple result of social dynamics that as he was running to escape the when people are lined up. If the first Wise organizations know what they don’t South Canyon fire he kept looking for a person in a line of people moving up an know. They know two things: first, they place to put the saw down so it wouldn’t escape route keeps his or her tools, then know that they have not experienced all get burned, a search which undoubtedly the second person in line who sees this possible failure modes and second, they slowed his progress. In his words, “at may conclude that the first person is not know that their technology is still capable some point, about 300 yds. up the scared. Having concluded that there is of generating surprises (Schulman, hill....I then realized I still had my saw no cause for worry or that I’m not going 1993). Thus, when they act on the basis over my shoulder! I irrationally started to be the only one who goes back of their past experience, wise organiza- looking for a place to put it down where without tools, the second person also tions act as if that experience is both it wouldn’t get burned. I found a place I retains his or her tools and is observed credible and limited. They simultane- it (sic) didn’t, though the others’ saws to do so by the third person in line who ously believe and doubt they know what did. I remember thinking I can’t believe similarly infers less danger than may is up. Consider the case of a near miss I’m putting down my saw.” These words exist. Each person individually may be or a close call. The fascinating thing have even more impact when it is fearful, but mistakenly concludes that about a near miss is that, “Every time a recalled that, among the fatalities, everyone else is calm. Thus, the pilot avoids a collision, the event firefighter #10 (Putnam, 1994) was situation appears to be safe except that provides evidence both for the threat found with a saw handle still in his hand. no one actually believes that it is. The and for its irrelevance. It is not clear To discard one’s tools may signify more actions of the last person in line, the whether the learning should emphasize than giving up control, it may also be an one whose back feels most intensely how close the organization came to admission of failure which, in a “can do” the heat of the blowup are observed by disaster, thus the reality of danger in the culture, is a devastating thing to admit. no one, which means it is tough to guise of safety, or the fact that disaster convey the gravity of the situation back was avoided, thus the reality of safety in There is a further complication with the up to the front of the line. the guise of danger” (March, Sproull, and seemingly simple act of dropping one’s Tamuz, 1991, p. 10). If the moment is tools. If people drop their tools, they What hasn’t changed in 45 years is the interpreted as safety in the guise of still face a tough choice, namely, do I power of symbols. Packs and saws may danger, then learning should be now run faster or do I stop and deploy be heavy and slow one’s pace. But that diminished because “more thorough my shelter? It is tough to do both may be one of their less important investigations, more accurate reporting, although some people at South Canyon qualities. More significant may be their deeper imagination, and greater sharing tried. Running faster and stopping to ability to reduce one’s sense of danger. of information” are all discouraged deploy are incompatible and uncertainty If throwing tools is a sign of surrender, (Sagan, 1993, p. 247). The attitude of about which one to do may compel keeping them may be a sign of a wisdom sees a near miss as evidence people simply to keep doing more of standoff or victory. It may be important that the system is both safe and vulner- what they are already doing, namely, for trainers to emphasize, “Look people, able, that people must remain alert, and running with tools. To keep running is you’re going to want to hang onto this that a safe environment is not measured to postpone having to make a tougher stuff. Don’t! It could cost you your life. by an absence of accidents (that out- choice, especially if the person feels come is largely dependent on luck), but both exhausted and uncertain how safe Wisdom. The fourth aspect of my is the result of active identification of the shelter really is. People may also Mann Gulch analysis that fits South hazards and their elimination (Allinson, hold onto tools because their predictions Canyon centers on the idea of wisdom. 1993, p. 186). of fire behavior suggest that the fire To understand why the idea of wisdom won’t reach them. This is a clear fits here, you need to understand first At Mann Gulch, people believed they possibility at South Canyon. As the fire that wisdom is a mixture of knowledge were fighting a fire that would be out by moved toward the hotshots and jumpers and ignorance. When one of them 10:00 the next morning and failed to moving North along the fireline, it grows, so does the other. To know raise questions about whether this repeatedly was channeled toward the something better is also to discover that expectation remained accurate. At ridgeline along draws that ran at right new questions about it are raised. South Canyon people believed they angles to their movement. This fire Wisdom is an attitude that what you could “hook” the fire before the winds

44 Part 4 of 4 would build and they presumed that recklessness, and they are able “to say troublesome discrepancy. Building lookouts and a commander had the big ‘no’ without sustaining dents in their direct line downhill is dangerous. picture even though the firefighters had machismo” (p. 28). That’s what wisdom Longanecker said, “going downhill direct seen no evidence of this. is about, and why it’s worth striving for. is a bad deal” (Report, 1994, A5-52). Archuleta asks, why are we punching in The attitude of wisdom is one way to line? Erickson asks, “Where are the remain alert, because it leads people to safe areas?” and hears the answer, remain open to what is happening and to Differences Between Mann “there really aren’t any.” Rhoades, rely cautiously on their past experience. Gulch and South Canyon Doehring, and Shelton overhear this I’ve always been struck by evidence conversation. But the decision is made suggesting that there are certain periods Discrepancies Between Beliefs and to build a direct line anyway, which during a person’s career, when they are Actions. Having suggested at least 4 leaves everyone tense. They believe most in danger of getting injured or ways in which dynamics of organizing that the action is dangerous, yet they killed. Police, for example, are in most in South Canyon replay themes that are doing it. What makes this really danger of being shot during their 5th unfolded earlier in Mann Gulch, I now troubling is that the decision is a public, year on the force. Firefighters are in want to explore some additional issues irrevocable, choice. There is good most danger of fireline accidents either that were less visible in Mann Gulch but research evidence (e.g., O’Reilly and in their first 2 years or after 10-15 years that stand out in South Canyon. Caldwell, 1981; Salancik, 1977) that of experience (Pyne, 1984, p. 391). when people make choices of this kind, Young firefighters are vulnerable The first of these is the unusually large they are more likely to change their because of their inability to recognize number of inconsistencies between beliefs so that they become consistent hazardous situations. The more beliefs and actions at South Canyon. with the action they are now committed experienced firefighters are vulnerable I want to dwell on these because they to. In this case people should begin to because they presume they’ve seen it suggest one reason why people believe that building direct line downhill all, they have less openness to new persisted so long doing things that is safe after all in order to justify what data, thus the validity of their models violated fire orders and watch outs. they are actually doing. decreases. The unexpected gets them. A recurring belief among people fighting And that’s what seemed to happen. Crews and commanders need to keep wildland fires is that some of the fires Listen to how Quentin Rhoades in his learning and updating their models. they fight are on worthless land. This own words, handled things: “I resolved This won’t happen if they presume that was a prominent issue at Mann Gulch. not to go down that hill digging line . . . nothing about fires can surprise them, if As Earl Cooley (1984) put it, “One of the Smokejumpers arrived and started near misses are treated as testimonials main questions was why we risked lives digging line. I remember thinking that I to safe practices, and if they are certain and spent many thousands of dollars to must have missed something. I hadn’t that they’ve experienced all possible save scrubby timber and cheatgrass” been on a fire since August 18, 1992 ways in which a system can fail. These (p. 91). A basic discrepancy that and I felt a little green.” Rhoades attitudes won’t change if they reflect firefighters and overhead face over and convinces himself that the main reason similar attitudes in top management. over is between their belief that the land the situation seems dangerous is that You may recall that Maclean felt “the is worthless and the reality that they are it’s his fault, he’s rusty, he’s missed Forest Service wanted to downplay the risking their lives to defend it. The action something, which means the situation is explosive nature of the Mann Gulch fire of defending is inconsistent with the not as dangerous as it looks. Other to protect itself against public charges belief that the area is worthless. people resolve the discrepancy in other that its ignorance of fire behavior was Contradictions such as this cause ways. They convince themselves that responsible for the tragedy” (Maclean, tension and continue to do so until the the leaders know what they’re doing, 1992, p. 125). The key word there is person either changes the belief—the that it won’t take long to cut the line, that “ignorance.” The service doesn’t want land is more valuable than it looks—or the predicted weather front won’t be to appear ignorant. Nor do it’s crews. changes the action, and uses low priority that strong, that they can “hook the fire The price of creating this impression suppression tactics. Either change before the front passed” (Report, 1994, may be a loss in vigilance, learning, reduces the inconsistency. p. A5-53), that the crews are really on and wisdom. top of this job, and that more resources Let’s extend this scenario to South are coming (Report, 1994, p. A5-47). It is tempting in a world of boldness and Canyon and a key decision, the decision There is a grain of truth in all of those aggressive attacks, to conclude that made at 9:30 the morning of July 6 to explanations. But people also have a there is no place for doubt. But as cut a direct fireline, downhill (Report, stake in needing them to be true, since Thoele (1994) has suggested the best 1994, p. A4-6). What is noteworthy they reduce the tension associated with firefighters do not confuse risk with about this decision is that it involves a doing something they believe to be 45 Part 4 of 4

dangerous. The trouble is, they now locked into public irrevocable, volitional somewhat chilling connotation to the have a vested interest in not seeing actions, and had to justify those actions. personnel category, “Overhead.” warning signals. If they do notice these These justifications made them more Experience is unevenly distributed signals, then their whole sense of what committed to those actions, which led across the several activities at South is happening collapses. Listen again to them to persist longer in executing Canyon and does not always line up what Rhoades says: “My ditty bag those actions despite growing dangers. with authority. There are no clear contained a copy of standard fire orders Notice that the people who would be mechanisms to mobilize and focus and and watch situations. I considered spared from this process of escalation implement the experience that is looking at it, but didn’t. I knew we were would be those who were forced to cut scattered around. And finally, everyone violating too many to contemplate.” line (there is low choice), people who is accessing their experience under saw escape routes, (the action is increasing amounts of stress, which When people take public, irrevocable revocable) and people who did not means they are likely to fall back on actions for which they feel responsibility, express their views in public (the those habits and understandings they their mind set is to justify those actions decision is not linked to them as have overlearned (Weick, 1990, pp. and to assemble evidence that shows individuals). 576-577). Unfortunately, these may be the action makes sense (Ross & Staw, the very habits and understandings that 1986). They are not indifferent toward Levels of Experience. Earlier I are least relevant to the unique evidence that raises doubt about the mentioned that experience has both an conditions in South Canyon. action. Instead they avoid, discredit, upside and a downside. The upside is ignore, or minimize this contrary that it gives you more patterns that can There are at least three reasons we evidence and keep looking for positive be retrieved and matched with current need to tackle the issue of experience reasons that justify continuing the action. puzzles to make sense of them. The and how it is mobilized. First, an People who justify their actions persist, downside is that more experience can important finding from studies of high or in the words of the investigating team, sometimes lead to less openness to reliability organizations is that they have “strategy and tactics were not adjusted novel inputs and less updating of the multiple structures. Aircraft carriers, for to compensate for observed and models one uses. Failures to revise example, have a bureaucratic potential extreme fire behavior” (p. 35). often produce ugly surprises. hierarchical structure for normal functioning during slack times, a I have dwelt on this one decision at I want to dig deeper into the issue of different structure built around expertise South Canyon to show how people experience levels at South Canyon, for “high tempo” periods of extended justify their actions and in doing so, partly because the accident investiga- flight operations, and a third structure become more committed to continuing tion team seemed reluctant to do so. I explicitly designed for emergencies. those actions. There are several other say this because if you look at the Fire High tempo structures are especially discrepancies that could be analyzed Entrapment Investigation and Review relevant for wildland firefighting where the same way, such as the belief that Guidelines (Report, 1994, pp. A12-3 to rank in the formal hierarchy does not this was a low priority fire yet Type 1 A12-11) which they followed religiously always coincide with technical expertise. crews were put on it; the policy that two in structuring their report, the only LaPorte and Consolini (1991) describe or three trees burning is a standard category out of the 28 that they omitted a high tempo structure on carriers this smokejumper dispatch (French), yet was category 23, “V. Involved personnel way: “Contingencies may arise that jumpers were not dispatched profiles - Experience levels” (Report, threaten potential failures and increase immediately; the belief that this is a 1994, p. A12-7). This omission may be the risk of harm and loss of operational potentially serious fire, yet a crew walks due to the fact that, on paper everyone capacity. In the face of such surprises, off it the night of the 5th; the belief that is qualified. But just because they’re there is a need for rapid adjustment retardant works only at certain stages qualified on paper, doesn’t mean that that can only rarely be directed from of a fire, yet requests for it at that stage their experience is deployed well in this hierarchical levels that are removed are refused; aerial reconnaissance that incident or sufficient to handle its from the arena of operational problems. spots fingers of fire in west drainage on changing character or easily adapted to As would be expected, superiors have July 6, yet these are not drawn on the it. Issues of experience levels at South difficulty in comprehending enough map (Report, 1994, pp. 26, A5-70). My Canyon are complicated, difficult to about the technical or operational situa- point is not simply that there were untangle, and touchy when untangled. tion to intervene in a timely, confident discrepancies at South Canyon. Life is But that’s no reason to avoid them. way. In such times, organizational full of discrepancies and people manage norms dictate noninterference with to deal with them by sizing up pro and The overall level of relevant experience operators, who are expected to use con evidence. My point is that, key for leadership appears to be low. considerable discretion. discrepancies at South Canyon seemed Several people appear to be in over their to occur in a context where people got heads, which gives a whole new and

46 Part 4 of 4

Authority patterns shift to a basis of Requisite variety that is more adequate He starts with a sloppy hand-off the functional skill. Collegial authority (and can be illustrated by a crew of smoke- evening of July 5 and an unfinished decision) patterns overlay bureaucratic jumpers who have had prior experience project which he is unable to continue. ones as the tempo of operations as hotshots. Such a crew has the He’s dropped on unfamiliar terrain, at increases. Formal rank and status capability to function either in a more twilight, with rolling debris and steep declines as a reason for obedience. independent jumper mode or a more slopes. The crew is unable to get much Hierarchical rank defers to the technical disciplined hotshot mode, which gives sleep. The resources (two Type 1 crews) expertise often held by those of lower them a larger variety of ways to cope that Mackey requests the night of the 5th formal rank. Chiefs (senior noncom- with a larger variety of fire behaviors. arrive in small numbers at unpredictable missioned officers) advise commanders, intervals the next day (8 jumpers at gently direct lieutenants, and cow The notion of requisite variety also 10:00 a.m., 10 hotshots at 12:30 p.m., ensigns. Criticality, hazards, and sophis- alerts us to a hidden danger in another 10 hotshots at 3:00 p.m.) and tication of operations prompt a kind of successful firefighting. There is growing Mackey is not even sure they’ll come at functional discipline, a professionaliza- evidence that success leads to system all since he’s been told his fire is low tion of the work teams. Feedback and simplification (Miller, 1993), which means priority. When there is disagreement (sometimes conflictual) negotiations successful systems steadily become about building line direct and downhill, increase in importance; feedback about less sensitive to complex changes the incident commander does not “how goes it” is sought and valued” around them. This insensitivity resolve it and the hotshot superintendent (p.32). culminates in a sudden string of failures does not seem to question the strategy and the horrifying realization that one when he arrives around noon (Report, People in South Canyon did not seem to has become obsolete and faces a nasty, 1994 pp. A4-6, A4-7). have the capability to form a high tempo prolonged period of playing catch-up. structure where influence flowed from At some level Mackey knows the downhill expertise and experience, rather than Again, the lesson from high reliability strategy is risky because, in response to from the formal chain of command. In organizations such as the Diablo a flare-up at 10:35 AM, he begins to pull part, the problem was that it was never Canyon nuclear power plant is the need the crew out (Report, 1994 p. A5-70) clear where the relevant expertise was to cultivate diverse experiences, variety, only to have that decision questioned by located so that the structure could form multiple points of view, and conceptual Longanecker who suggests doing bucket around it. Furthermore, there was no slack (Schulman, 1993) so that people drops. The drops are made and the clear chain of command that could defer have a better sense of the complexity crew resumes cutting line. Not long to more experienced people nor was they face. And, there also need to be after this Rhoades observed that “Don there a clearly understood set of signals well-learned, trusted, procedures to looked terrible.” Still later, when the saw by which such a shift in structure could handle the inevitable conflicts that arise Rhoades is using breaks down, Mackey be conveyed immediately and when people make different interpreta- offered to sharpen it and help him cut unequivocally to everyone. tions, such as when a Fire Management line. This looks like a clear instance of Officer and a Hotshot superintendent a person falling back on overlearned A second reason the issue of experience differ on how the fire should be fought. behavior when that person is under is important is because it has the pressure. Mackey discards the less potential to create a smarter system that The third and final nuance of experience familiar activity of keeping your head up senses more. A key idea in system that I want to raise is the question of and supervising for the more familiar design is the notion of requisite variety: what happens when you are at the limits activity of keeping your head down and it takes a complex system to compre- of your experience where demands cutting line. hend a complex environment (Miller, exceed capabilities? And what can be 1993). Analyses of South Canyon that done about it? I mention this example to make the point are consistent with this principle have that when demands exceed capabilities, already begun to appear. For example, For the sake of illustration, let’s look at which is the basic condition under which Topic 3.5 in the IMRT review states that jumper Mackey who was jumper-in- people experience stress (McGrath, managers should “match qualified inci- charge at South Canyon and who had 1976), this is seldom simply the fault of dent commanders with the complexity just recently been given a permanent an individual. The buck doesn’t stop with of incidents” (Wildfire, Vol. 3, No. 4, Dec. appointment. What’s interesting and that person. Instead, the buck stops 1994, p. 46). That’s requisite variety. troubling about Mackey’s position is that everywhere (Allinson, 1993). The people Inadequate requisite variety occurs when the system makes it hard for him to do a around Mackey made his assignment a less complex incident commander, or good job on this fire. If we put ourselves harder and reduced his capabilities to a less complex jumper crew, or a less in Mackey’s shoes we discover that he handle it. The resulting pressure made complex dispatcher, cannot adequately is in a bad spot almost from the start. it harder for Mackey to gain access to comprehend a more complex event. the experience he already had, which

47 Part 4 of 4

increased pressure when his decisions Contributing Cause was listed as flaws work at a level of stress which is already were questioned, which gave him even in the decision making process. The quite high. It doesn’t take much less access to his experience until he implication of such an analysis is that additional stress before the quality of was caught in a vicious circle where he people should devote the brunt of their their judgment and thinking may begin did what he had always done on fires, energy to correcting the main cause. to suffer. Radio discipline is practically namely cut line rather than supervise. non-existent (Report, 1994, p. A5-37). The Hotshots had no idea something Allinson (1993, p. 111) among others Dispatch keeps reminding people that like this might be developing, and when has argued that the Challenger South Canyon is a low priority fire, that they saw Mackey, he seemed to be investigating team had their priorities there is nothing out of the ordinary moving around and checking, which is reversed. The failure was set in motion (LaDou statement), and they keep what overhead is supposed to do. by actions and choices that said it was saying “Roger” to all requests for safe to launch and by the decision to resources without any feedback as to The system let Mackey down. It did little launch itself. The defect in the O-ring how and when the request will be to remove or redistribute pressures, it can’t harm anyone as long as that handled, if at all. Requests for retardant did little to simplify his assignment, and defect stays on the ground. The fact are denied, weather briefings are it did little to monitor the fact that he and that a defective design even “existed at unevenly distributed, and no one takes others had less and less energy to cope all was the result of previous decisions responsibility for better distribution. The with growing complexities. The crew to select this design. That it was allowed IC is invisible (Report, 1994, p. A5-67) was losing variety and alertness, and no to continue to exist was the result of and there is no guidance for helicopter one spotted this or slowed the loss, or previous decisions not to alter it, despite use which means that people compete altered the work so that whatever repeated warnings. That it was allowed continuously (Byers statement) for its alertness remained was sufficient. to be in use in unsuitable weather services (Report, 1994, p. A5-22). conditions was also the result of These poorly integrated managerial Communication. In the preceding decisions made to allow it to operate decisions are spread over the period discussion of levels of experience, I despite the danger that the weather from July 2nd through the 6th and may steadily enlarged the size of the relevant conditions represented. [Allinson reflect even earlier decisions about organizational unit from jumpers and concludes by saying] It seems more safety and how people are to be treated. hotshots and South Canyon overhead, appropriate, then to describe the to the system in general including technical defect of the Challenger with The questionable decisions continue dispatcher, interagency coordinators, the term “proximate cause” and after the blowup, suggesting that the and top management. I did so in the management’s decision to launch the incident within the incident is hope that we would not fall into the trap Challenger without an adequate regard mishandled. Aircraft are kept circling of glibly saying that South Canyon is for safety, the ‘primary cause’ ” (p. 113). above the blowup for 45 minutes in 50 another instance of operator error, but knot winds (Ferneau statement). The would instead incorporate a larger, Since the South Canyon report focuses governor is allowed to tie up a key phone earlier, higher set of design decisions on the crews cutting line, it is difficult to connection for 15 minutes which delays as significant contributors to the spot earlier administrative decisions that rescue efforts for the people deployed incident. are potentially significant. But there is in shelters. Helicopter pilot Good, who certainly no shortage of possibilities. seems to have a stunning amount of You may recall that the team Crews at South Canyon are told to be endurance and resilience, is still being investigating South Canyon felt that aggressive but are given little support ordered around at 9:00 at night, this time “Management support and dispatch to do so and later are faulted for being to fly body bags in. He speaks for a coordination” were not “significant too aggressive. Prineville Hotshots are larger group when he refuses, saying, contributors” to the disaster, but merely requested and then treated poorly when “I’ve had enough” (Report, 1994, p. A5- “influenced” it (Report, 1994, p. 33). I they arrive at the Glenwood Springs 50). mention this partly because not office at 8:00 a.m. on the 6th (Report, everyone agrees with this assessment 1994, p. A5-80) where they are forced These are all symptoms of problems far (e.g., OHSA, IMRT), and partly because to look around to find tools and then go removed from the crew boss on the this is the same kind of questionable to the 7-11 to get food. Their ground, and our job is to diagnose assessment that was made by the team understandable agitation at being symptoms of what. Many would say investigating the Challenger disaster. handled this way probably does not these are symptoms of problems in In the Challenger report, the Main disappear the moment they get to communication. This is what the Cause of the disaster was listed as South Canyon. Instead, much like the Hotshots said: “The crew wants to know “failure in the joint between the 2 lower married pilot who takes command of an where the communications broke down segments of the right solid rocket motor” airplane shortly after an intense with the red flag warning” (Report, 1994, (Allinson, 1993, p. 111) and the domestic quarrel, the crew starts their p. A5-81). The answer to their question 48 Part 4 of 4 about a breakdown is that the commun- treated as a positive message that not desired, or would not be passed on. ications broke down everywhere, which things are OK. Notice, that if things are They may not have asked because is an inevitable diagnosis when you not OK and people are preoccupied and they thought they had all the answers argue that the buck stops everywhere. unable to send a message, this too will or wouldn’t get them anyway. And they result in a failure to report. may not have passed on information Here’s what good communication looks because they assumed it would not like. The example comes from Thus, no reporting can mean either receive a hearing. If any of these Winston Churchill. When he discovered things are OK or things are not OK. The possibilities are true, and if people also to his horror that Singapore was Zebrugge Ferry disaster on March 6, believe that no news is good news, vulnerable to a Japanese land invasion 1987 involved this very misunder- then wildland firefighting is a thousand during WWII, Churchill said, “I ought to standing. The person responsible for administrative accidents waiting to have known. My advisers ought to have closing the bow doors of the ferry did happen and is even more dangerous known and I ought to have been told and not report any deficiency to the Captain, than people realize. Fire is not the I ought to have asked” (Allinson, 1993, not because there was none, but problem. The problems are alertness, p. 11). Notice how much complexity because he had fallen asleep before trust, trustworthiness, respect, candor, Churchill has described. There is no one closing the doors. The Captain steamed and “the will to communicate” (Allinson, cause for this disaster. Churchill could into the channel unaware that the doors 1993, p. 41), a list that fits Mann Gulch have known. Others should have known. were open and water was flowing into as much as it fits South Canyon. The Those who should have known, should the vessel. Five minutes after leaving difference is that in South Canyon, the have informed Churchill without his the coast of Zebrugge, the ferry Herold list applies to a more dispersed set of asking. If others did not know, they of Free Enterprise capsized, sank, and people with a more diverse set of should have found out and informed 193 lives were lost. The buck stops interdependent tasks. Churchill on their own without waiting for everywhere on this incident. Virtually him to ask. If they didn’t know, Churchill, the same scenario happened 5 years Safety attitudes are inherent in good by inquiring of them, might have prodded earlier on October 29, 1983 aboard the management practice rather than them to find out. If they had known but ferry Pride, but was caught before the something that are tacked on. Free flow failed to speak up, Churchill, by inquiring, ship capsized. At that time, the Master of information is good management may have been given the necessary urgently communicated with practice, gets things done, and saves information. Any of these eventualities management requesting that there be lives. If people fail to pass along might have changed the course of some indication on the bridge whether information, fail to listen attentively, and events (adapted from Allinson 1993, the watertight doors were closed or not fail to elicit information actively, that is pp. 11-12). (Allinson, 1993, p. 203). Management bad management and unsafe did not listen. Their responses to this management. I suspect Stephen Pyne It is everyone’s responsibility to request are preserved in the accident (1984, p. 394) has it about right when challenge and to respond to the investigation and included remarks he said that “All too often ‘safety’ is a challenges in a trustworthy manner, and such as, “Nice, but don’t we already cosmetic, a mandated and barely to listen carefully and respectfully to the pay someone?”; “Assume the guy who tolerated veneer of declarations, response. When people fail to engage shuts the doors tells the bridge if there memorandums, task force reports, in respectful interactions (Weick, 1993, is a problem”; “My goodness.” People safety officers, and exhortations that pp. 642-644), things can get dangerous. at the top didn’t feel it was part of their has little relevance to the conduct of Let me suggest why that happens. job to inquire, or to listen attentively, or practical affairs. Something is taught to pass along information. So there is as a ‘safe’ procedure rather than the One possibility in wildland firefighting is no reason for the Masters’ of the vessels only procedure. Safety is something that a norm has developed which says to act differently if this is the preferred added to a program, not something essentially, no news is good news. communication style at the Peninsula integral to it . . . Most safety programs Partly because people on crews are and Oriental Steam Navigation fail at the bottom because they are not independent, adventuresome, take- Company (Allinson, 1993, p. 195). truly practiced at the top.” charge people; partly because radio traffic is so hard to control; partly People associated with South Canyon Watch Outs for Administrators. In the because there are no detailed and didn’t know a lot of things they should context of a closer look at administrators, systematic communication protocols for have known. This raises at least 3 it makes sense to look at the 10 fire dispatchers and crew leaders to questions: why weren’t they told, why orders and 18 Watch Outs that are exchange information about changes in didn’t they ask, and why didn’t they tell potential guidelines for firefighters, fire status; and partly because people what they knew? They may not have guidelines that remained on a card presume the basic task itself is been told because others thought the inside Rhoades’ ditty bag, untouched straightforward, a failure to report is information would have no effect, was and unread. I want to make two points 49 Part 4 of 4

about these two lists. First, I think is to maintain the big picture. The most Safe areas for administrators are firefighters should begin to compile a list effective aircraft cockpit crews are those created by such things as clear norms of Watch Outs for administrators. In the in which, during an emergency, the about the relationship between failure same way that the current 18 Watch aircraft is flown by the first officer (co- and learning, secret ballots, anonymous Outs alert crews to increased hazards pilot) not the captain and the captain reporting of near misses, access to at the site of the fire itself, administrator plans how to deal with the emergency brainstorming where evaluation of ideas Watch Outs would alert crews to and tracks progress. is intentionally suspended, the conditions back at headquarters that equivalent of a penalty box where are just as hazardous as the fire itself. Although, I have already discussed people who commit glaring errors are Recall that Longanecker (Report, 1994, communication, a good way to put for a finite period of time after which p. A5-54) proposed just such a watch illustrate it is by a surprising finding in they rejoin the action, and availability of out in his statement after South Canyon: studies of captains who lead the best 3rd parties to mediate conflicts that are Watch out “when you don’t receive the aircraft crews. Investigators found that difficult to resolve. It is the very resources that you need or you are these captains readily acknowledge that availability of these safe areas that debating with the dispatcher about the their decision making ability is not as allows administrators to act in a candid resources you need.” A handful of other good in times of emergency as it is at manner that can then be mirrored on Watch Outs might include, Watch out, other times (Helmreich, Foushee, the fireline. Benson, & Russini, 1986). Captains 1. When the governor is in town (Report, who are the worst leaders, say that If a firecrew sees that management is 1994, p. AF-64); their decision making ability is just as violating its own version of LCES, they 2. When interagency ties are strained good in time of emergency as it is at should be just as wary and alert as if (Report, 1994, p. A5-63); other times. Poor leaders don’t listen they saw themselves violating LCES at 3. When dispatchers keep track of things because they don’t think they need to. the fire itself. The dangers, in either in their head rather than on paper; Good leaders don’t fall into that trap. case, are real, immediate, and serious. 4. When the norms for radio discipline Recall an earlier point I made that a are loose (Report, 1994, p. A5-37); potential trap when people gain 5. When people are reluctant to ask for experience is that they lose openness help; to new information. Here we see clear Moving Toward Solutions 6 When administrators are getting on- evidence that good pilots—and by the-job training; extension, good leaders in general— My analysis so far has been largely 7. When administrators say “keep it don’t let that happen. speculative and has consisted of simple;” extrapolations from what is known 8. When the overhead is tough to find Escape routes for administrators about high reliability organizations to (Caballero statement); and consist of things like options, revocable seemingly analogous circumstances in 9. When you don’t know which office to actions, pulling the plug, seeing the South Canyon. Given the tentative report to, you think about it, and temptation to escalate a commitment to quality of this diagnosis, it is premature having thought about it you then go salvage a losing cause and then to talk about remedies. Nevertheless, to the wrong one (Taft statement). avoiding it. The scary thing about remedies have already been implied in administrative escape routes, is that what I’ve said and I want to illustrate The second point I want to make is that sometimes they are used to deny briefly some directions in which those a good place to start in developing a list individual responsibility and to pass the implications point. of administrative Watch Outs is with buck. That’s the mind set that we want existing efforts to boil the ten fire orders to undercut with a culture where the 1. If leadership is an issue, then it down to the acronym LCES (Gleason, buck stops everywhere. Managers seems important to look more closely at 1991). If lookouts, communication, responsible for treating people with the possible pathways by which one escape routes, and safe areas, are good respect need to have the welfare of can become a smokejumper foreman, enough for firefighters, they are good those people in mind and not just their whether the route is through expertise enough for administrators. The own reputation, when they vow never to with parachutes, or with leadership, or principles are essentially the same in get into anything without having a way with fires. Depending on which route is either case. For example, the out for everyone. Safe flight operations favored, people in the field could have administrative counterpart of lookouts is on aircraft carriers are made possible very different habits they fall back on a person with the big picture. In nuclear because that’s precisely what managers when put under pressure. power plant control rooms, there is a believe and put into practice (Weick & person called the shift foreman (Weick, Roberts, 1993). 1987, p. 116) whose sole responsibility

50 Part 4 of 4

2. If people feel there are too many rules it possible for live cases to be made a investigation team, p. 3), enough binding on firefighters (10 fire orders + regular part of training. Our research on violations that Rhoades was scared to 18 Watch Out situations + 4 LCES + 4 socialization of newcomers on aircraft count them, yet Ryerson is quoted in the common denominators + 3 sources of carriers suggests that old hands who Wall Street Journal (8-22-94) as saying judgment error [ignorance, casualness, tell war stories are an invaluable source “it happened fast enough that none of us distraction] + 9 guidelines for indirect/ of training. Remember, we’re talking knew we were in danger . . . It happened downhill line construction = 48, in about organizations in which it is hard to in a matter of seconds” (Page A1, Gleason, 1994, pp. 24-25), and if fire- learn by trial and error. The next error column 1) and Blanco called dispatch fighters say they need to violate orders may be the last trial. If trial and error shortly before the blowup “and told to keep fires from growing (Rhoades, learning is limited, then case studies them that things looked good” (Report, 1994, p. 22), then clearly some priority become very important. 1994, p. A5-11)? I realize that Ryerson setting is in order. It is here where I probably means the blowup itself think it makes sense to talk about simul- 5. I think there is a key training lesson happened fast, yet conditions had been taneous centralization and decentraliza- in the recent experience with airline steadily worsening and the blowup was tion. What you want to do is centralize training in cockpit crew management. not the first moment people sensed everyone in terms of 3 or 4 key values This training didn’t have much effect or danger. People either weren’t keeping which are treated as non-discretionary credibility until the people being trained score of the number of violations, or and imperative (LCES?), and were put in flight simulators where they didn’t want to know the score, or decentralize the others issues so that solved in-flight problems and were because they arrived at different times they serve as guidelines and a platform video-taped doing so (Helmreich and with different information had a different for improvisation to meet unanticipated Foushee, 1993, p. 28). Pilots saw sense of the number of violations. local conditions. I have no idea what the themselves actually committing the final partitioning of Gleason’s 48 guides errors that up to then, had only been A different set of questions concerns would look like. I do know that discus- described in dry classroom lectures. the role that groups play. Why didn’t sions to hammer out such a partitioning And what may have been most crucial the Prineville Hotshots speed up, look would strengthen the will to communicate. in this Line Oriented Flight Training is back, drop their tools? Perhaps they that each videotape was erased didn’t think they were in great danger. 3. I would pay close attention to what immediately after the performance had The fire could have burned straight people overlearn during their training, been critiqued. Videotapes of crew uphill toward the lunchspot. But what since this is what they are most likely to interaction during fires, of dispatchers we may also be seeing here is the flip do when put under pressure. For allocating scarce resources, or of side of what I think happened at Mann example, the 23 people (Report, 1994, administrators briefing local property Gulch. At Mann Gulch the group p.14), who fled from the ridgeline did owners, all could prove to be a valuable disintegrated, which led to a loss of not take the shorter, safer, more direct window on just how well the struggle for meaning and then to something route used by Haugh, Hipke, and alertness is being waged. approximating panic. At South Canyon Erickson, but instead ran out the same the group remained together (Report, way they had hiked in, which exposed I know these are all small solutions to 1994, p. A4-10) and things stayed them to more danger for a longer period. potentially big problems. But they are a meaningful, but people held onto the If firefighters haven’t practiced and over- start, they can be done in parallel, they wrong meaning. Imagine what a typical learned shelter deployment, or dropping can be done simultaneously in different hotshot might be thinking. Erickson their tools, or using a checklist, or places, and they may stimulate a better and Haugh are strangers and jumpers watching out for the safety of a buddy, set of starting points. to boot; they are saying “run,” but this or running from fire as fast as possible has been a sloppy operation from the (Maclean, 1992, p. 272) over and over, start. Furthermore, we didn’t hear then it’s a safe bet they won’t do those anything about a weather front nor did things either when they are under Lingering Questions we hear the argument about cutting intense pressure. direct line downhill, so presumably we’re Even though I have some hunches safe and they’re probably exaggerating. 4. I think Dave Thomas (1994, pp. 45- about what might have been going on 48) is right in his insistence that fire in South Canyon, there are some It may be that group ties were too tight stories and case studies are a crucial questions that continue to baffle me. among the hotshots, the level of concern means to extend people’s repertoire of For example, how is it possible that so was too low, and the meaning persisted, experience, even if that experience is many fire orders and Watch Outs were like it did at Mann Gulch, that this is just second-hand. There certainly are being violated (20/28 were violated one more 10:00 fire. enough “old fire dogs” around to make according to the South Canyon

51 Part 4 of 4

Perhaps there is such a thing as a group well-spaced, steady march up the function of deflecting blame from admini- being too disciplined and too cohesive. fireline toward the ridge with tools in strators onto crews, and are only inci- High cohesion wards off panic, but it hand, may represent the behavior of a dentally relevant to safe practice. With also encourages groupthink and wards group under enormous pressure rather this many guidelines in place, it’s fairly off more disturbing and more varied than that of a group that is relatively easy after the fact for administrators to meanings of what may be happening. calm and thinks this is just another fire, spot at least one violation that occurred Variety may have been crucial to albeit one that has been has been and to spotlight it as THE cause of the surviving this incident. The 12 people managed a bit more poorly than usual. accident. climbing up the fireline toward the ridge all did the same thing and perished. A further puzzle at South Canyon My point here is not to be cynical. The other 37 people on the mountain concerns the possibility that this fire fell Instead, I want to raise the possibility did different things, most of which in a kind of “no man’s land” at a crucial that the system may know less about worked. Three ran to the top of the period. Jumpers who dropped on the firefighting than it thinks it does. The fireline; 8 ran above the lunch spot and fire the night of August 5th found a fire multiple guidelines give the impression deployed shelters; 1 stayed at the that seemed larger than an initial attack that much is known, but the guidelines lunchspot; 23 headed for Helitack 2 but fire for which they are experts. When may be redundant, they may say the then stopped and went down various the shots began trickling in around same thing in several different ways. portions of the east drainage; and all of noon on August 6th, they found a fire The result may be that when people these people lived. Two people tried to that seemed smaller than fires for take these guidelines seriously they make it to Helitack 2, but failed. To put which they are experts. The result is a reduce their ability to sense subtle it in the most extreme form, the fuzzy situation where the fire is too big variations in fire behavior and therefore hotshots didn’t panic and that may have for some, too small for others, and too undertake more dangerous actions. been their problem. If they had come foreign to the experience of the people The guidelines may shield management, closer to doing so they might have in charge. The problem may not be but they also may create blindspots for lived. I know how bizarre that sounds. that a transition was mishandled and firefighters. I think that possibility needs But it’s important to realize that we are resulted in fatalities. Rather, the to be explored carefully. dealing with strong, competing, human problem at South Canyon may have tendencies toward independence and been that the complexity of the fire fell If it turns out that the 48 guidelines say conformity. That lies behind respectful outside the scope of everyone who just a handful of different things and interaction. People need sufficient social tried to control it. If that’s plausible then anticipate a relatively limited set of support to stay calm and sufficient it suggests the need for rethinking the variations in fire behavior, then efforts independence to be innovative. People adequacy of existing fire categories and should be made to develop a more who fight wildland fires aren’t freed from their matchup with training and comprehensive, more varied set of this dilemma simply because they are expertise. Problems may occur not guides. If it turns out that all 48 are bold. As long as crews and danger and only when fires move from one different, varied, and necessary, then it different experiences mix together, we category to another, but also when they would seem important either to prioritize can expect puzzling outcomes. defy categorization in the first place. them as mentioned earlier, or divide up responsibility for them among the crew. Notice that we can take a totally different As a final lingering question, I wonder if If there are 48 guidelines and 10 crew, approach in analyzing the Hotshots’ 48 guidelines might be too few then each crew member would be behavior. Earlier, I argued that because guidelines to be of much help to assigned 5 guidelines to monitor, they were poorly treated in Glenwood firefighters? There seems to be lots of champion, and communicate. Springs, they may have been under overlap and similarity among the some stress when they got to South guidelines, so much so in fact that if we Canyon. If, in addition, they had doubts study them closely, we might discover about the safety of what they were that they have too little variety to match Conclusion doing, then the level of stress might have the large amount of variety in wildland been quite high when they were ordered fires. If that were possible, then it Something that both Mann Gulch and to retreat to the ridge. If, during hotshot would explain why firefighters feel they South Canyon share in common is a training, people overlearn paramilitary have to violate orders. They do so to series of events in which something very discipline, regimentation, and obedience, regain the variety of attack they feel is small escalated into something mon- then we would expect this pattern of necessary to combat the variety in the strous. A good example of two events discipline to be especially visible under fire they face. The possibility that 48 that can be caught in an escalating spiral high stress. The general idea is that guidelines actually reduce requisite that starts small and ends monstrous are when stress increases, people fall back variety is also consistent with the idea the events of “fear” and “understanding.” on overlearned habits. Thus, the brisk, that these guidelines may serve the As fear increases, understanding 52 Part 4 of 4 decreases, which causes fear to Helmreich, R. L., Foushee, H. C., Rumsey, W. (1949). Testimony. Mann increase even more, which leads to Benson, R., & Russini, W. (1986). Gulch Transcript (pp. 97-109). even less understanding, and this Cockpit management attitudes: Sagan, S. D. (1993). The limits of escalation increases until something Exploring the attitude-performance safety. Princeton University Press. explodes. That could be what happens linkage. Aviation, Space, and Salancik, G. R. (1977). Commitment as people discuss how to prevent more Environmental Medicine, 57, 1198- and the control of organizational South Canyons. But if the discussion 1200. behavior and belief. In B. M. Staw leads to more understanding, then we Maclean, N. (1992). Young men and and G. R. Salancik (Eds.), New create a world where more fear leads to fire. Chicago: University of Chicago directions in organization behavior. more discussion which leads to more Press. (pp 1-54). Chicago: St Clair. understanding which leads to less fear. March, J. G., Sproull, L. S., & Tomuz, Sallee, R. (1949). Testimony. Mann My remarks should be understood as M. (1991). Learning from samples Gulch Transcript (pp. 69-89). an invitation to discussions that improve of one or fewer. Organization Washington, D.C.: U.S. Forest our understanding and lessen our fears. Science, 2, 1-13. Service. McGrath, J. E. (1976). Stress and Schulman, P. R. (1993). The behavior in organizations. In M. D. negotiated order of organizational Dunnette (Ed.), Handbook in reliability. Administration and References industrial and organizational Society, 25, 353-372. psychology (pp. 1351-1395). Thoele, M. (1994). Firefighters Allinson, R. E. (1993). Global Chicago: Rand-McNally. emphasize safety, but. Wildfire, 3 disasters. NY: Prentice-Hall. Miller, D. (1993). The architecture of (3), 27-29. Cooley, E. (1984). Trimotor and trail. simplicity. Academy of Thol, H. J. (1949). Testimony. Mann Missoula, Mont.: Missoula Press. Management Review, 18, 116-138. Gulch Transcript (pp. 183-202). Dodge, W. (1949). Testimony. Mann O’Reilly, C. A., & Caldwell, D. F. Washington, D.C.: U.S. Forest Gulch Transcript (pp. 117-125). (1981). The commitment and job Service. Service Washington, DC: US Forest tenure of new employees: Some Thomas, D. (1994). A case for fire Service. evidence of postdecisional behavior case studies. Wildfire, 3 Fite, F. (1949). Testimony. Mann justification. Administrative Science (3), 45-47. Gulch Transcript (pp. 26-32). Quarterly, 26, 597-616. Weick, K. E. (1987). Organizational Washington, DC: US Forest Putnam, T. (1994). Analysis of escape culture as a source of high Service. efforts and personal protective reliability. California Management Gleason, P. (1991). LCES-A key to equipment on the South Canyon Review, 29 (2), 112-127. safety in the wildland fire fire. USDA Forest Service, Weick, K. E. (1990). The vulnerable environment. Fire Management Missoula Technology and system: An analysis of the Tenerife Notes, 52 (4), 9. Development Center. air disaster. Journal of Gleason, P. (1994). Unprepared for the Pyne, S. J. (1984). Introduction to Management, 16, 571-593. worst case scenario. Wildfire, 3 (3), wildland fire. NY: Wiley. Report of Weick, K. E. (1993). The collapse of 23-26. the South Canyon Fire Accident sensemaking in organizations:The Helmreich, R. L., & Foushee, H.C. Investigation Team. August 17, Mann Gulch disaster. (1993). Why crew resource 1994. Administrative Science Quarterly, management? Empirical and Rhoades, Q. (1994). Effective fire 38, 628-652. theoretical bases of human factors fighting calls for bending the rules Weick, K. E., & Roberts, K. (1993). training in aviation. In sometimes. Wildfire, 3 (3), 22. Collective mind in organizations: E. Weiner, B. G. Kanki, & R.L. Ross, J., & Staw, B. M. (1986). Expo Heedful interrelating on flight decks. Helmreich (Eds.), Cockpit resource 86: An escalation prototype. Administrative Science Quarterly, management. (pp. 3-45). San Administrative Science Quarterly, 38, 357-381. Diego: Academic Press. 31:274-297.

53 Appendix E—Related Reports Part 4 of 4

The Collapse of Decisionmaking— People are not always aware of which factors dominate their decision process. Decisionmaking and A Telling Model Although we say “safety first,” this does Organizational Structure not mean it’s necessarily first in actual on Storm King Mountain Human thinking and decisionmaking decisions. Also, people are seldom have been studied and modeled. The aware of the few factors they actually are Ted Putnam, Ph.D., Protective Clothing and decision process is essentially additive: processing, so they tend to be overcon- Equipment Specialist, Missoula Technology A+B+C. For example, a decision to build fident in their decisionmaking ability. and Development Center, May 1996 fireline may be characterized by fire- Although people are unable to use all the fighters (FFa, FFb, FFc, FFd) basing available information for decisionmaking, Stress, fear, and panic predictably lead their choice on these factors: especially when under stress, computers to the collapse of clear thinking and have no such limitation. Computers organizational structure. While these FB—fire behavior process information interactively, AxBxC, psychological and social processes have W—weather and can use most of the available infor- been well studied by the military and the FL—fuels mation for better decisions. People are aircraft industry (Cockpit Resource E—equipment very good at determining the state of Management) (Weick 1990 and Wiener, P—personnel, experience, skill each factor, the inputs, but not so good Kanki, and Helmrich 1993), the wildland S—safety at integrating all the factors to make a fire community has not supported similar M—expectations of management decision. While computers are of help to research for the fireline. The fatal wild- incident management teams, normally land fire entrapments of recent memory Numerous studies show no matter how they aren’t available for extended initial have a tragic common denominator— many factors are important, the human attack. human error. The lesson is clear: study- mind normally can handle only about ing the human side of fatal wildland fire seven factors (e.g., seven-digit telephone So when fireline conditions are routine, accidents is overdue. numbers). People differ both as to how most people would reach similar deci- many factors they use and the value sions because they are more aware and Historically, wildland fire fatality investi- placed on these factors. In this modeling, take more information into account. gations focus on external factors like fire the first factor is the one each firefighter When fireline conditions worsen, behavior, fuels, weather, and equipment. pays the most attention to with the other decisions are more at the mercy of the Human and organizational failures are factors added in decreasing level of one or two factors individuals are still seldom discussed. When individual importance. processing and their level of experience. firefighters and support personnel are In the example above, under stressful singled out, it’s often to fix blame in the So the decisionmaking processing conditions even though each firefighter’s same way we blame fire behavior or leading to fireline building could be main factors differ, if they readily fuels. This is wrong-headed and danger- modeled: communicate as a crew, most of the ous, because it ignores what I think is an factors are still present. Although underlying cause of firefighter deaths— FFa = M+W+FB+S+P+E+FL individual decisions are additive, where the difficulty individuals have to consist- FFb = S+P+M+FB good communications exist, the group ently make good decisions under stress. FFc = FB+P+E decision can approach the better inter- FFd = P+E+S+FB+W active process. There’s no question individuals must be held accountable for their performance. Although their decisions were the same, Studies also show that our linear thinking But the fire community must begin they arrived at them through quite tends to underestimate hazards, particu- determining at psychological and social different factor evaluations. larly if the hazard is increasing at a levels why failures occur. The goal logarithmic or exponential rate as can should not be to fix blame. Rather, it However, in situations that create stress, happen on the fireline. An example should be to give people a better under- fear, and panic, minds regress toward would be estimating rates of fire spread. standing of how stress, fear, and panic simpler, more habitual thinking. This A computer would give the better combine to erode rational thinking and regression could be modeled: decision in a heartbeat. People would how to counter this process. Over the tend to underestimate the rate of spread years, we’ve made substantial progress FFa = M+W (Get the work done, weather and have difficulty deciding on an in modeling and understanding the permitting) appropriate course of action. And so it external factors in wildland fire suppres- FFb = S (Safety first) is important to understand the limits of sion, and too little in improving thinking, FFc = FB (Fire behavior most important) how we process information and the leadership, and crew interactions. FFd = P+E (People and equipment common types of errors that can occur. dominant)

54 Part 4 of 4

Leadership and Group Risk-Taking in Collapse of Behavior Wildland Firefighting Decisionmaking on Storm King Mountain Stress, fear, and panic take their toll at First, wildland fires cannot be fought all levels of the wildland firefighting without risk. Making decisions while at On the South Canyon Fire the first organization. Under stress, leadership risk assumes firefighters can evaluate decision failures occurred at the BLM becomes more dogmatic and self- the likelihoods of various states of nature. (Bureau of Land Management) district centered. It regresses toward more On larger fires, with structured incident level. Although the fire started July 2 in habituated behavior. Groups tend to management teams (IMT), specialists, a fire exclusion zone, resources did not fragment under stress into smaller units and portable weather stations, etc., the reach the fire until July 5. It was the or to stick together and follow their leader likelihoods are more objective and out- worst fire season in years and local without joining the decisionmaking pro- comes are better predicted. An excellent resources were stressed. Holding costs cess. Either way, most of the information study of leadership under stress on a down and making do with local available for the best decisions is not larger (IMT) scale is Taynor, Klein, and resources dominated decisionmaking. utilized. Thordsen’s 1987 article, Distributed From our earlier analysis, we can predict Decisionmaking in Wildland Firefight- a tendency to fall back on habituated An extensive 12-year study of Forest ing. They describe the IMT as a very tactics, such as letting the fire go until a Service field crews conducted by sociol- robust organization due to lengthy local crew is available. Although many ogist Jon Driessen (1990) showed there experience levels, the common crews were available nationally, the is an inverse correlation between crew experience of working together, district did not request help until July 5. cohesion and accident rates. The study excellent communication structure, and The longer initial attack was delayed, the also identified factors fostering cohesion. well-defined, well-practiced roles. In greater the risk the firefighters faced. Driessen found it takes about 6 weeks contrast, on smaller fires, the likelihoods for good crew cohesion to take effect. are more subjective, based on skill and An incident commander (IC) from the So firefighting crews are predisposed experience rather than instruments. local BLM district arrived on the fire the toward accidents until they become When small fires grow larger and more morning of July 5. But because of cohesive units. Unfortunately, this type complex, such subjective estimates mechanical problems with their chain of information is not normally considered become less accurate, and decision- saws, the IC and crew left the fire that even when sending crews to riskier fires. making regresses to a reliance on fewer evening as a load of smokejumpers were and fewer factors. The result is a failure dropped onto a nearby ridge. The first An excellent case study of leadership to keep up with rapidly changing condi- person out the door of the jumper aircraft under stress on a smaller scale is Dr. tions, and people on the fireline are put became the jumper-in-charge (JIC). Via Karl E. Weick’s The Collapse of Sense- at greater risk. radio the IC turned the fire over to the making in Organizations: The Mann JIC. This situation raises two immediate Gulch Disaster (Weick 1993). Although Second, risk-taking is subject to leadership questions: Why did the IC the leadership and organizational struc- perceived and actual rewards and leave the fire? Was first experienced ture discussed are based on Norman punishments. When we attach a stigma person out the door the best way to Maclean’s Young Men and Fire, Weick’s to deploying a fire shelter, we bias fire- choose the JIC? analysis is thought-provoking. It is also fighters into taking more risks to escape. haunting because the South Canyon If there’s a stigma associated with The jumpers fought the fire most of the Fire Investigation report shows the dropping packs and tools, firefighters will night as it continued to grow in size. In human and organizational failures on carry everything while trying to outrun a response, the JIC ordered two more Storm King Mountain are similar to those fire. If a stigma is attached to abandoning Type I crews. The IC returned with his he hypothesizes happened at Mann a fire or the fireline, firefighters will take crew the morning of July 6. By 10:30 Gulch 45 years earlier. more risks to control a fire. The various a.m., a second load of jumpers arrived, payoffs associated with risk-taking are and the JIC of that plane load became not necessarily those managers claim the line scout (LS). The IC and his crew are operating. We need professionals stayed on top of the ridge building specializing in the study of decisionmak- fireline, while the jumpers began ing under stress to interview managers constructing fireline downhill on the west and firefighters, so we can begin to flank. At 12:30 p.m., 10 members of the better understand actual risk-taking on Prineville Hotshots (PHS), including the fireline.

55 Part 4 of 4 their superintendent, arrived at the fire. • Decisions questioned. higher number of fatalities (Putnam The IC, JIC, and PHS superintendent 1994). agreed to send 9 PHS down to help • Most experienced people not consulted build fireline on the west flank. At 3:00 and locked out of decision process. • Although firefighters knew what fire p.m., the remaining 10 PHS arrived at shelters were and how to open them, the fire and stayed on top of the ridge • Poor communication concerning they clearly did not know how to use with their superintendent to help the IC deteriorating conditions—especially them effectively or where they would and his local crew. among groups. work best.

So the organization structure before the • Continued fragmentation into smaller blowup was: groups.

Location Local National • Decreased talking within groups. Resources Resources Training to Make Ridgetop 9 BLM District 11 PHS • Failure to integrate vital, available Decisions Under Stress 2 USFS District 2 Helitack information when changes occurred. Courses such as Cockpit Resource West flank None 9 PHS • Failure to act on the weight of the Management train crews to counteract 8 Missoula SJ 4 McCall SJ evidence. the natural tendencies for behavioral 2 North Cascades SJ regression. Countermeasures mentioned 1 West Yellowstone SJ 1 Grangeville SJ • Underestimating the current and by Weick and others include: potential fire behavior. • Nonstop communication, both verbal All the ingredients were in place for a Once the blowup occurred, in the ensu- and nonverbal is crucial, especially catastrophe. Three local crews (BLM, ing stress, fear, and panic, people’s when people first come together. USFS, Helitack), the Prineville crew split actions followed classic lines of into two groups, and jumpers from five regressing to more habituated patterns • Survival goals (threat recognition, different bases led by two somewhat of behavior: escape, shelter use) must be over- randomly selected JIC’s were thrown learned through repeated practice or together and asked to perform as a team • On the ridgetop all but two people ran they will not be dominant in dangerous under increasingly unstable conditions. out the east drainage, a potential death situations. Neither leadership roles nor a cohesive trap. This was not a matter of thought organizational structure stabilized before as much as regression—going back • Cross-train in roles. the blowup. the way you had come in. • Value wisdom and openness. On the west flank, a group of nine • The two helitack refused to go into the smokejumpers split off to construct fire- east drainage and ran back along the • Initiate respectful face-to-face line to the southwest, forming a third ridge they had been dropped off on, encounters between crew members group. These three groups began to possibly looking for a copter pickup site. and between crews. focus on their own immediate problems and communications among them • The west flank SJ and PHS went back • Remain curious and observant. continued to decline. As the wind picked up the fireline they had been digging. up after 3:00 p.m., so did fire activity and • If things don’t make sense, speak up. firefighter stress levels. And, predictably, • Virtually all the escaping firefighters decisionmaking and organization col- carried their tools and packs even • Avoid overconfidence and overcau- lapsed inward, with fatal consequences. though it cost many of them their lives tiousness. (Putnam, 1994). From the South Canyon Fire Investi- • When situations deteriorate, pay more gation report and witness testimony, we • Even when the firefighters were yelled attention to leadership, perceptions, can find signs of collapse similar to those at to drop their tools and equipment, and group interactions. Strengthen ties. Weick identified in his analysis of Mann they did not. This deeply ingrained Gulch, including: response pattern resulted in fatalities. • Group dynamics before a crisis affect survival during a crisis. • Leadership questioned and challenged • Even though their lives were at stake, (for incident commander, jumper-in- very few firefighters made any attempt • Expect everyone to work safely, com- charge, and line scout). to use their fire shelters, resulting in a municate effectively, and cooperate. 56 Part 4 of 4

• Talk to other crew members and crews. LCES) in crisis situations. To link the into question the very process and struc- Expect them to talk to you—then listen. human factors involved in firefighting to ture by which we investigate fatalities and the classic Look Up, Look Down, Look communicate the results to the fire com- • Be especially wary of accepting incre- Around, we can add Look Inside. And munity. We can and ought to do better. ments of worsening conditions. It is we could change LCES to I-LCES, deceptive to accept the increments where the “I” means Inside, Inner, and rather than the entire change. Interpersonal.

It is apparent from this list that to be Patrick Withen, a smokejumper and adequately prepared requires training, sociologist, has discussed firefighter Discussion overlearning, and using these skills attitudes and has pointed out (Withen, routinely before a crisis strikes. It is also 1994) that there is no way to “just say There is no intent here to blame the clear these skills are a necessary no” in firefighting that doesn’t carry formal individual firefighters and managers for prerequisite for effective decisionmaking or informal sanctions. The onus is on the what they did or didn’t do related to the concerning integrating fire behavior, individual firefighter—not management— fire on Storm King Mountain. The real weather, fuels, equipment, and human to justify the decision. Routinely, there is issue is that we are not preparing our factors. a stigma attached to leaving the fireline. firefighters and managers to operate with maximal effectiveness under While looking at the firefighter from psy- known stressful, risky conditions. The chological and sociological perspectives processes and papers cited, when is encouraging, this idea has not been considered in the light of the South well received by many in the wildland Canyon Fire Investigation report, clearly A Start fire community. When suggested to the demonstrate that an almost automatic South Canyon Fire Investigation Team collapse of decisionmaking and organi- Within the wildland fire agencies, aware- and the follow-up Review Board as a zational structure occurred. It should also ness is growing about the value of possible causal factor, the suggestion be clear that we are not unique in oper- cockpit resource management type was dropped from further consideration. ating under stressful, risky conditions. training and the need to pay more atten- Their strongest recommendations should Other organizations have reduced fatali- tion to psychological and sociological come as no surprise—improve fire ties through training using techniques aspects of fighting fires. Paul Gleason, behavior prediction, improve weather with a proven track record. Paying more a seasoned hotshot superintendent, forecasting, develop better fuel invento- attention to the psychological and believes that the 10 Fire Orders, 18 ries, and look at our firefighting institution sociological processes of our people is Watchout Situations, and 9 Downhill/ from the external perspective. These long overdue. Indirect Line Construction Guidelines tried-and-true solutions simply fail to can be information overload for the deal with a major cause of the fatalities. It is clear that even our best crews are firefighter on the line. For this reason he not adequately trained in escape believes four of the key factors should We lost firefighters on Storm King procedures and fire shelter use. This is be constantly emphasized: Lookouts, Mountain because decision processes a reflection of the prevailing attitude Communications, Escape routes, and naturally degraded. At this time we do among managers that if we give fire- Safety zones (LCES) as central to safe not have training courses that give fire- fighters more training and better firefighting (Gleason 1991,1994 ). We fighters the knowledge to counter these predictions for fire behavior, fuels, know from our previous model that 30+ processes. Both the Investigation Team weather, and tactics, entrapments won’t warnings are an overload under normal and Review Board recommended happen. So why plan for them? conditions (seven is the practical limit) so creating a passion for safety but did not Individual firefighters agree with their LCES, while based on the others, is an acknowledge that this passion is deter- managers and also have the attitude that excellent system because it is manage- mined by psychological and sociological it won’t happen to me, so why practice able in crisis situations. Since LCES is processes. The type and skill level of for an entrapment. These attitudes easy to use, firefighters can constantly investigation team members and review caught up with our best and brightest reevaluate their situation. Gleason con- boards (typically they include IMT firefighters on Storm King Mountain and cludes that a change in training content personnel, a fire weather forecaster, fire were a causal factor in the fatalities. is not needed and that we need to better behaviorist, fuels specialist, equipment practice what we already know. specialist, but no psychologist or sociol- Since 1990, extended droughts and ogist) predisposes them to focus on the more severe fire behavior have However, I’m arguing that a different traditional inputs, which effectively shortened the time firefighters have to kind of training is needed to be able to excludes other types of input, hence decide whether to try to escape or to use our existing knowledge (including predetermining the outcome. This calls deploy shelters. Some 23 firefighters 57 Part 4 of 4 have perished trying to escape uphill ➌ Develop a training program to com- Gleason, Paul. 1994. Unprepared for carrying packs and equipment. Estimates municate these new skills to person- the Worst Case Scenario. Wildfire. 3 show most would have lived had they nel such as Incident Management (3): 23-26. simply dropped their gear and run for Teams, Type I and II crews, strike safety carrying only fire shelters. team leaders, and others at risk or Maclean, Norman. 1992. Young Men who make decisions under stress. and Fire. Chicago: University of This is why mandatory training for shelter Chicago Press. use, escape, decisionmaking under ➍ Analyze the organizational structure stress, and stress-resistant organiza- of initial attack and extended initial Putnam, Ted. 1994. Analysis of Escape tional characteristics should become attack crews and how these crews Efforts and Personal Protective national priorities. interrelate to form an effective Equipment on the South Canyon Fire. organization with optimal leadership Missoula, MT: USDA Forest Service, Everyone agrees our top priority should and decisionmaking capabilities. Missoula Technology and Develop- be reducing the number of entrapments ment Center. by practicing safety and LCES. But we ➎ Develop professional requirements, also need to face the reality that on best skills mix, and organizational South Canyon Fire Investigation Report. average 30 firefighters are trapped each structure for fatality investigation 1994. NIFC/NFES. season, and that we have not taught teams and review boards. Form IMT- them how to escape, how to use fire type teams before fatalities occur so Taynor, Janie; Gary Klein, and Marvin shelters effectively, or the concepts investigation teams are trained and Thordsen. 1987. Distributed discussed here. Clearly, firefighters need ready for dispatch. Decisionmaking in Wildland this type of training. Better personal and Firefighting. Alexandria, VA: U.S. interpersonal skills will enable firefighters ➏ Consider adding a Look Inside Army Research Institute for the to use all their training and experience component to Look Up, Look Down, Behavioral and Social Sciences. optimally under risky, stressful conditions. Look Around and an “I” to LCES. Incorporate an inner check list into the Weick, Karl. 1990. The Vulnerable Fireline Safety Reference Notebook. System: Analysis of the Tenerife Air Disaster. Journal of Management. 16: 571-593.

Recommendations Weick, Karl. 1993. The Collapse of Sensemaking in Organizations: The ➊ Implement recommendations in fire Literature Cited Mann Gulch Disaster. Ithaca, NY: shelter training stemming from the Cornell University. analysis of protective clothing and Driessen, Jon. 1990. The Supervisor equipment and its use on the South and the Work Crew. USDA Forest Wiener, Earl, Barbara Kanki, and Robert Canyon Fire (Putnam, 1994). Service, Missoula Technology and Helmrich. 1993. Cockpit Resource Development Center, Missoula, MT. Management. San Diego: Academic ➋ Convene a task group of firefighters, Press. fire training and safety officers, Fireline Safety Reference. 1992. NFES psychologists, sociologists, and others 2243. National Interagency Fire Withen, Patrick. 1994. Fire Culture. who will recommend specific actions Center. Boise, ID. Inner Voice. September/October for individuals and groups that will 1994: 12-13. maximize their resistance to decision Gleason, Paul. 1991. LCES—A Key To and organizational collapse under Safety In the Wildland Fire stressful conditions. Environment. Fire Management Notes. 52 (4): 9.

58 Part 4 of 4

The Collapse of Sense- events in Mann Gulch and simply review might be a “death trap” (p. 64). They them to provide a context for the told the second in command, William making in Organizations: analysis. Hellman, to take the crew across to the The Mann Gulch Disaster north side of the gulch and march them toward the river along the side of the hill. Karl E. Weick. Reprinted from The Collapse While Hellman did this, Dodge and Harri- of Sensemaking in Organizations: The Mann son ate a quick meal. Dodge rejoined Gulch Disaster by Karl E. Weick published in Administrative Science Quarterly Volume 38 the crew at 5:40 p.m. and took his (1993): 628-652 by permission of Administra- The Incident position at the head of the line moving tive Science Quarterly. © 1993 by Cornell toward the river. He could see flames University 0001-8392/93/3804-0628. As Maclean puts it, at its heart, the Mann flapping back and forth on the south This is a revised version of the Katz- Gulch disaster is a story of a race (p. slope as he looked to his left (p. 69). Newcomb lecture presented at the University 224). The smokejumpers in the race of Michigan, April 23-24, 1993. The 1993 (excluding foreman “Wag” Wagner At this point the reader hits the most lecture celebrated the life of Rensis Likert, the Dodge and ranger Jim Harrison) were chilling sentence in the entire book: founding director of the Institute for Social ages 17-28, unmarried, seven of them “Then Dodge saw it!” (p. 70). What he Relations. All three people honored at the lecture—Dan Katz, Ted Newcomb, and Ren were forestry students (p. 27), and 12 of saw was that the fire had crossed the Likert—were born in 1903, which meant this them had seen military service (p. 220). gulch just 200 yards ahead and was lecture also celebrated their 90th birthdays. They were a highly select group (p. 27) moving toward them (p. 70). Dodge I am grateful to Lance Sandelands, Debra and often described themselves as turned the crew around and had them Meyerson, Robert Sutton, Doug Cowherd, and Karen Weick for their help in revising professional adventurers (p. 26). A angle up the 76-percent hill toward the early drafts of this material. I also want to lightning storm passed over the Mann ridge at the top (p. 175). They were soon thank John Van Maanen, J. Richard Gulch area at 4 p.m. on August 4, 1949 moving through bunch grass that was Hackman, Linda Pike, and the anonymous and is believed to have set a small fire two and a half feet tall and were quickly ASQ reviewers for their help with later drafts. in a dead tree. The next day, August 5, losing ground to the 30-foot-high flames 1949, the temperature was 97 degrees that were soon moving toward them at The death of 13 men in the Mann Gulch and the fire danger rating was 74 out of 610 feet per minute (p. 274). Dodge fire disaster, made famous in Norman a possible 100 (p. 42), which means yelled at the crew to drop their tools, and Maclean’s Young Men and Fire, is “explosive potential” (p. 79). When the then, to everyone’s astonishment, he lit analyzed as the interactive disintegration fire was spotted by a forest ranger, the a fire in front of them and ordered them of role structure and sensemaking in a smokejumpers were dispatched to fight to lie down in the area it had burned. No minimal organization. Four potential it. Sixteen of them flew out of Missoula, one did, and they all ran for the ridge. sources of resilience that make groups Montana at 2:30 p.m. in a C-47 transport. Two people, Sallee and Rumsey, made less vulnerable to disruptions of sense- Wind conditions that day were turbulent, it through a crevice in the ridge unburned, making are proposed to forestall and one smokejumper got sick on the Hellman made it over the ridge burned disintegration, including improvisation, airplane, didn’t jump, returned to the horribly and died at noon the next day, virtual role systems, the attitude of base with the plane, and resigned from Dodge lived by lying down in the ashes wisdom, and norms of respectful inter- the smokejumpers as soon as he landed of his escape fire, and one other person, action. The analysis is then embedded (“his repressions had caught up with Joseph Sylvia, lived for a short while and in the organizational literature to show him,” p. 51). The smokejumpers and then died. The hands on Harrison’s that we need to reexamine our thinking their cargo were dropped on the south watch melted at 5:56 (p. 90), which has about temporary systems, structuration, side of Mann Gulch at 4:10 p.m. from been treated officially as the time the 13 nondisclosive intimacy, intergroup 2000 feet rather than the normal 1200 people died. dynamics, and team building. feet, due to the turbulence (p. 48). The parachute that was connected to their After the fire passed, Dodge found The purpose of this article is to reanalyze radio failed to open, and the radio was Sallee and Rumsey, and Rumsey stayed the Mann Gulch fire disaster in Montana pulverized when it hit the ground. The to care for Hellman while Sallee and described in Norman Maclean’s (1992) crew met ranger Jim Harrison who had Dodge hiked out for help. They walked award-winning book Young Men and been fighting the fire alone for four hours into the Meriwether ranger station at Fire to illustrate a gap in our current (p. 62), collected their supplies, and ate 8:50 p.m. (p. 113), and rescue parties understanding of organizations. l want supper. About 5:10 p.m. (p. 57) they immediately set out to recover the dead to focus on two questions: Why do started to move along the south side of and dying. All the dead were found in an organizations unravel? And how can the gulch to surround the fire (p. 62). area of 100 yards by 300 yards (p. 111). organizations be made more resilient? Dodge and Harrison, however, having It took 450 men five more days to get Before doing so, however, l want to strip scouted ahead, were worried that the the 4,500-acre Mann Gulch fire under Maclean’s elegant prose away from the thick forest near which they had landed control (pp. 24, 33). At the time the crew 59 Part 4 of 4 jumped on the fire, it was classified as five years after the fire (p. 106). Maclean lapse of time, the destructive forces of a Class C fire, meaning its scope was located and interviewed both living wit- nature over 28 years, and the power of between 10 and 99 acres. nesses of the blaze, Sallee and Rumsey, a blowup fire to melt and displace every- and persuaded both to accompany him thing in its path, discovery of these The Forest Service inquiry held after the and Laird Robinson, a guide at the traces is surprising as well as helpful in fire, judged by many to be inadequate, smokejumper base, on a visit back to reconstructing events. concluded that “there is no evidence of the site on July 1, 1978. Maclean also disregard by those responsible for the knew Dodge’s wife and had talked to her Archival records are crucial to the devel- jumper crew of the elements of risk informally (p. 40). He attempted to inter- opment of the case, although the Forest which they are expected to take into view relatives of some who lost their Service made a considerable effort after account in placing jumper crews on lives but found them too distraught 27 its inquiry to scatter the documents (p. fires.” The board also felt that the men years later to be of much help (p. 154). 153) and to classify most of them “Confi- would have been saved had they He also attempted to interview (p. 239) dential” (p. 158), perhaps fearing it would “heeded Dodge’s efforts to get them to a member of the Forest Service inquiry be charged with negligence. Records go into the escape fire area with him” team, A. J. Cramer who, in 1951, had used by Maclean included statistical (quoted in Maclean, p. 151). Several persuaded Sallee, Rumsey, and ranger reports of fire suppression by smoke- parents brought suit against the Forest Robert Jansson to alter their testimony jumpers in Forest Service Region 1 Service, claiming that people should not about the timing of key incidents. (e.g., p. 24); the report of the Forest have been jumped in the first place (p. Cramer was the custodian of seven or Service Board of Review issued shortly 149), but these claims were dismissed eight watches that had been removed after the incident (dated September 29, by the Ninth Circuit U.S. Court of from victims (p. 233), only one of which 1949, which many felt was too soon for Appeals, where Warren E. Burger arg- (Harrison’s) was released and used as the board to do an adequate job); state- ued the Forest Service’s case (p. 151). the official time of the disaster (5:56 ments made to the board by people such p.m.). To this day it remains unclear as the C-47 pilot, parents of the dead Since Mann Gulch, there have been no why the Forest Service made such a crew (p. 150), and the spotter on the deaths by burning among Forest Service strong effort to locate the disaster closer aircraft (p. 42); court reports of litigation firefighters, and people are now to 6:00 p.m. than to 5:30, which was brought by parents of smokejumpers equipped with backup radios (p. 219), suggested by testimony from Jansson, against the Forest Service; photographs, better physical conditioning, the tactic of who was near the river when the fire virtually all of which were retrieved for building an escape fire, knowledge that blew up, and from a recovered watch him by women in the Forest Service who fires in timber west of the Continental that read 5:42. Maclean had continuing were eager to help him tell the story (p. Divide burn differently than do fires in access to two Forest Service insiders, 160); early records of the smokejumpers grass east of the Divide, and the insist- Bud Moore and Laird Robinson (p. 162). organization, which was nine years old ence that crew safety take precedence He also interviewed experts on prece- at the time of the disaster; reports of the over fire suppression. dents for the escape fire (p. 104) and 1957 task force on crew safety (p. 221); on the nature of death by fire (p. 213). and contemporary reports of the disaster in the media, such as the report in the The use of trace records, or physical August 22, 1949 issue of Life magazine. evidence of past behaviors, is illustrated by the location during a 1979 trip to the Direct observation occurred during The Methodology gulch, of the wooden cross that had Maclean’s three visits to Mann Gulch in been placed in 1949 to mark the spot 1976, 1977, and 1978 (p. 189), trips Among the sources of evidence Maclean where Dodge lit his escape fire (p. 206). made much more difficult because of used to construct this case study were The year before, 1978, during the trip the inaccessibility of the area (pp. 191- interviews, trace records, archival into the gulch with Sallee and Rumsey, 192). The most important of these three records, direct observation, personal Maclean located the rusty can of visits is the trip to the gulch with Sallee experience, and mathematical models. potatoes that had been discarded after and Rumsey, during which the latter pair Hellman drank its salty water through reenacted what they did and what they Since Maclean did not begin to gather two knife slits Rumsey had made in the saw intermittently through the dense documents on Mann Gulch until 1976 can (p. 173). He also located the flat smoke. When their accounts were (p. 156) and did not start to work in rocks on which Hellman and Sylvia had matched against subsequent hard data earnest on this project until his seventy- rested while awaiting rescue, the juniper (e.g., their estimation of where Dodge fourth birthday in 1977, the lapse of tree that was just beyond the crevice lit his escape fire compared against almost 28 years since the disaster made Sallee and Rumsey squeezed through discovery of the actual cross planted in interviewing difficult, especially since on the ridge (p. 207), and Henry Thol, 1949 to mark the spot), it was found that Dodge had died of Hodgkin’s disease Jr.’s flashlight (p. 183). Considering the their reconstruction of events prior to 60 Part 4 of 4 the time they made it to safety through between fire and men unfolded, Maclean Let me first be clear about why I think the the crevice is less accurate than their taught himself mathematics and turned crew of smokejumpers at Mann Gulch memory for events and locations after to mathematical modeling. He worked was an organization. First, they have a they made it to safety. This suggests to with two mathematicians, Frank Albini series of interlocking routines, which is Maclean that “we don’t remember as and Richard Rothermel, who had built crucial in Westley’s (1990: 339) definition exactly the desperate moments when mathematical models of how fires of an organization as “a series of inter- our lives are in the balance as we spread. The group ran the predictive locking routines, habituated action remember the moments after, when the models in reverse to see what the fire patterns that bring the same people balance has tipped in our favor” (p. 212). in Mann Gulch must have been like to together around the same activities in Direct observation also occurred when generate the reports on its progress that the same time and places.” The crew at Maclean and Robinson themselves were found in interviews, reports, and Mann Gulch have routine, habituated hiked the steep slopes of Mann Gulch actual measurements. It is the combina- action patterns, they come together from under summer conditions of heat and tion of output from the model and a common pool of people, and while slippery, tall grass that resembled the subjective reports that provide the this set of individual smokejumpers had conditions present in the disaster of revealing time line of the final 16 not come together at the same places or 1949. The two men repeatedly compared minutes (pp. 267-277). times, they did come together around photos and maps from 1949 with physi- the same episodes of fire. Westley’s cal outcroppings in front of them to see If these several sources of evidence are definition suggests it doesn’t take much more clearly what they were looking at combined and assessed for the to qualify as an organization. The other (e.g., photos misrepresent the steep- adequacy with which they address side is, it also may not take much to stop ness of the slope, p. 175). There were “sources of invalidity,” it will be found being one. also informal experiments, as when Rod that they combat 12 of the 15 sources Norum, an athlete and specialist on fire listed by Runkel and McGrath (1972: Second, the Mann Gulch crew fits the behavior, retraced Dodge’s route from 191) and are only “moderately five criteria for a simple organizational the point at which he rejoined the crew, vulnerable” to the other three. Of course, structure proposed by Mintzberg (1983: moved as fast as possible over the route an experienced woodsman and 158). These five include coordination by Dodge covered, and was unable to storyteller who has “always tried to be direct supervision, strategy planned at reach the grave markers as fast as the accurate with facts” (p. 259) would the top, little formalized behavior, organic crew did (p. 67). During these trips, expect that. The rest of us in organiza- structure, and the person in charge Maclean took special note of prevailing tional studies may be pardoned, tending to formulate plans intuitively, winds by observing their effect on the however, if we find those numbers a meaning that the plans are generally a direction in which rotted timber fell. good reason to take these data seriously. direct “extension of his own personality.” These observations were used to build Structures like this are found most often a theory of how wind currents in the in entrepreneurial firms. gulch could have produced the blowup (p. 133). And third, the Mann Gulch crew has “generic subjectivity” (Wiley, 1988), Personal experience was part of the Cosmology Episodes in meaning that roles and rules exist that case because, in 1949, Maclean had Mann Gulch enable individuals to be interchanged visited the Mann Gulch fire while it was with little disruption to the ongoing still burning (p. 1). Maclean also was a Early in the book (p. 65), Maclean asks pattern of interaction. In the crew at Forest Service firefighter (not a smoke- the question on which I want to focus: Mann Gulch there were at least three jumper) at age 15 and nearly lost his life “what the structure of a small outfit roles: leader, second in command, and in the Fish Creek fire, a fire much like should be when its business is to meet crewmember. The person in the lead the one in Mann Gulch (p. 4). Maclean sudden danger and prevent disaster.” sizes up the situation, makes decisions, also reports using his practical experi- This question is timely because the work yells orders, picks trails, sets the pace, ence as a woodsman to suggest initial of organizations is increasingly done in and identifies escape routes (pp. 65-66). hypotheses regarding what happened small temporary outfits in which the The second in command brings up the at Mann Gulch (e.g., he infers wind stakes are high and where foul-ups can rear of the crew as it hikes, repeats patterns in the gulch from observations have serious consequences (Heyde- orders, sees that the orders are under- of unusual wave action in the adjacent brand, 1989; Ancona and Caldwell, stood, helps the individuals coordinate Missouri River, p. 131). 1992). Thus, if we understand what their actions, and tends to be closer to happened at Mann Gulch, we may be the crew and more of a buddy with them Having collected data using the above able to learn some valuable lessons in than does the leader. And finally, the sources, but still feeling gaps in his how to conceptualize and cope with crew clears a fire line around the fire, understanding of precisely how the race contemporary organizations. cleans up after the fire, and maintains 61 Part 4 of 4 trails. Thus, the crew at Mann Gulch is an Stated more informally, a cosmology There have been at least three distinct organization by virtue of a role structure episode feels like vu jàdé—the opposite responses to these problems. First, there of interlocking routines. of déjà vu: I’ve never been here before, has been a shift, reminiscent of Neisser l have no idea where I am, and I have no and Winograd’s (1988) work on memory, I want to argue that the tragedy at Mann idea who can help me. This is what the toward examining naturalistic decision Gulch alerts us to an unsuspected smokejumpers may have felt increas- making (Orasanu and Connolly, 1993), source of vulnerability in organizations. ingly as the afternoon wore on and they with more attention to situational Minimal organizations, such as we find lost what little organization structure they assessment and sensemaking (Klein, in the crew at Mann Gulch, are suscep- had to start with. As they lost structure 1993). Second, people have replaced tible to sudden losses of meaning, which they became more anxious and found it an interest in decision making with an have been variously described as funda- harder to make sense of what was interest in power, noting, for example, mental surprises (Reason, 1990) or happening, until they finally were unable that “power is most strategically deployed events that are inconceivable (Lanir, to make any sense whatsoever of the in the design and implementation of 1989), hidden (Westrum, 1982), or one thing that would have saved their paradigmatic frameworks within which incomprehensible (Perrow, 1984). Each lives, an escape fire. The disaster at the very meaning of such actions as of these labels points to the low proba- Mann Gulch was produced by the inter- ‘making decisions’ is defined” (Brown, bility that the event could occur, which is related collapse of sensemaking and 1978: 376). And third, people are replac- why it is meaningless. But these explan- structure. If we can understand this ing the less appropriate normative ations say less about the astonishment collapse, we may be able to forestall models of rationality (e.g., Hirsch, of the perceiver and even less about similar disasters in other organizations. Michaels, and Friedman, 1987) based the perceiver’s inability to rebuild some on asocial “economic man” (Beach and sense of what is happening. Lipshitz, 1993) with more appropriate models of rationality that are more To shift the analytic focus in implausible Sensemaking in Mann sophisticated about social relations, such events from probabilities to feelings and as the model of contextual rationality social construction, l have borrowed the Gulch (White, 1988). term “cosmology” from philosophy and stretched it. Cosmology refers to a Although most organizational analyses Reed (1991) described contextual branch of philosophy often subsumed begin and end with decision making, rationality as action motivated to create under metaphysics that combines there is growing dissatisfaction with this and maintain institutions and traditions rational speculation and scientific evi- orthodoxy. Reed (1991) showed how that express some conception of right dence to understand the universe as a far the concept of decision making has behavior and a good life with others. totality of phenomena. Cosmology is the been stretched, singling out the patching Contextual rationality is sensitive to the ultimate macro perspective, directed at that James G. March has done in recent fact that social actors need to create issues of time, space, change, and discussions of decision making. March and maintain intersubjectively binding contingency as they relate to the origin (1989: 14) wrote that “decision making normative structures that sustain and and structure of the universe. Integra- is a highly contextual, sacred activity, enrich their relationships. Thus, organi- tions of these issues, however, are not surrounded by myth and ritual, and as zations become important because they just the handiwork of philosophers. much concerned with the interpretive can provide meaning and order in the Others also make their peace with these order as with the specifics of particular face of environments that impose ill- issues, as reflected in what they take for choices.” Reed (1991: 561) summarized defined, contradictory demands. granted. People, including those who are March this way: “decision making prefer- smokejumpers, act as if events cohere ences are often inconsistent, unstable, One way to shift the focus from decision in time and space and that change and externally driven; the linkages making to meaning is to look more unfolds in an orderly manner. These between decisions and actions are closely at sensemaking in organizations. everyday cosmologies are subject to loosely-coupled and interactive rather The basic idea of sensemaking is that disruption. And when they are severely than linear; the past is notoriously unre- reality is an ongoing accomplishment disrupted, l call this a cosmology episode liable as a guide to the present or the that emerges from efforts to create order (Weick, 1985: 51-52). A cosmology future; and…political and symbolic and make retrospective sense of what episode occurs when people suddenly considerations play a central, perhaps occurs. Recognition-primed decision and deeply feel that the universe is no overriding, role in decision making.” making, a model based in part on com- longer a rational, orderly system. What Reed wondered aloud whether, if March mand decisions made by firefighters, has makes such an episode so shattering is is right in these descriptions, decision features of sensemaking in its reliance that both the sense of what is occurring making should continue to set the on past experience, although it remains and the means to rebuild that sense agenda for organizational studies. At grounded in decision making (Klein, collapse together. some point a retreat from classic 1993). Sensemaking emphasizes that principles becomes a rout.

62 Part 4 of 4 people try to make things rationally 6. As the fire gains on them, Dodge says, making is different. Sensemaking is accountable to themselves and others. “Drop your tools,” but if the people in about contextual rationality. It is built out Thus, in the words of Morgan, Frost, and the crew do that, then who are they? of vague questions, muddy answers, Pondy (1983: 24), “individuals are not Firefighters? With no tools? and negotiated agreements that seen as living in, and acting out their attempt to reduce confusion. People in lives in relation to, a wider reality, so 7. The foreman lights a fire that seems Mann Gulch did not face questions like much as creating and sustaining images to be right in the middle of the only where should we go, when do we take a of a wider reality, in part to rationalize escape route people can see. stand, or what should our strategy be? what they are doing. They realize their Instead, they faced the more basic, the reality, by reading into their situation 8. The foreman points to the fire he has more frightening feeling that their old patterns of significant meaning.” started and yells, “Join me,” whatever labels were no longer working. They that means. But his second in com- were outstripping their past experience When the smokejumpers landed at mand sounds like he’s saying, “To and were not sure either what was up or Mann Gulch, they expected to find what hell with that, I’m getting out of here” who they were. Until they develop some they had come to call a 10:00 fire. A (p. 95). sense of issues like this, there is nothing 10:00 fire is one that can be surrounded to decide. completely and isolated by 10:00 the 9. Each individual faces the dilemma, l next morning. The spotters on the must be my own boss yet follow orders aircraft that carried the smokejumpers unhesitatingly, but I can’t comprehend “figured the crew would have it under what the orders mean, and I’m losing Role Structure in Mann control by 10:00 the next morning” my race with the advancing fire (pp. (Maclean, p. 43). People rationalized 219-220). Gulch this image until it was too late. And because they did, less and less of what As Mann Gulch loses its resemblance to Sensemaking was not the only problem they saw made sense: a 10:00 fire, it does so in ways that make in Mann Gulch. There were also it increasingly hard to socially construct problems of structure. It seems plausible 1. The crew expects a 10:00 fire but reality. When the noise created by wind, to argue that a major contributor to this grows uneasy when this fire does not flames, and exploding trees is deafening; disaster was the loss of the only struc- act like one. when people are strung out in a line and ture that kept these people organized, relative strangers to begin with; when their role system. There were two key 2. Crewmembers wonder how this fire they are people who, in Maclean’s words, events that destroyed the organization can be all that serious if Dodge and “love the universe but are not intimidated that tied these people together. First, Harrison eat supper while they hike by it” (p. 28); and when the temperature when Dodge told Hellman to take the toward the river. is approaching a lethal 140 degrees (p. crew to the north side of the gulch and 220), people can neither validate their have it follow a contour down toward the 3. People are often unclear who is in impressions with a trusted neighbor nor river, the crew got confused, the spaces charge of the crew (p. 65). pay close attention to a boss who is also between members widened appreciably, unknown and whose commands make and Navon—the person taking pictures 4. The flames on the south side of the no sense whatsoever. As if these were (p. 71)— made a bid to take over the gulch look intense, yet one of the not obstacles enough, it is hard to make leadership of the group (p. 65). Notice smokejumpers, David Navon is taking common sense when each person sees what this does to the role system. There pictures, so people conclude the fire something different or nothing at all is now no one at the end of the line can’t be that serious, even though because of the smoke. repeating orders as a check on the their senses tell them otherwise. accuracy with which they are under- The crew’s stubborn belief that it faced stood. Furthermore, the person who is 5. Crewmembers know they are moving a 10:00 fire is a powerful reminder that leading them, Hellman, is more familiar toward the river where they will be positive illusions (Taylor, 1989) can kill with implementing orders than with safe from the fire, only to see Dodge people. But the more general point is constructing them or plotting possible inexplicably turn them around, away that organizations can be good at escape routes. So the crew is left for a from the river, and start angling decision making and still falter. They crucial period of time with ill-structured, upslope, but not running straight for falter because of deficient sensemaking. unacknowledged orders shouted by the top. Why? (Dodge is the only one The world of decision making is about someone who is unaccustomed to being who sees the fire jump the gulch strategic rationality. It is built from clear firm or noticing escape routes. Both ahead of them.) questions and clear answers that routines and interlocking are beginning attempt to remove ignorance (Daft and to come apart. The second, and in some Macintosh, 1981). The world of sense- way more unsettling threat to the role 63 Part 4 of 4 system occurred when Dodge told the as McDougall’s (1920), which had as they knew it, disintegrated. As their retreating crew “throw away your tools!” argued that panic leads to group disinte- group disintegrated, the smokejumpers (p. 226). A fire crew that retreats from a gration, Freud, reversing this causality, became more frightened, stopped fire should find its identity and morale argued that group disintegration precip- thinking sooner, pulled apart even more, strained. If the retreating people are then itates panic. By group disintegration, and in doing so, lost a leader-follower also told to discard the very things that Freud meant “the cessation of all the relationship as well as access to the are their reason for being there in the feelings of consideration which the novel ideas of other people who are a lot first place, then the moment quickly turns members of the group otherwise show like them. As these relationships disap- existential. If I am no longer a firefighter, one another” (p. 29). He described the peared, individuals reverted to primitive then who am l? With the fire bearing mechanism involved this way: “If an tendencies of flight. Unfortunately, this down, the only possible answer individual in panic fear begins to be response was too simple to match the becomes, An endangered person in a solicitous only on his own account, he complexity of the Mann Gulch fire. world where it is every man for himself. bears witness in so doing to the fact that Thus, people who, in Maclean’s words, the emotional ties, which have hitherto What holds organization in place may had perpetually been almost their own made the danger seem small to him, be more tenuous than we realize. The boss (p. 218) suddenly became com- have ceased to exist. Now that he is by recipe for disorganization in Mann Gulch pletely their own boss at the worst himself in facing the danger, he may is not all that rare in everyday life. The possible moment. As the entity of a crew surely think it greater.” recipe reads, Thrust people into dissolved, it is not surprising that the final unfamiliar roles, leave some key roles command from the “crew” leader to jump It is certainly true in Mann Gulch that unfilled, make the task more ambiguous, into an escape fire was heard not as a there is a real, palpable danger that can discredit the role system, and make all legitimate order but as the ravings of be seen, felt, heard, and smelled by the of these changes in a context in which someone who had “gone nuts” (p. 75). smokejumpers. But this is not the first small events can combine into something Dodge’s command lost its basis of legiti- time they have confronted danger. It monstrous. Faced with similar conditions, macy when the smokejumpers threw may, however, be the first time they organizations that seem much sturdier away their organization along with their have confronted danger as a member of may also come crashing down (Miller, tools. a disintegrating organization. As the crew 1990; Miles and Snow, 1992), much like moved toward the river and became Icarus who overreached his competence more spread out, individuals were as he flew toward the sun and also isolated and left without explanations or perished because of fire. emotional support for their reactions. As the ties weakened, the sense of danger Panic In Mann Gulch increased, and the means to cope be- came more primitive. The world rapidly With these observations as background, shifted from a cosmos to chaos as it we can now look more closely at the became emptied of order and rationality. From Vulnerability to process of a cosmology episode, an Resilience interlude in which the orderliness of the It is intriguing that the three people who universe is called into question because survived the disaster did so in ways that The steady erosion of sense and struc- both understanding and procedures for seem to forestall group disintegration. ture reached its climax in the refusal of sensemaking collapse together. People Sallee and Rumsey stuck together, the crew to escape one fire by walking stop thinking and panic. What is inter- their small group of two people did not into another one that was intentionally esting about this collapse is that it was disintegrate, which helped them keep set. A closer look at that escape fire discussed by Freud (1959: 28) in the their fear under control. As a result, they allows us to move from a discussion of context of panic in military groups: “A escaped through a crack in the ridge that what went wrong at Mann Gulch, to a panic arises if a group of that kind the others either didn’t see or thought discussion of what makes organizations [military group] becomes disintegrated. was too small to squeeze through. Wag more resilient. l want to discuss four Its characteristics are that none of the Dodge, as the formal leader of a group sources of resilience: (1) improvisation orders given by superiors are any longer he presumed still existed, ordered his and bricolage, (2) virtual role systems, listened to, and that each individual is followers to join him in the escape fire. (3) the attitude of wisdom, and (4) only solicitous on his own account, and Dodge continued to see a group and to respectful interaction. without any consideration for the rest. think about its well-being, which helped The mutual ties have ceased to exist, keep his own fear under control. The rest and a gigantic and senseless fear is set of the people, however, took less notice free.” Unlike earlier formulations, such of one another. Consequently, the group,

64 Part 4 of 4

Improvisation and beyond what you currently think.” With While improvised fire fighting may sound this as background, it now becomes improbable, in fact, Park Service fire- Bricolage relevant that Dodge was an experienced fighters like those stationed at the Grand woodsman, with lots of hands-on exper- Canyon approximate just such a style. The escape fire is a good place to start ience. He was what we now would call Stephen Pyne (1989), a Park Service in the search for sources of resilience a bricoleur, someone able to create firefighter, observed that people like him simply because it is clear evidence that, order out of whatever materials were at typically have discretion to dispatch minimal though the organization of the hand (e.g., Levi-Strauss, 1966; Harper, themselves, which is unfathomable to crew might have been, there still was a 1987). Dodge would have known at the Forest Service crews that rely on solution to the crisis inside the group. least two things about fires. He would dispatchers, specialization, regimenta- The problem was, no one but Dodge have known the famous fire triangle— tion, rules, and a conscious preference recognized this. The question then you must have oxygen, flammable for the strength of the whole rather than becomes, How could more people either material, and temperature above the the versatility and resourcefulness of the see this escape fire as a solution or point of ignition to create a fire (Maclean, parts. Forest Service people marvel at develop their own solution? This is not p. 35). A shortage of any one of these the freedom of movement among the an easy question to answer because, would prevent a fire. In his case, the Park people. Park Service people marvel from everything we know, Dodge’s escape fire removed flammable material. at how much power the Forest Service invention of burning a hole in a fire And since Dodge had been with the is able to mobilize on a fire. Pyne (1989: should not have happened. It should not Forest Service longer than anyone else 122) described the Park Service fire have happened because there is good on the crew, he would also have known operations as a nonstandard “eclectic evidence that when people are put under more fully their four guidelines at that assembly of compromises” built of pressure, they regress to their most time for dealing with fire emergencies discretion and mobility. In contrast to the habituated ways of responding (e.g., (p. 100). These included (1) start a Forest Service, where people do every- Barthol and Ku, 1959). This is what we backfire if you can, (2) get to the top of thing by the book, “The Park Service see in the 15 people who reject Dodge’s a ridge where the fuel is thinner, (3) turn has no books; it puts a premium on the order to join him and who resort instead into the fire and try to work through it, individual. Its collective behavior is tribal, to flight, a more overlearned tendency. and (4) don’t allow the fire to pick the and it protects its permanent ranks.” If What we do not expect under life- spot where it hits you. Dodge’s invention, improvisation were given more attention threatening pressure is creativity. if we stretch a bit, fits all four. It is a in the job description of a crew person, backfire, though not in the conventional that person’s receptiveness to and The tactic of lighting a fire to create an sense of a fire built to stop a fire. The generation of role improvisations might area where people can escape a major escape fire is lit near the top of a ridge, be enhanced. As a result, when one prairie fire is mentioned in James Feni- Dodge turns into the main fire and works organizational order collapses, a substi- more Cooper’s 1827 novel The Prairie, through it by burning a hole in it, and he tute might be invented immediately. Swift but there is no evidence Dodge knew chooses where the fire hits him. The 15 replacement of a traditional order with this source (Maclean, p. 104). Further- who tried to outrun the fire moved toward an improvised order would forestall the more, most of Dodge’s experience had the ridge but by not facing the fire, they paralysis that can follow a command to been in timbered country where such a allowed it to pick the spot where it hit “drop your tools.” tactic wouldn’t work. In timber, an them. escape fire is too slow and consumes too much oxygen (p. 105). And the fire The collapse of role systems need not that Dodge built did not burn long result in disaster if people develop skills enough to clear an area in which people in improvisation and bricolage (see could move around and dodge the fire Janowitz, 1959: 481). Bricoleurs remain Virtual Role Systems as they did in the prairie fire. There was creative under pressure, precisely just room enough to lie down in the because they routinely act in chaotic Social construction of reality is next to ashes where the heat was less intense conditions and pull order out of them. impossible amidst the chaos of a fire, (p. 104). Thus, when situations unravel, this is unless social construction takes place simply normal, natural trouble for brico- inside one person’s head, where the role While no one can say how or why the leurs, and they proceed with whatever system is reconstituted and run. Even escape fire was created, there is a line materials are at hand. Knowing these though the role system at Mann Gulch of argument that is consistent with what materials intimately, they then are able, collapsed, this kind of collapse need not we know. Bruner (1983: 183) described usually in the company of other similarly result in disaster if the system remains creativity as “figuring out how to use skilled people, to form the materials or intact in the individual’s mind. If each what you already know in order to go insights into novel combinations. individual in the crew mentally takes all

65 Part 4 of 4 roles and therefore can then register become big as hell, the ordinary can overcautious shun curiosity for fear it escape routes and acknowledge suddenly become monstrous, and the will only deepen their uncertainties. commands and facilitate coordination, upgulch breezes can suddenly turn to Both the cautious and the confident are then each person literally becomes a murder” (p. 217). To state the point more closed-minded, which means neither group (Schutz, 1961). And, in the manner generally, what most organizations miss, makes good judgments. It is this sense of a holograph, each person can recon- and what explains why most organiza- in which wisdom, which avoids extremes, stitute the group and assume whatever tions fail to learn (Scott, 1987: 282), is improves adaptability. role is vacated, pick up the activities, and that “Reality backs up while it is run a credible version of the role. approached by the subject who tries to A good example of wisdom in groups is Furthermore, people can run the group understand it. Ignorance and knowledge the Naskapi Indians’ use of caribou in their head and use it for continued grow together” (Meacham, 1983: 130). shoulder bones to locate game (Weick, guidance of their own individual action. To put it a different way, “Each new 1979). They hold bones over a fire until domain of knowledge appears simple they crack and then hunt in the direc- It makes just as much sense to talk from the distance of ignorance. The more tions to which the cracks point. This about a virtual role system as it does to we learn about a particular domain, the ritual is effective because the decision talk about a virtual anything else (e.g., greater the number of uncertainties, is not influenced by the outcomes of past Bruner, 1986: 36-37). An organization doubts, questions and complexities. hunts, which means the stock of animals can continue to function in the imagina- Each bit of knowledge serves as the is not depleted. More important, the final tion long after it has ceased to function thesis from which additional questions or decision is not influenced by the inevit- in tangible distributed activities. For the antithesis arise” (Meacham, 1983: 120). able patterning in human choice, which Mann Gulch fire, this issue has bearing enables hunted animals to become on the question of escape routes. In our The role system best able to accept the sensitized to humans and take evasive research on accidents in flight operations reality that ignorance and knowledge action. The wisdom inherent in this off nuclear carriers (Weick and Roberts, grow together may be one in which the practice derives from its ambivalence 1993), Karlene Roberts and I found that organizational culture values wisdom. toward the past. Any attempt to hunt for people who avoid accidents live by the Meacham (1983: 187) argued that caribou is both a new experience and credo, “never get into anything without wisdom is an attitude rather than a skill an old experience. It is new in the sense making sure you have a way out.” At the or a body of information: that time has elapsed, the composition very last moment in the Mann Gulch of the hunter band has changed, the tragedy, Dodge discovered a way out. To be wise is not to know particular facts caribou have learned new things, and so The point is that if other people had been but to know without excessive forth. But the hunt is also old in the sense able to simulate Dodge and/or his role confidence or excessive cautiousness. that if you’ve seen one hunt, you’ve seen in their imagination, they too might have Wisdom is thus not a belief, a value, a them all: There are always hunters, been less puzzled by his solution or set of facts, a corpus of knowledge or weapons, stealth, decoys, tracks, odors, better able to invent a different sensible information in some specialized area, or and winds. The practice of divination solution for themselves. a set of special abilities or skills. Wisdom incorporates the attitude of wisdom is an attitude taken by persons toward because past experience is discounted the beliefs, values, knowledge, informa- when a new set of cracks forms a crude tion, abilities, and skills that are held, a map for the hunt. But past experience tendency to doubt that these are neces- is also given some weight, because a sarily true or valid and to doubt that they seasoned hunter “reads” the cracks and The Attitude of Wisdom are an exhaustive set of those things injects some of his own past experience that could be known. into an interpretation of what the cracks To understand the role of wisdom (Bige- mean. The reader is crucial. If the low, 1992) as a source of resilience, we In a fluid world, wise people know that reader’s hunches dominate, randomiza- need to return to the crew’s belief that they don’t fully understand what is tion is lost. If the cracks dominate, then all fires are 10:00 fires. This belief was happening right now, because they have the experience base is discarded. The consistent with members’ experience. never seen precisely this event before. cracks are a lot like the four guidelines As Maclean put it, if the major purpose Extreme confidence and extreme for fire emergencies that Dodge may of your group is to “put out fires so fast caution both can destroy what organiza- have relied on when he invented the they don’t have time to become big ones” tions most need in changing times, escape fire. They embody experience, (p. 31), then you won’t learn much about namely, curiosity, openness, and but they invite doubt, reassembly, and fighting big fires. Nor will you learn what complex sensing. The overconfident shaping to fit novelties in the present. Maclean calls the first principle of reality: shun curiosity because they feel they “little things suddenly and literally can know most of what there is to know. The

66 Part 4 of 4

Respectful Interaction without deprecating them or yourselves likely to be mitigated than the speech of (self-respect) (adapted from Campbell, captains. It was also found that topics The final suggestion about how to 1990: 45-46). introduced in mitigated speech were less counteract vulnerability makes explicit likely to be followed-up by other crew the preceding focus on the individual and Earlier I noted a growing interest in members and less likely to be ratified by social interaction. Respectful interaction contextual rationality, understood as the captain. Both of these effects relate depends on intersubjectivity (Wiley, actions that create and maintain institu- directly to the situation in which a subor- 1988: 258), which has two defining tions and traditions that express some dinate crew member makes a correct characteristics: (1) intersubjectivity conception of right behavior and a good solution that is ignored… The value of emerges from the interchange and life with others (Reed, 1991). Campbell’s training in unmitigated speech is strongly synthesis of meanings among two or maxims operationalize this good life with suggested by these results. (O’Hare and more communicating selves, and (2) the others as trust, honesty, and self-respect Roscoe, 1990: 219) self or subject gets transformed during in moment-to-moment interaction. This interaction such that a joint or merged triangle of trust, honesty, and self-respect If a role system collapses among people subjectivity develops. It is possible that is conspicuously missing (e.g., King, for whom trust, honesty, and self-respect many role systems do not change fast 1989: 46-48) in several well-documented are underdeveloped, then they are on enough to keep up with a rapidly disasters in which faulty interaction their own. And fear often swamps their changing environment. The only form processes led to increased fear, dimin- resourcefulness. If, however, a role that can keep up is one based on face- ished communication, and death. For system collapses among people where to-face interaction. And it is here, rather example, in the Tenerife air disaster trust, honesty, and self-respect are more than in routines, that we are best able to (Weick, 1990), the copilot of the KLM fully developed, then new options, such see the core of organizing. This may be aircraft had a strong hunch that another as mutual adaptation, blind imitation of why interaction in airline cockpit crews, 747 airplane was on the takeoff runway creative solutions, and trusting compli- such as discussed by Foushee (1984), directly in front of them when his own ance, are created. When a formal strikes us so often as a plausible micro- captain began takeoff without clearance. structure collapses, there is no leader, cosm of what happens in much larger But the copilot said nothing about either no roles, no routines, no sense. That is systems. In a cockpit under crisis, the the suspicions or the illegal departure. what we may be seeing in Mann Gulch. only unit that makes sense (pun intend- Transient cockpit crews, tied together by Dodge can’t lead because the role ed) is face-to-face synthesis of meaning. narrow definitions of formal responsibili- system in which he is a leader disap- ties, and headed by captains who pears. But what is worse, Dodge can’t Intersubjectivity was lost on everyone at mistakenly assume that their decision- rely on his crew members to trust him, Mann Gulch, everyone, that is, but Sallee making ability is unaffected by increases question him, or pay attention to him, and Rumsey. They stuck together and in stress (Helmreich et al., 1985), have because they don’t know him and there lived. Dodge went his own individual way few protections against a sudden loss is no time to change this. The key ques- with a burst of improvisation, and he too of meaning such as the preposterous tion is, When formal structure collapses, lived. Perhaps it’s more important that possibility that a captain is taking off what, if anything, is left? The answer to you have a partner than an organization without clearance, directly into the path that question may well be one of life or when you fight fires. A partner makes of another 747. death. social construction easier. A partner is a second source of ideas. A partner Even when people try to act with strengthens independent judgment in the honesty, trust, and self respect, if they face of a majority. And a partner enlarges do so with little social support, their the pool of data that are considered. efforts are compromised. For example, Partnerships that endure are likely to be linguists who analyzed the conversations those that adhere to Campbell’s three at Tenerife and in the crash of Air Florida Structures For imperatives for social life, based on a flight 90 in Washington concluded that Resilience reanalysis of Asch’s (1952) conformity the copilots in both cases used “devices experiment: (1) Respect the reports of of mitigation” to soften the effects of their While the answer to that question is not others and be willing to base beliefs requests and suggestions: a matter of life or death for organiza- and actions on them (trust); (2) Report tional theorists, they do have an interest honestly so that others may use your A mitigated instruction might be phrased in how it comes out. A theorist who hears observations in coming to valid beliefs as a question or hedged with qualifica- Maclean’s question, “what the structure (honesty); and, (3) Respect your own tions such as “would” or “could.” …(I)t of a small outfit should be when its perceptions and beliefs and seek to was found that the speech of subordi- business is to meet sudden danger and integrate them with the reports of others nate crew members was much more prevent disaster,” might come back with

67 Part 4 of 4 a series of follow-up questions based on participant.” While the smokejumpers “recreated in interaction,” “constituted,” thinking in organizational studies. l look have the obvious superordinate goal of and “constitutive” directs attention away briefly at four such questions to link containing fires, their group ties may not from losses of frameworks and losses of Mann Gulch with other concepts and to be sufficiently developed for this to be a meaning. For example, Ranson, Hinings, suggest how these linkages might guide group goal that overrides self-interest. Or and Greenwood (1980: 5) asserted that further research. Bass’s proposition itself may be incom- the “deep structure of schema which are plete, failing to acknowledge that unless taken for granted by members enables First, there is the follow-up question, Is superordinate goals are overlearned, them to recognize, interpret, and negoti- “small” necessarily a key dimension, they will be discarded in situations of ate even strange and unanticipated since this group is also young and danger. situations, and thus continuously to transient? Maclean calls the 16-person create and reenact the sense and smokejumper crew “small,” except that Second, there is the follow-up question, meaning of structural forms during the it is conventional in the group literature Is “structure” what we need to under- course of interaction.” The Mann Gulch to treat any group of more than 10 stand in Mann Gulch, or might structuring disaster is a case in which people were people as large (Bass, 1990: 604). also be important? By structure, l mean unable to negotiate strangeness. Because there is so little communication “a complex medium of control which is Frameworks and meanings destroyed within the crew and because it operates continually produced and recreated in rather than constructed one another. largely through obtrusive controls like interaction and yet shapes that interac- rules and supervision (Perrow, 1986), tion: structures are constituted and This fugitive quality of meaning and it acts more like a large formal group constitutive…of interpersonal cognitive frameworks in Mann Gulch suggests that with mediated communication than a processes, power dependencies, and the process of structuring itself may be small informal group with direct commu- contextual constraints” (Ranson, Hinings, more unstable than we realized. Struc- nication. and Greenwood, 1980: 1, 3). Structuring, turing, understood as constitutive then, consists of two patterns and the relations between meaning and frame- It is striking how little communication relationships between them. The first works, may be a deviation-amplifying occurred during the three and a half pattern, which Ranson et al. variously cause loop (Maruyama, 1963; Weick, hours of this episode. There was little described as informal structure, agency, 1979) capable of intensifying either an discussion during the noisy, bumpy or social construction, consists of interac- increase or decrease in either of the two plane ride, and even less as individuals tion patterns that stabilize meaning by connected elements. Typically, we see retrieved equipment scattered on the creating shared interpretive schemes. l instances of increase in which more north slope. After a quick meal together, refer to this pattern as shared provinces shared meanings lead to less elaborate people began hiking toward the river but of meaning, or meaning. The second frameworks of roles, which lead to further quickly got separated from one another. pattern, variously described as configur- developments of shared meaning, etc. Then they were suddenly turned around, ation, contextual constraints, or a vehicle What we fail to realize is that, when told to run for the ridge, and quickly ran that embodies dominant meanings, elements are tied together in this direct out of breath as they scaled the steep refers to a framework of roles, rules, manner, once one of them declines, this south slope. The minimal communication procedures, configured activities, and decline can also spread and become is potentially important because of the authority relations that reflect and facili- amplified as it does so. Fewer shared growing evidence (e.g., Eisenhardt, tate meanings. l refer to this second meanings lead to less elaborate frame- 1993: 132) that nonstop talk, both vocal pattern as structural frameworks of works, less meaning, less elaborate and nonverbal, is a crucial source of constraint, or frameworks. frameworks, and so on. Processes that coordination in complex systems that are mutually constitute also have the capabil- susceptible to catastrophic disasters. Meanings affect frameworks, which ity to mutually destroy one another. affect meaning. This is the basic point of The lack of communication, coupled with the growing body of work on structura- If structuration is treated as a deviation- the fact that this is a temporary group in tion (e.g., Riley, 1983; Poole, Seibold, amplifying process, then this suggests the early stages of its history, should and McPhee, 1985), understood as the the kind of structure that could have heighten the group’s vulnerability to mutual constitution of frameworks and prevented the Mann Gulch disaster. disruption. As Bass (1990: 637) put it, meanings (Ranson, Hinings, and Green- What people needed was a structure in “Groups that are unable to interact easily wood, 1980) or relations and typifications which there was both an inverse and a or that do not have the formal or informal (DiMaggio, 1991) or structures and direct relationship between role systems structure that enables quick reactions structuring (Barley, 1986). Missing in and meaning. This is the only pattern are likely to experience stress (Bass, this work is attention to reversals of that can maintain resilience in the face 1960). Panic ensues when members of structuration (Giddens, 1984). The use of crisis. The resilience can take one of a group lack superordinate goals—goals of descriptive words in structuration two forms. Assume that we start with an that transcend the self-interests of each theory such as “continually produced,” amplifying system like the one in Mann 68 Part 4 of 4

Gulch. The role system lost its structure, meaning temporarily, and as social the crew. Given the constantly changing which led to a loss of meaning, which led relations become clearer, their attention composition of the smokejumping crews, to a further loss of structure, and so on. shifts back to meanings. In the second the task largely structured their relations. This is the pattern associated with a scenario, when social relations decline, Simply acting in concert was enough, deviation-amplifying feedback loop in people pay more attention to meaning, and there was no need to know each which an initial change unfolds they ignore frameworks temporarily, and other well in addition. This social form unchecked in the same direction. One as meanings become clearer, attention resembles what Eisenberg (1990: 160) way to prevent this amplification is to shifts back to frameworks. Both scen- called nondisclosive intimacy, by which retain the direct relation between struc- arios illustrate operations of wisdom: In he meant relationships rooted in collec- ture and meaning (less role structure Meacham’s words, ignorance and know- tive action that stress “coordination of leads to less meaning, more structure ledge grow together. Either of these two action over the alignment of cognitions, leads to more meaning) but create an controlled patterns should reduce the mutual respect over agreement, trust inverse relation between meaning and likelihood of disaster in Mann Gulch. As over empathy, diversity over homoge- structure (less meaning, more structure, the smokejumpers begin to lose struc- neity, loose over tight coupling, and and vice versa). This inverse relationship ture they either also lose meaning, which strategic communication over unrestric- can be understood as follows: When alerts them to be more attentive to the ted candor.” Nondisclosive intimacy is a meaning becomes problematic and structure they are losing, or they gain sufficient ground for relating as long as decreases, this is a signal for people to individual meaning, which leads them to the task stays constant and the environ- pay more attention to their formal and realign structure. The second alternative ment remains stable. informal social ties and to reaffirm and/or may be visible in the actions taken by reconstruct them. These actions produce Dodge and Rumsey and Sallee. What the Mann Gulch disaster suggests more structure, which then increases is that nondisclosive intimacy may limit meaning, which then decreases the This may seem like a great deal of fret- the development of emotional ties that attention directed at structure. Puzzle- ting about one single word in Maclean’s keep panic under control in the face of ment intensifies attentiveness to the question, “structure.” What I have tried obstacles. Closer ties permit clearer social, which reduces puzzlement. to show is that when we transform this thinking, which enables people to find word from a static image into a process, paths around obstacles. For example, The other form of control arises when a we spot what looks like a potential for when Rumsey squeezed through a change in structure, rather than a change collapse in any process of social crevice in the ridge just ahead of the fire, in meaning, is responsible for counter- sensemaking that is tied together by he collapsed “half hysterically” into a acting the fluctuations in sensibleness. constitutive relations. And we find that juniper bush, where he would have soon In this variation, less structure leads to social sensemaking may be most stable burned to death. His partner Sallee more meaning, and more meaning then when it is simultaneously constitutive stopped next to him, looked at him coldly, produces more structure. The inverse and destructive, when it is capable of never said a word, and just stood there relationship between structure and increasing both ignorance and know- until Rumsey roused himself, and the meaning can be understood this way: ledge at the same time. That seems two then ran together over the ridge and When social ties deteriorate, people try like a fair return for reflecting on a down to a rock slide where they were harder to make their own individual single word. better able to move around and duck the sense of what is happening, both socially worst flames (Maclean, p. 107). Sallee’s and in the world. These operations Third, there is the follow-up question, Is surprisingly nuanced prodding of his increase meaning, and they increase the “outfit” the best way to describe the partner suggests the power of close ties tendency to reshape structure consistent smokejumpers? An outfit is normally to moderate panic. with heightened meaning. Alienation defined as “a group associated in an intensifies attentiveness to meaning, undertaking requiring close cooperation, One might expect that the less threaten- which reduces alienation. as a military unit” (Random House, ing the environment, the less important 1987: 1374). The smokejumpers are tied are relational issues in transient groups, What is common to both of these together largely by pooled interdepend- but as Perrow (1984) emphasized in his controlled forms is an alternation ence, since the job of each one is to normal accident theory, there are few between attention to frameworks and clear adjacent portions of a perimeter safe environments. If events are increas- attention to meanings. More attention to area around a blaze so that the fire stops ingly interdependent, then small one leads to more ignorance of the for lack of fuel. Individual efforts to clear unrelated flaws can interact to produce other, followed by efforts to correct this away debris are pooled and form a fire something monstrous. Maclean saw this imbalance, which then creates a new line. What is significant about pooled clearly at Mann Gulch: The colossal fire imbalance. In the first scenario, when interdependence is that it can function blowup in Mann Gulch was “shaped by meaning declines, people pay more without much cohesion (Bass, 1990: little screwups that fitted together tighter attention to frameworks, they ignore 622). And this is what may have trapped and tighter until all became one and the 69 Part 4 of 4 same thing—the fateful blowup. Such is by looking at a single set of actions is sense when time is extended, demands much of tragedy in modern times and doomed to failure because it does not change, and there is no formal leader at probably always has been except that reflect how needs change as a crisis the beginning of the episode. But there past tragedy refrained from speaking of unfolds, nor does it reflect how different is a leader in Mann Gulch, the foreman. its association with screwups and blow- coherent groupings form to meet the There is also a second in command and ups” (Maclean, 1992: 92). new needs. the remaining crew, which means there is a top (foreman), middle (second in Nondisclosive intimacy is not the only The team in the Andes had 10 weeks command), and bottom (remaining crew). alternative to “outfit” as a way to describe and changing threats of bleeding, If we take this a priori structure seriously, the smokejumpers. Smith (1983) argued hygiene, starvation, avalanche, expedi- then the Mann Gulch disaster can be that individual behaviors, perceptions of tion, rescue, and accounting, whereas understood as a dramatic failure of lead- reality, identities, and acts of leadership the team in Mann Gulch had more like ership, reminiscent of those lapses in are influenced by intergroup processes. 10 minutes and the increasingly singular leadership increasingly well documented Of special relevance to Mann Gulch is threat of being engulfed in fire. Part of by people who study cockpit/crew Smith’s reanalysis of the many groups the problem in Mann Gulch is the very resource management in aircraft that formed among the 16 members of inability for intergroup structures to form. accidents (e.g., Wiener, Kanki, and the Uruguayan soccer team who The inability to form subgroups within Helmreich, 1993). survived for 10 weeks in an inaccessible the system may be due to such things as region of the Chilean Andes mountains time pressure, the relative unfamiliarity The captain of an aircrew, who is analo- after their aircraft, carrying 43 people, of the smokejumpers with one another gous to a player-coach on a basketball crashed (see Read, 1974 for the original compared with the interdependent team (Hackman, 1993: 55) can often account of this event). Aside from the members of a visible sports team, the have his or her greatest impact on team eerie coincidence that both disasters inability to communicate, the articulation functioning before people get into a tight, involved 16 young males, Smith’s of a common threat very late in the time-critical situation. Ginnett (1993) has analysis makes the important point that smokejumpers’ exposure to Mann shown that aircraft captains identified by 16 people are not just an outfit, they are Gulch, and ambiguity about means that check airmen as excellent team leaders a social system within which multiple would clearly remove the threat, spent more time team building when the groups emerge and relate to one compared with the relative clarity of the team first formed than did leaders judged another. It is these intergroup relation- means needed by the soccer players to as less expert. Leaders of highly effec- ships that determine what will be seen deal with each of their threats. tive teams briefed their crewmembers on as acts of leadership and which people four issues: the task, crew boundaries, may be capable of supplying those acts. The point is, whatever chance the standards and expected behaviors In the Andes crash, demands shifted smokejumpers might have had to sur- (norms), and authority dynamics. from caring for the wounded, in which vive Mann Gulch is not seen as clearly Captains spent most time on those of two medical students took the lead, to if we view them as a single group rather the four that were not predefined by the acquiring food and water, where the team than as a social system capable of organizational context within which the captain became leader, to articulating differentiating into many different sets crew worked. Typically, this meant that that the group would not be rescued of subgroups. The earlier discussion of excellent captains did not spend much and could sustain life only if people virtual role systems suggested that an time on routine tasks, but less-excellent consumed the flesh of the dead, to exe- intergroup perspective could be simu- captains did. Crew boundaries were cuting and resymbolizing this survival lated in the head and that this should enlarged and made more permeable by tactic, to selecting and equipping an heighten resilience. Smith makes it clear excellent captains when, for example, expeditionary group to hike out and look that, virtual or not, intergroup dynamics they regarded the flight attendants, gate for help, and finally to finding someone affect survival, even if we overlook them personnel, and air traffic controllers as able to explain and rationalize their in our efforts to understand the group or members of the total flight crew. This decisions to the world once they had the “outfit.” contrasts with less-excellent captains, been rescued. who drew a boundary around the people As a fourth and final follow-up question, in the cockpit and separated them from What Smith shows is that this group of If there is a structure that enables people everyone else. 16 forms and reforms in many different to meet sudden danger, who builds and directions during its history, each time maintains it? A partial answer is Ken Excellent captains modeled norms that with a different coherent structure of Smith’s intergroup analysis, suggesting made it clear that safety, effective people at the top, middle, and bottom, that the needed structure consists of communication, and cooperation were each with different roles. What also many structures, built and maintained expected from everyone. Of special becomes clear is that any attempt to by a shifting configuration of the same interest, because so little communica- pinpoint the leader or to explain survival people. As I said, this perspective makes tion occurred at Mann Gulch, is how the 70 Part 4 of 4 norm, “communication is important,” talk less, listen more, and resort less to Barley, Stephen R. 1986. “Technology was expressed. Excellent crews expect “canned presentations.” as an occasion for structuring: one another to enact any of these four Evidence from observations of CT exchanges: “(1) I need to talk to you; Taken together, all of these team- scanners and the social order of (2) l listen to you; (3) I need you to talk building activities increase the probability radiology departments.” Administra- to me; or even (4) I expect you to talk to that constructive, informed interactions tive Science Quarterly, 31: 78-108. me” (Ginnett, 1993: 88). These four can still occur among relative strangers complement and operationalize the spirit even when they get in a jam. If we com- Barthol, R. P., and N. D. Ku. 1959. of Campbell’s social imperatives of trust, pare the leadership of aircraft captains “Regression under stress to first honesty, and self-respect. But they also to leadership in Mann Gulch, it is clear learned behavior.” Journal of Abnormal show the importance of inquiry, advo- that Wag Dodge did not build his team of and Social Psychology, 59: 134-136. cacy, and assertion when people do not smokejumpers in advance. Furthermore, understand the reasons why other people members of the smokejumper crew did Bass, Bernard M. 1960. Leadership, are doing something or ignoring some- not keep each other informed of what Psychology, and Organizational thing (Helmreich and Foushee, 1993: 21). they were doing or the reasons for their Behavior. New York: Harper. actions or the situational model they Issues of authority are handled differ- were using to generate these reasons. Bass, Bernard M. 1990. Bass and ently by excellent captains. They shift These multiple failures of leadership may Stogdill’s Handbook of Leadership. their behaviors between complete be the result of inadequate training, New York: Free Press. democracy and complete autocracy inadequate understanding of leadership during the briefing and thereafter, which processes in the late ‘40s, or may be Beach, Lee R., and Raanan Lipshitz. makes it clear that they are capable of attributable to a culture emphasizing 1993. “Why classical decision theory a range of styles. They establish compe- individual work rather than group work. is an inappropriate standard for evalu- tence and their capability to assume Or these failures of leadership may ation and aiding most human decision legitimate authority by doing the briefing reflect the fact that even the best leaders making.” In Gary A Klein, Judith in a rational manner, comfortably, with and the most team-conscious members Orasanu, Roberta Calderwood, and appropriate technical language, all of can still suffer when structures begin to Caroline E. Zsambok (eds.), Decision which suggests that they have given pull apart, leaving in their wake sense- Making in Action: Models and some thought to the upcoming flight and lessness, panic, and cosmological Methods: 21-35. Norwood, NJ: Ablex. have constructed a framework within questions. If people are lucky, and which the crew will work. interpersonally adept, their exposure to Bigelow, John. 1992. “Developing questions of cosmology is confined to an managerial wisdom.” Journal of Less autocratic than this enactment of episode. If they are not, that exposure Management Inquiry, 1: 143-153. their legitimate authority is their willing- stretches much further. Which is just ness to disavow perfection. A good about where Maclean would want us Brown, Richard Harvey. 1978. “Bureauc- example of a statement that tells crew- to end. racy as praxis: Toward a political members they too must take responsi- phenomenology of formal organiza- bility for one another is this: “I just want tions.” Administrative Science you guys to understand that they assign Quarterly, 23: 365-382. the seats in this airplane based on seniority, not on the basis of compe- Bruner, Jerome. 1983. In Search of tence. So anything you can see or do References Mind New York: Harper. that will help out, I’d sure appreciate hearing about it” (Ginnett, 1993: 90). Ancona, Deborah G., and David F. Bruner, Jerome. 1986. Actual Minds, Notice that the captain is not saying, l Caldwell. 1992. “Bridging the Possible Worlds. Cambridge, MA: am not competent to be the captain. boundary: External activity and perfor- Harvard University Press. Instead, the captain is saying, we’re all mance in organizational teams.” fallible. We all make mistakes. Let’s keep Administrative Science Quarterly, 37: Campbell, Donald T. 1990. “Asch’s an eye on one another and speak up 634-665. moral epistemology for socially shared when we think a mistake is being made. knowledge.” In Irwin Rock (ed.), The Asch, Solomon. 1952. Social Psychol- Legacy of Solomon Asch: Essays in Most democratic and participative is the ogy. Englewood Cliffs, NJ: Prentice- Cognition and Social Psychology: 39- captain’s behavior to engage the crew. Hall. 52. Hillsdale, NJ: Erlbaum. Briefings held by excellent captains last no longer than do those of the less- excellent captains, but excellent captains 71 Part 4 of 4

Daft, Richard L., and Norman B. Harper, Douglas. 1987. Working Know- Levi-Strauss, Claude. 1966. The Savage Macintosh. 1981. “A tentative explor- ledge: Skill and Community in a Small Mind. Chicago: University of Chicago ation into the amount and equivocality Shop. Chicago: University of Chicago Press. of information processing in organiza- Press. tional work units.” Administrative Maclean, Norman. 1992. Young Men Science Quarterly, 26: 207-224. Helmreich, Robert L., and Clayton and Fire. Chicago: University of Foushee. 1993. “Why crew resource Chicago Press. DiMaggio, Paul. 1991. “The micro-macro management? Empirical and theoreti- dilemma in organizational research: cal bases of human factors training in March, James G. 1989. Decisions and Implications of role-system theory.” In aviation.” In Earl L. Wiener, Barbara Organizations. Oxford: Blackwell. Joan Huber (ed.), Micro-macro G. Kanki, and Robert L. Helmreich Changes in Sociology: 76-98. (eds.), Cockpit Resource Management: Maruyama, Magorah. 1963. “The Newbury Park, CA: Sage. 3-45. San Diego: Academic Press. second cybernetics: Deviation- amplifying mutual causal process.” Eisenberg, Eric M. 1990. “Jamming: Helmreich, Robert L., Clayton H. American Scientist, 51: 164-179. Transcendence through organizing.” Foushee, R. Benson, and W. Russini. Communication Research, 17: 139-164. 1985. “Cockpit resource management: McDougall, William. 1920. The Group Exploring the attitude-performance Mind. New York: Putnam. Eisenhardt, Kathleen M. 1993. “High linkage.” Paper presented at Third reliability organizations meet high Aviation Psychology Symposium, Meacham, John A. 1983. “Wisdom and velocity environments: Common Ohio State University. the context of knowledge.” In D. Kuhn dilemmas in nuclear power plants, and J. A. Meacham (eds.), Contribu- aircraft carriers, and microcomputer Heydebrand, Wolf V. 1989. “New tions in Human Development, 8: 111- firms. ” In Karlene H. Roberts (ed.), organizational forms.” Work and 134. Basel: Karger. New Challenges to Understanding Occupations, 16: 323-357. Organizations: 117-135. New York: Miles, Ray E., and Charles C. Snow. Macmillan. Hirsch, Paul, Stuart Michaels, and Ray 1992. “Causes of failure in network Friedman. 1987. “ ‘Dirty hands’ vs. organizations.” California Management Foushee, H. Clayton. 1984. “Dyads and ‘clean models’. Is sociology in danger Review, 34(4): 53-72. triads at 35,000 feet.” American of being seduced by economics?” Psychologist, 39: 885-893. Theory and Society, 16: 317-336. Miller, Danny. 1990. The Icarus Paradox. New York: Harper. Freud, Sigmund. 1959. Group Psychol- Janowitz, Morris. 1959. “Changing ogy and the Analysis of the Ego. (First patterns of organizational authority: Mintzberg, Henry. 1983. Structure in published in 1922.) New York: Norton The military establishment.” Admini- Fives: Designing Effective Organiza- . strative Science Quarterly, 3: 473-493. tions. Englewood Cliffs, NJ: Prentice- Giddens, Anthony. 1984. The Constitu- Hall. tion of Society. Berkeley: University of King, Jonathan B. 1989. “Confronting California Press. chaos.” Journal of Business Ethics, Morgan, Gareth, Peter J. Frost, and 8: 39-50. Louis R. Pondy. 1983. “Organizational Ginnett, Robert C. 1993. “Crews as symbolism.” In L. R. Pondy, P. J. Frost, groups: Their formation and their Klein, Gary A. 1993. “A recognition- G. Morgan, and T. C. Dandridge (eds.), leadership.” In Earl L. Wiener, Barbara primed decision (RPD) model of rapid Organizational Symbolism: 3-35. G. Kanki, and Robert L. Helmreich decision making ” In Gary A. Klein, Greenwich, CT: JAI Press. (eds.), Cockpit Resource Management: Judith Orasanu, Roberta Calderwood, 71-98. San Diego: Academic Press. and Caroline E. Zsambok (eds.), Neisser, Ulric, and Eugene Winograd. Decision Making in Action: Models 1988. Remembering Reconsidered: Hackman, J. Richard. 1993. “Teams, and Methods: 138-147. Norwood, NJ: Ecological and Traditional Approaches leaders, and organizations: New Ablex. to the Study of Memory. New York: directions for crew-oriented flight Cambridge University Press. training.” In Earl L. Wiener, Barbara Lanir, Zvi. 1989. “The reasonable choice G. Kanki, and Robert L. Helmreich of disaster: The shooting down of O’Hare, David, and Stanley Roscoe. (eds.), Cockpit Resource Management: the Libyan airliner on 21 February 1990. Flightdeck Performance: The 47-69. San Diego: Academic Press . 1973.” Journal of Strategic Studies, Human Factor. Ames, IA: lowa State 12: 479-493. University Press.

72 Part 4 of 4

Orasanu, Judith, and Terry Connolly. Reed, M. 1991. “Organizations and Weick, Karl E. 1985. “Cosmos vs. chaos: 1993. “The reinvention of decision rationality: The odd couple.” Journal of Sense and nonsense in electronic making.” In Gary A. Klein, Judith Management Studies, 28: 559-567. contexts.” Organizational Dynamics, Orasanu, Roberta Calderwood, and 14(Autumn): 50-64. Caroline E. Zsambok (eds.), Decision Riley, Patricia. 1983. “A structuration- Making in Action: Models and alist account of political culture.” Weick, Karl E. 1990. “The vulnerable Methods: 3-20. Norwood, NJ: Ablex. Administrative Science Quarterly, 28: system: Analysis of the Tenerife air 414-437. disaster.” Journal of Management, Perrow, Charles. 1984. Normal Acci- 16: 571-593. dents. New York: Basic Books. 1986. Runkel, Phillip J., and Joseph E. Complex Organizations, 3rd ed. New McGrath. 1972. Research on Human Weick, Karl E., and Karlene H. Roberts. York: Random House. Behavior. New York: Holt, Rinehart, 1993. “Collective mind in organiza- and Winston. tions: Heedful interrelating on flight Poole, M. Scott, David R. Seibold, and decks.” Administrative Science Robert D. McPhee. 1985. “Group Schutz, William C. 1961. “The ego, FIRO Quarterly, 38: 357-381. decisionmaking as a structurational theory and the leader as completer.” process.” Quarterly Journal of Speech, In Louis Petrullo and Bernard M. Bass Westley, Frances R. 1990. “Middle mana- 71: 74-102. (eds.), Leadership and Interpersonal gers and strategy: Microdynamics of Behavior: 48-65. New York: Holt, inclusion.” Strategic Management Pyne, Stephen. 1989. Fire on the Rim. Rinehart, and Winston . Journal, 11: 337-351. New York: Weidenfeld & Nicolson. Scott, W. Richard. 1987. Organizations: Westrum, Ron. 1982. “Social intelligence Random House. 1987. Dictionary of the Rational, Natural, and Open Systems. about hidden events.” Knowledge. 3: English Language, 2d ed.: Unabridged. Englewood Cliffs, NJ: Prentice-Hall. 381-400. New York: Random House. Smith, Ken K. 1983. “An intergroup White, S. K. 1988. The Recent Work of Ranson, Stewart, Bob Hinings, and perspective on individual behavior.” Jurgen Habermas: Reason, Justice, Royston T. Greenwood. 1980. “The In J. Richard Hackman, Edward E. and Modernity. Cambridge: structuring of organizational Lawler, and Lyman M. Porter (eds.), Cambridge University Press. structures.” Administrative Science Perspectives on Behavior in Quarterly, 25: 1-17. Organizations: 397-408. New York: Wiener, Earl L., Barbara G. Kanki, and McGraw-Hill. Robert L. Helmreich. 1993. Cockpit Read, P. P. 1974. Alive. London: Pan Resource Management. San Diego: Books. Taylor, Shelby E. 1989. Positive Illusions. Academic Press. New York: Basic Books. Reason, James. 1990. Human Error. Wiley, Norbert. 1988. “The micro-macro New York: Cambridge University Weick, Karl E. 1979. The Social problem in social theory.” Sociological Press. Psychology of Organizing, 2d ed. Theory, 6: 254-261. Reading, MA: Addison-Wesley.

73 Part 4 of 4

Notes

74 Part 4 of 4

Notes

75 Part 4 of 4

Additional single copies of this document may be ordered from:

USDA-FS, Missoula Technology & Development Center Building 1, Fort Missoula Missoula, MT 59804-7294 Phone: (406) 329-3900 Fax: (406) 329-3719

For further technical information, contact Ted Putnam at the above address.

Phone: (406) 329-3965 Fax: (406) 329-3719 DG: T.Putnam:R01A E-mail: /s=t.putnam/[email protected]

Library Card

Putnam, Ted. 1995. Findings from the wildland firefighters human factors workshop; 12-16 June 1995; Missoula, MT. Tech. Rep. 9551-2855-MTDC. Missoula, MT: U.S. Department of Agriculture, Forest Service, Missoula Technology and Development Center. 74 p.

This document includes an overview and the findings of a workshop held to address the human factors involved in firefighter safety. The deaths of 34 firefighters during the 1994 fire season, including the 14 firefighters who died on the South Canyon Fire in Colorado, helped point out the need for the workshop. Participants explored firefighter psychology, interactions among firefighters and among fire crews, and better ways to organize firefighters. Appen- dixes include the four keynote presentations and two related papers addressing human factors involved in fire- fighter safety.

Keywords: group behavior, organizational behavior, psychological factors, safety, sociological analysis.

-End of Document- 76