Upload your Term Project Paper here. You may write on any topic related to technology security.You must produce a minimum of a 10 pages paper. You must use a minimum of 5 references.- Assure you are c

http://hfs.sagepub.com/ Ergonomics Society of the Human Factors and Human Factors: The Journal http://hfs.sagepub.com/content/52/2/295 The online version of this article can be found at:   DOI: 10.1177/0018720810371689 published online 23 July 2010 2010 52: 295 originally Human Factors: The Journal of the Human Factors and Ergonomics Society Jamie C. Gorman, Nancy J. Cooke and Polemnia G. Amazeen Training Adaptive Teams     Published by: http://www.sagepublications.com On behalf of:

    Human Factors and Ergonomics Society can be found at:

Society Human Factors: The Journal of the Human Factors and Ergonomics Additional services and information for         http://hfs.sagepub.com/cgi/alerts Email Alerts:   http://hfs.sagepub.com/subscriptions Subscriptions:   http://www.sagepub.com/journalsReprints.nav Reprints:   http://www.sagepub.com/journalsPermissions.nav Permissions:   http://hfs.sagepub.com/content/52/2/295.refs.html Citations:   at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from What is This?   - Jul 23, 2010 OnlineFirst Version of Record  - Sep 14, 2010 Version of Record >> at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training Adaptive T\Geams Jamie C. G\frman and Nancy J. C\f\f\be, Arizona State Uni\lversity–Polyte\fhni\f,\l Mesa, Arizona, and P\flemnia G. Amazeen\G, Arizona State Uni\lversity, \bempe, Arizona Objective: We report an experiment in which three training approaches are compared with the goal o\f training adaptive teams. Background: Cross\btraining is an established method in which team members are trained with the goal o\f building shared knowledge.

Perturbation training is a new method in which team interactions are constrained to pro\b vide new coordination experiences during task acquisition. These two approaches, and a more traditional procedural approach, are compared. Met\fod: Assigned to three training conditions were 26 teams. Teams \flew nine simulated uninhabited air vehicle missions; three were critical tests o\f the team’s ability to adapt to novel situations. Team per\for\b mance, response time to novel events, and shared knowledge were measured. Results:

Perturbation\btraine\ld teams signi\ficantly outper\formed teams in the other conditions in two out o\f three critical test missions. Cross\btraining resulted in signi\ficant increases in shared teamwork knowledge and highest mean per\formance in one critical test.

Procedural training led to the least adaptive teams. \bonclusion: Perturbation training allows teams to match coordination variability during training to demands \for coordina\b tion variability during posttraining per\formance. Although cross\btraining has adaptive bene\fits, it is suggested that process\boriented approaches, such as perturbation training, can lead to more adaptive teams. Application: Perturbation training is amenable to sim\b ulation\bbased training, where perturbations provide interaction experiences that teams can trans\fer to nov\lel, real\bworld situa\ltions. Address correspondence to Jamie C. Gorman, Cognitive Engineering Research Institute, 5810 S. Sossaman Rd., Ste. 106, Mesa, AZ 85212; [email protected]. HUMAN FACTORS, Vol. 52, No. 2, April 2010, pp. 295–307.

DOI: 10.1177/00187208103716\l89. Copyright © 201\l0, Human Factors a\lnd Ergonomics Society. INTRODUCTION In settings ranging \from business and manu\b \facturing to military and medical operations, there are many tasks that are too cognitively demanding to be per\formed by individuals working alone. An example is a surgical task, which requires a set o\f highly trained individuals, including two surgeons, an anesthesiologist and two nurses, each o\f whom brings di\f\ferent cognitive capabilities to the team. But it is not enough to bring together a set o\f highly trained individuals. To \function as a team, individuals must coordinate their activities. Adaptive teams have the ability to coordinate their activities not only under routine conditions but also under novel conditions \for which they have not been explicitly trained. Adaptation is the altering o\f structure in accordance with changes in the environment. Because they have the ability to change their interactions to match the changing demands o\f the environment, adaptive teams can per\form at a high level under novel task conditions.

A number o\f relatively recent tragic system \failures can be at least partially attributed to poor coordination o\f a team\blevel response to environmental uncertainty. System \failures attrib\b utable to poor team skills at Three Mile Island and Chernobyl (Gaddy & Wachtel, 1992), social pathogens behind the 1986 launch decision o\f the space shuttle Challenger (Vaughan, 1996), and lack o\f communication in the Operation Provide Com\fort \friendly \fire incident (Gorman, Cooke, & Winner, 2006; Snook, 2002) each implicate, in di\f\ferent ways, de\ficiencies in inter\b action and coordination that result in a \failure to adapt to changes in the task environment. These incidents, and others like them, highlight the need \for training that addresses limitations and SPECIAL ISSUE at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 296 April 2010 - Human Factors de\ficiencies at the team level in responding to novel patterns o\f e\lvents and threats.

Approaches to Tra\fn\fng Adapt\f\be Teams A challenging problem \for training team cognition (i.e., training teams to decide, plan, think, and act as an integrated unit; Cooke, Gorman, & Winner, 2007) is how to balance training \for high per\formance under routine task conditions with training to adapt to novel task demands (Marks, Zaccaro, & Mathieu, 2000; Stachowski, Kaplan, & Waller, 2009).

These training goals can be approached with varying theoretical motives. In this article, we report an experiment in which three training approaches, each with a di\f\ferent underlying theoretical motive, were investigated with the goal o\f training teams that per\form at a high level under novel task conditions. The training approaches include cross\btraining, procedural training, and pert\lurbation training. Cross-training. In cross\btraining, team mem\b bers are trained on each other’s roles and respon\b sibilities (e.g., Blickensder\fer, Cannon\bBowers, & Salas, 1998). Cross\btraining is theoretically aligned with the idea that team cognition is the shared knowledge o\f the team members and is \found widely in the team training literature (see Salas et al., 2008, and Salas, Nichols, & Driskell, 2007, \for recent meta\banalyses). The goal o\f cross\btraining is the development o\f shared, or interpositional, knowledge (Cannon\bBowers, Salas, Blickensder\fer, & Bowers, 1998; Cooke et al., 2003; Volpe, Cannon\bBowers, Salas, & Spector, 1996). Positional clari\fication (receiv\b ing in\formation on other roles), positional mod\b eling (observing other roles), and positional rotation (\firsthand experience per\forming di\f\b \ferent roles) (Blickensder\fer et al., 1998), which are types o\f cross\btraining, have been e\f\fective in the development o\f shared knowledge, ulti\b mately improving coordination and team per\for\b mance (e.g., Marks, Sabella, Burke, & Zacarro, 2002). Cross\btraining has a \firm empirical grounding and a record o\f success in the team training literature, making it a good point o\f com\b parison with the o\lther approaches in\l this study. One o\f the potential bene\fits o\f cross\btraining \for shared knowledge is a high level o\f team per\formance under stress (high workload, time pressure). Drawing on shared knowledge, team members anticipate each other’s needs to com\b municate e\f\ficiently, or coordinate implicitly, under stress (Cannon\bBowers et al., 1998; Entin & Ser\faty, 1999; Stout, Cannon\bBowers, Salas, & Milanovich, 1999). It is thought that shared expectations, resulting \from the development o\f shared knowledge, allow team members to gen\b erate predictions \for appropriate behavior under novel conditions, enabling them to quickly adapt to the changing demands o\f the task environ\b ment (Fiore, Salas, & Cannon\bBowers, 2001). A possible drawback o\f a shared set o\f expecta\b tions, however, is the habituation o\f team member interaction, which could result in dys\functional consequences i\f the situation is highly novel (e.g., Gersick & Hackman, 1990; Gorman, Cooke, & Winner, 2006).

Whereas cross\btraining is \feasible \for rela\b tively small, homogeneously skilled teams, it can become impractical as teams grow in diver\b sity and size. For example, it would be impracti\b cal to cross\btrain the surgeon and nurse positions o\f an emergency room team (Cooke et al., 2003; Marks et al., 2002). Also, cross\btraining may negatively impact individual\blevel per\formance due to the demands o\f training \for multiple team member roles (Cannon\bBowers et al., 1998), which is problematic as teams grow in size.

Although there are adaptive bene\fits o\f cross\b training, there are practical limitations to its applicability. Procedural training. Procedural training is a \form o\f process training in which operators in complex systems are positively rein\forced (through \feedback) to \follow a standard sequence o\f actions (a procedure) each time a particular stimulus is encountered. The assumption behind procedural training is that i\f the procedure is always \fol\b lowed, then errors resulting \from human inter\b action will be reduced and per\formance will be enhanced, particularly under conditions o\f stress and high workload (e.g., Hockey, Sauer, & Wastell, 2007; Sauer, Burkholter, Kluge, Ritzmann, & Schuler, 2008). Procedural train\b ing is widely used in aviation, military, medical, manu\facturing, and business settings, in which deviations \from complicated procedures can be catastrophic. The prevalence o\f procedural train\b ing \for coordination in highly critical team tasks at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 297 (e.g., emergency response; Ford & Schmidt, 2000; Stachowski et al., 2009) make it a good point o\f comparison \for the other training meth\b ods in this study. Procedural training is compatible with the concept o\f “overlearning”: continuation o\f prac\b tice beyond mastery that leads to automatic responding. Drilling a standard team interaction pattern, \for a speci\fic class o\f event, over the entirety o\f training can lead to an automatic response that a team can rely on under stress.

The goal o\f procedural training, as operational\b ized in the current study, is to overlearn a team coordination procedure. Ideally, due to over\b learning, procedurally trained teams per\form under stress by automatically (re\flexively) reacting with an a\l priori coordinate\ld response. Procedural training does not impose the practical limitations o\f cross\btraining but may limit a team’s ability to trans\fer training to novel situations. Similar to the concept o\f a “set e\f\fect” (Luchins, 1942), procedural training may set teams up to coordinate in a routine \fashion under a novel condition. We argue, there\fore, that like habituation (Gersick & Hackman, 1990), rigid proceedure\b\following\l during task acquisition can lead to poor per\formance when posttraining conditions do not \lmatch training con\lditions. Perturbation training. Perturbation training is a \form o\f process training introduced in this study. Adopted \from the dynamic systems liter\b ature, a perturbation is an extrinsic application o\f \force that brie\fly disrupts a dynamic process, \forcing the reacquisition o\f a new stable trajec\b tory, and is typically used to probe the stability o\f that process (Gorman, Amazeen, & Cooke, in press). The concept o\f perturbation can be applied to team training by disrupting standard coordination procedures multiple times during task acquisition, \forcing teams to coordinate in novel ways to achieve their objective. Unlike training in which the situation or objectives are varied (e.g., training \for low\b vs. high\b\frequency circumstances), in perturbation training, critical coordination links are disrupted while the team objective remains constant. The goal o\f pertur\b bation training is to counteract habituation and procedural rigidity associated with team interac\b tions—possible outcomes o\f cross\btraining and procedural training, respectively—allowing teams to acquire \flexible interaction processes that will trans\fer to nov\lel task conditions.\l Perturbation training is theoretically inspired by \findings in the motor\b and verbal\blearning literatures that suggest that introducing di\f\ficul\b ties \for the learner, such as practice condition variability, \facilitates per\formance under novel posttraining conditions (Schmidt & Bjork, 1992). Perturbation training thus shares some \features o\f motor schema theory (Schmidt, 1975) and desirable di\f\ficulties (Bjork, 1994) but \for coordination and \for teams. According to motor schema theory, varying the conditions o\f practice during motor skill acquisition enhances the “rules” that relate movements to external task demands. In verbal learning, desir\b able di\f\ficulties are unpredictable and variable conditions o\f practice that cause di\f\ficulty \for the learner but ultimately enhance the trans\fer o\f concepts to new contexts. Whereas those approaches employ equally probable but ran\b domly varying training conditions to introduce practice condition variability, perturbation training employs abrupt but \focused disruptions to team coordinati\lon. Bjork (1994) argued that varying practice conditions exercises more elaborate encoding and retrieval processes needed in the posttraining environment. Perturbation training extends this idea to team process: When coordination is per\b turbed, all team member interactions (not just those directly a\f\fected by the perturbation) must readjust to accommodate the perturbation in such a manner that the team objective is never\b theless met (Turvey, 1990). We suggest that simi\b lar to the e\f\fects o\f practice condition variability, perturbations exercise the team processes needed to adapt in the po\lsttraining environ\lment. A major limitation o\f perturbation training is that it has not previously been applied and its e\f\fec\b tiveness is unknown. An experiment described by Gorman, Cooke, Pedersen, et al. (2006) provided some empirical grounding \for perturbation training. Teams were initially trained and per\b \formed a repetitive command\band\bcontrol \l task during a 3\bhr experimental session. Participants returned \for a second session a\fter a retention interval, a\fter which they were either intact (kept the same team members) or mixed (same role on the team but di\f\ferent team members). at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 298 April 2010 - Human Factors As expected, intact teams outper\formed mixed teams under routine conditions, but the e\f\fect was short\blived. Mixed teams, however, per\b \formed better on tests o\f situation awareness, had higher process ratings (Gorman, Cooke, Pedersen, et al., 2006), and had more \flexible coordination dynamics (Gorman et al., in press).

Those bene\fits were not attributable to increased shared knowledge or procedural rigidity but to increased variatio\ln in interaction e\lxperience.In some sense, mixing up the team members \ferturbed rigid coordination patterns, ultimately leading to a more \flexible and adaptive team.

Perturbation training, as operationalized in the current study, does not involve mixing team members but was designed have the same e\f\fect.

By disrupting standard coordination procedures during task acquisition, perturbation training increases interaction experience in intact command\band\bcontrol tea\lms.

The Current Study We compared the three di\f\ferent training approaches in an uninhabited air vehicle (UAV) simulator with the goal o\f producing teams that per\form at a high level under novel task condi\b tions and that respond rapidly to novel events. In the UAV task, three team members (navigator, photographer, and pilot) coordinate to take pic\b tures o\f stationary ground targets. Three training protocols were developed \for cross\btraining, perturbation training, and procedural training o\f UAV teams. The \following hypotheses are based on prior results and existing literature. Hy\fothesis \b: By \focusing on introducing varied inter\b action experiences during task acquisition, perturba\b tion training will result in per\formance scores and response times to novel events that are as good as or better than cross\btraining and superior to procedural training.

Hy\fothesis 2: Because o\f its \focus on training team members to know each other’s roles and responsi\b bilities, cross\btraining will result in higher levels o\f shared knowledge compared with both procedural and perturbation training.

Hy\fothesis 3: By training teams to rigidly \follow a procedure, procedural training will result in the least adaptive teams (i.e., poor per\formance and slow response to novel events) compared with both perturbation and cross\btraining. METHOD Part\fc\fpants We recruited 32 three\bperson teams (96 par\b ticipants) \for participation \from Mesa, Arizona, and surrounding areas. The team members had no prior experience working together. Participants ranged in age \from 18 to 54 ( M = 28), and 71 were male. The experiment occurred during two 3\b to 4\bhr sessions. Because o\f scheduling con\flicts \for Session 2, a total o\f 26 teams (78 participants) completed both experimental sessions. Participants were paid $10 per hour, and each member o\f the highest\bper\forming \l team received a $1\l00 bonus. Mater\fals and Appara\ttus The experiment was conducted in a UAV synthetic task environment (UAV\bSTE) \for teams (Cooke & Shope, 2005). Each o\f the three team members was seated at a workstation in \front o\f three computer monitors with a key\b board and a mouse. To interact, team members wore aviation\bquality headsets and communi\b cated by holding down push\bto\btalk buttons.

The workstations were located in the same room, con\figured in a U shape with team members backs to each other. With the team members donning headsets, the UAV\bSTE did not a\f\ford \face\bto\b\face interac\ltion. The team’s task was to take reconnaissance photographs o\f stationary ground targets during a series o\f nine 40\bmin missions divided across two experimental sessions. There were 11 to 12 targets per mission except \for one high\b workload mission that had 20 targets. The three team member roles—navigator, photographer, and pilot—were each associated with di\f\ferent, yet interdependent tasks, in\formation resources, and needs. Measures Team \ferformance. Per\formance was mea\b sured \for each UAV\bSTE mission as the weighted composite o\f several team\blevel mission param\b eters, including number o\f missed targets, time to process targets, and time spent with unaddressed warnings and alarms. Cooke, Gorman, Pedersen, et al. (2007) report the parameter weights, which were established in previous experiments to maximize score sensitivity. Teams started each at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 299 mission with a per\formance score o\f 1,000, and points were subtracted on the basis o\f \final values o\f the mission parameters. This team per\for\b mance score has been validated against other measures o\f team process and per\formance (Cooke, Gorman, Duran, & Taylor, 2007; Cooke, Gorman, Pedersen, \let al., 2007).Res\fonse time to novel events. Novel events were introduced within UAV missions by intro\b ducing roadblocks. Roadblocks are novel changes in the task environment that have to be jointly recognized by two or more team members who take action to overcome them (e.g., a new target is introduced, equipment \fails, an enemy threat appears; Cooke, Gorman, & Rowe, 2009; Gorman, Cooke, & Winner, 2006). Time to over\b come roadblocks, de\fined as the time \from the initiation o\f the roadblock to the time that action is taken that overcomes the roadblock, was the measure o\f response time to novel events. Inter\fositional taskwork knowledge. This measure assessed a team’s average knowledge o\f the taskwork associated with the other two roles. To measure taskwork knowledge, related\b ness ratings (1 = com\fletely unr elated to 5 = com\fletely r elated) were elicited \for 55 pairs o\f concepts \from the UAV task (e.g., airspeed, altitude). Individual team member ratings were analyzed using the Path\finder algorithm (Schvaneveldt, 1990), which translates related\b ness ratings across pairs o\f concepts into a graph\b ical network representation o\f conceptual interrelatedness. Individual networks were compared with expert role re\ferent networks. The re\ferents were derived empirically \from the top \five individual per\formers at each role in previous UAV\bSTE experiments (Cooke, Gorman, Pedersen, et al., 2007). Each team member was scored against the other two role re\ferents on the basis o\f the pro\b portion o\f shared links (0 = n o similarity to 1 = ex actly similar ). Team\blevel interpositional task\b work knowledge was taken as the average o\f these two scores across each o\f the three team members. Scores closer to 1 indicated a higher level o\f interpositional taskwork knowledge across team members. Inter\fositional teamwork knowledge. This measure assessed a team’s average knowledge o\f the teamwork associated with the other two roles.

Interpositional teamwork knowledge was elicited with the use o\f a questionnaire that consisted o\f 16 items related to which communications were nec\b essary to achieve a given scenario goal (e.g., “For a priority target, must the photographer communi\b cate camera settings to the navigator, the pilot, or both?”). Items that were necessary had to be indi\b cated by individual team members using check marks. To calculate teamwork knowledge, indi\b vidual responses were compared with role\bspeci\fic answer keys that were generated by experimenters \familiar with the task, and points were awarded \for correct answers (Cooke, Gorman, Pedersen, et al., 2007).

To measure interpositional teamwork knowl\b edge, each team member was scored on the basis o\f the proportion correct relative to the answer key \for each o\f the other two roles. Interpositional teamwork knowledge was calculated as the aver\b age number o\f these two scores across the three team members. Scores closer to 1 indicated a higher level o\f interpositional teamwork knowl\b edge across team m\lembers. Procedure When participants arrived \for the \first session, they were randomly assigned to a team member role and the team was assigned to one o\f the three training conditions. Participants received approximately 1 hr 45 min o\f training via three PowerPoint training modules and a hands\bon training mission. The \first two PowerPoint modules were identical \for all training condi\b tions and covered the general task and inter\face.

The third module and the hands\bon training mis\b sion di\f\fered on the basis o\f training condition.

(Procedures \for each training condition are described in the \fo\lllowing section.) Teams then completed Missions 1 through 5.

Knowledge measures were taken a\fter Mission 1. Missions 2 through 4 were condition\bspeci\fic \l training missions (Table 1). The \first roadblock was introduced dur\ling Mission 5, the\l \first post\b training mission. The roadblock consisted o\f cutting communication \from the navigator to the pilot \for 5 min, during which teams had to reroute navigator\bto\bpilot communications through the photographer to overcome the roadblock. Mission 5 was the \first o\f three criti\b cal missions that tested the teams’ ability to per\b \form under novel conditions. The completion o\f Mission 5 concluded\l the \first session.\l at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 300 April 2010 - Human Factors Teams returned a\fter 8 to 10 weeks \for the second session. All participants received re\fresher training on the so\ftware inter\faces, a\fter which teams completed \four additional UAV\bSTE mis\b sions. Roadblocks were introduced during each mission. Mission 6 was the second critical mis\b sion, which tested retention o\f team skill a\fter the break and included the second roadblock (a disguised target was hidden on the navigator map and photographer target list; teams had to recognize and photograph the target to over\b come the roadblock). Knowledge was measured \for a second time a\fter Mission 6 as a test o\f knowledge retention\l. Teams then completed their \final three mis\b sions (Missions 7 through 9). Mission 9 was the high\bworkload mission, in which the rate o\f tar\b gets per minute was almost doubled \from 0.28 to 0.5 and teams were exposed to a two\bpart roadblock (communication channel cut \for 5 min \from pilot to navigator and \from navigator to pilot; teams had to reroute their communica\b tions through open channels to overcome the roadblock). Mission 9 was the third o\f the three critical test missi\lons. Tra\fn\fng Procedure Cross-training. For the third PowerPoint train\b ing module, team members in the cross\btraining condition received training on the other two roles (positional clari\fication). Teams then completed a short training mission \followed by approximately 15 min o\f hands\bon experience per\forming all team member roles (positional rotation). A\fter Missions 2 through 4, teams in the cross\btraining condition were prompted to discuss how well they per\formed and to plan \for the next mission. Procedural training.

For the third PowerPoint training module, teams in the procedural training condition received train\b ing on the standard UAV\bSTE target photo\b graphing procedure: (a) The navigator provides target in\formation to the pilot, (b) the pilot and photographer negotiate altitude and air\b speed \for that target, and (c) the photogra\b pher provides \feedback on the status o\f the target photograph (Figure 1). Teams then completed a short training mission \followed by approximately 15 min o\f hands\bon train\b ing using the target photographing proce\b dure. A\fter Missions 2 through 4, teams in the procedural condition received experimenter \feedback on deviations \from the standard procedure. During training, team members in the procedural condition were provided with a hard copy o\f the target photographing procedure. Perturbation training. Teams in the pertur\b bation training condition received \filler PowerPoint training on the history and current uses o\f UAVs. Teams then completed a short training mission \followed by approximately 15 min o\f communications system testing in which they identi\fied the source o\f static in the UAV\bSTE communication system (i.e., which push\bto\btalk button was emanating static).

This training exercise provided experience on the use o\f multiple communication paths.

During Missions 2 through 4, teams in the per\b turbation condition received perturbations to the target\bphotographing procedure (Table 2) as they attempted to photograph targets.

Perturbations were less general than roadblocks and \forced teams to adjust speci\fic interactions TABLE 1: Experimental Pro\fedure Session 1 Session 2 Initial parti\fipant training Refresher training Mission 1 Mission 6 (retention test and se\fond roadblo\fk) C Knowledge measuresKnowledge measures Mission 2 \b Mission 7 Mission 3 \b Mission 8 Mission 4 \b Mission 9 (high workload and two-part roadblo\fk) C Mission 5 (first roadblo\fk) C Debriefing Note. \b = \fondition-spe\fifi\f \ltraining mission; \lC = \friti\fal test missi\lon. at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 301 relative to the in\formation\bnegotiation\b\feedback procedure (Figure 1). RESULTS O\f the 26 teams that completed the experi\b ment, there were 10 teams in the procedural condition and 8 teams each in the cross\btrain\b ing and perturbation conditions. Previous experiments in the UAV\bSTE exhibited low between\bsubjects power with a = .05 (M = .11, SD = .05) on tests o\f team per\formance due to small sample size. To increase statistical power, a signi\ficance level o\f a = .10 was used. For planned critical test mission com\b parisons, two conditions were pooled to \form a comparison against a single condition. These planned comparisons also served to increase power. Team Performance Team per\formance results are summarized in Table 3 and graphed in Figure 2. Team per\for\b mance was analyzed using a 3 (training) × 9 (mission) mixed ANOVA. The main e\f\fect o\f mission was signi\ficant, F(8, 184) = 22.14, \f < .001, MSE = 3535.25, η 2 = .53. No other e\f\fects in the omnibus tes\lt were signi\ficant.\l A repeated contrast on mission revealed that per\formance increased signi\ficantly during initial per\formance acquisition until Mission 4. There was a signi\fi\b cant drop in per\formance at Mission 6, a\fter the retention interval. Per\formance then improved signi\ficantly as teams reacquired the task, until Mission 8. Task reacquisition was \followed by a signi\ficant drop in per\formance at Mission 9, the high\bworkload mission. These results rein\force Missions 5, 6, and\l 9 as critical miss\lions. Figure \b. Standard photographing procedure \for the uninhabited air vehicle synthetic task environment.

TABLE 2: Perturbations to the Standard Uninhabited Air Vehi\fle Syntheti\f \bask Environment \barget- Photographing Pro\fedure Used for Perturbation \braining During Missions 2 \bhrough 4 Link in the Pro\fedure Perturbation Method of Introdu\fing Perturbatio\ln When Introdu\fed Information Photographer must provide target information to pilot Experimenter \falls in new target restri\ftions to photographer and disables \famera until restri\ftions are \fommuni\fated to pilot On\fe in Mission 2 On\fe in Mission 3 \bwi\fe in Mission 4 Negotiation Navigator/pilot must negotiate airspeed/altitude Experimenter \falls in new airspeed/altitude of \furrent target to navigator On\fe in Mission 2 On\fe in Mission 3 \bwi\fe in Mission 4 Feedba\fk Photographer does not provide feedba\fk to navigator and pilot Experimenter \falls in status of target photo to navigator and pilot and \futs all photographer \fommuni\fations On\fe in Mission 2 \bwi\fe in Mission 3 \bwi\fe in Mission 4 at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 302 April 2010 - Human Factors Planned comparisons were per\formed \for each o\f the three critical test missions to address which training condition resulted in the highest per\formance under novel conditions. As shown in Figure 2, perturbation\btraine\ld teams exhib\b ited better mean per\formance in two o\f the three critical test missions (Missions 5 and 9), whereas cross\btrained teams exhibited better mean per\formance in one o\f the critical test missions (Mission 6). Per\formance o\f the perturbation\btrained teams at Mission 5 ( M = 50 0.37, SD = 50.93) was signi\ficantly better than the other two conditions ( M = 459.31, SD = 54 .91), F(1, 24) = 3.23, \f = .085, MSE = 2892.21, η 2 = .12. Per\formance o\f perturbation\btrained t eams at Mission 9 ( M = 442.24, SD = 36.83) was also signi\ficantly better than the other two condi\b tions ( M = 379.62, SD = 60.00), F(1, 24) = 7.37, \f = .012, MSE = 2945.49, η 2 = .24. The cross\b t raining per\formance advantage at Mission 6 was not signi\ficant, F(1, 24) = .76, \f = .392, MSE = 83 79.75, η 2 = .03. TABLE 3: Mean \beam Performan\fe by \braining Condition Mission Cross-\brained Perturbation Pro\fedural 1 345.04 (65.80) 342.78 (54.23) 316.92 (78.88) 2 383.18 (72.89) 409.33 (80.94) 373.90 (65.65) 3 422.58 (74.39) 463.39 (80.69) 439.92 (54.71) 4 450.54 (77.71) 483.76 (59.83) 447.83 (54.26) 5 446.40 (64.41) 500.37 (50.93)* 469.63 (46.92) 6 435.99 (54.48) 380.30 (166.10) 383.76 (100.91) 7 477.79 (77.32) 471.77 (75.38) 421.38 (86.88) 8 513.25 (70.94) 547.06 (47.86) 502.47 (58.60) 9 389.13 (76.39) 442.24 (36.83)* 372.02 (46.00) Note. Standard deviations in pa\lrentheses.

*p < .10. 300 350 400 450 500 550 600 1 3 4 5 6 7 8 9 2 Mission Team Performance Cross-Trainin\f Proce\bural Perturbation Acquisition Reacquisition Critical Test 1 First Roa\bblock Critical Test 2 Retention Critical Test 3 Hi\fh Workloa\b Figure 2. Team per\formance \for each training condition across missions. Adapted \from Gorman et al. (2007). at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 303 Response T\fme to No\t\bel E\bents Time to overcome roadblocks was analyzed with a 3 (training) × 3 (critical mission) mixed ANOVA. One observation was missing \from the cross\btraining condition. There was a signi\ficant main e\f\fect o\f mission, F(1.53, 33.59) = 119.82, \f < .001, MSE = 16018.63, η 2 = .85 (Greenhouse\b G eisser correction used). The signi\ficant mission e\f\fect was attributed to di\f\ferences in the di\f\fi\b culty o\f roadblocks. There\fore, no \further analy\b ses were per\formed to isolate that e\f\fect. The main e\f\fect o\f training was also signi\ficant, F(2, 22) = 3.53, \f = .047, MSE = 19460.72, η 2 = .24 ( Figure 3). The planned comparisons at the critical mis\b sions revealed that procedural\btrained \l teams were signi\ficantly slower to overcome road\b blocks (M = 218.70, SD = 94.34) than were teams in the other two conditions (M = 146.67, SD = 103.52) at Mission 5, F(1, 23) = 3.1 1, \f = .091, MSE = 10005.89, η 2 = .12. Teams with procedural training were also signi\ficantly slower (M = 656.00, SD = 191.01) than those in the other two conditions (M = 533.07, SD = 98.35) at Mission 6, F(1, 23) = 4.50, \f = .045, MSE = 20164.71, η 2 = .16. The same compari\b s on at Mission 9 was not signi\ficant. The Training × Critical Mission interaction was not signi\ficant. Analysis o\f the measure \for time to overcome roadblock in the noncritical missions (i.e., Missions 7 and 8) did not reveal any sig\b ni\ficant di\f\ferences. Interpos\ft\fonal Know\tledge Interpositional teamwork and taskwork knowledge results are summarized in Table 4.

Interpositional teamwork and taskwork knowl\b edge were separately analyzed with 3 (training) × 2 (session) mixed ANOVAs. The taskwork ANOVA did not reveal any signi\ficant di\f\fer\b ences. There was a signi\ficant interaction e\f\fect \for interpositional teamwork knowledge, F (2, 23) = 2.70, \f = .089, MSE = .01, η 2 = .19 ( Figure 4). Pooled comparisons revealed that cross\btraining ( M = .87, SD = .07) led to signi\fi\b cantly higher interpositional teamwork knowl\b edge compared with the other two conditions ( M = .78, SD = .08) at Session 2, F(1, 24) = 7.04, \f = .014, MSE = .01, η 2 = .23. The same compari\b s on \for Session 1 was not signi\ficant. The main e\f\fect o\f session \for interpositional teamwork knowledge was also signi\ficant, F(1, 23) = 6.24, \f = .02, MSE = .01, η 2 = .21. Although teams in a ll training conditions exhibited some increase in interpositional teamwork knowledge across ses\b sions, teams in the cross\btraining condition exhibited a signi\ficantly greater increase. DISCUSSION Perturbation\btraine\ld teams signi\ficantly out\b per\formed teams in the other conditions in two out o\f three critical test missions, and their response times to overcome novel roadblock events were roughly equivalent to cross\btrained teams. These results lend support to our \first hypothesis that perturbation training leads to high per\formance un\lder novel condition\ls. The results suggest that something similar to the e\f\fects o\f practice condition variability (Schmidt & Bjork, 1992) contributed to trans\fer at the team level: Perturbation training allowed teams to generalize per\formance to novel con\b ditions by \forcing the teams to coordinate in new ways during task acquisition. However, whereas practice condition variability provides a range o\f task conditions \for the individual learner, perturbation training induced coordina\b tion variability across team members during repetitions o\f the same task. By training teams to \formulate and test new solutions to the prob\b lem o\f coordinating ground targets during task acquisition, perturbation training actively engaged 0 100 200 300 400 500 600 700 800 5 (First Roa\fbloc\b) Mission Time to Overcome R\loadbloc\f (seconds) Cross-Traine\fPerturbationProce\fural 9 (High Wor\bloa\f) 6 (Retention) Figure 3. Time to overcome roadblocks by training condition across the critical missions (error bars repre\b sent 90% con\fidence intervals). at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 304 April 2010 - Human Factors team processes that are needed to adapt to novel, but related, coordination problems in the posttraining environment. Because perturbation training builds on prior novelty, it may also have allowed teams to develop within a rich experiential learning environment (Kolb, 1984).Our second hypothesis was that cross\btraining would result in the highest levels o\f shared knowledge but that this would not necessarily result in the best per\formance under novel task conditions or the \fastest response times to novel events. Support \for that hypothesis was mixed.

Cross\btraining resulted in greater shared team\b work knowledge in the second session but not in the \first session. This is not a surprise given that the majority o\f condition\bspeci\fic \l training took place a\fter the \first knowledge measure\b ment session. However, shared taskwork knowledge did not change across experimental sessions, regardless o\f training condition. This may suggest a ceiling e\f\fect, such that task\b related concept relatedness (e.g., the association between airspeed and altitude) does not change a\fter initial parti\lcipant training.

Cross\btraining also resulted in highest mean per\formance at one o\f the critical missions (the retention test) and \faster response times \for overcoming roadblocks, although those di\f\fer\b ences were not signi\ficant. It is possible that with a larger sample size, or a less variable task environment, cross\btraining would have resulted in signi\ficant advantages. Marks et al. (2000) \found that development o\f a shared mental model pre\b dicted per\formance under novel task conditions better than under routine task conditions. The retention and roadblock tests are novel condi\b tions unique to our experiment, however, and \further empirical work is needed to better under\b stand the bene\fits o\f cross\btraining and shared knowledge under the\lse conditions. The results support our third hypothesis that procedural training should result in the least adaptive teams. Procedural training is arguably the most prevalent \form o\f training \for coordi\b nating highly critical team tasks, but its utility \for training adaptive teams has been increas\b ingly called into question (e.g., Ford & Schmidt, 2000; Grote, Kolbe, Zala\bMezö, Biene\feld\bSeall, & Künzle, 2010; Sta\lchowski et al., 200\l9). Procedural training need not be limited to a single, standardized coordination process; assum\b ing that the space o\f possible \future events is \finite, procedures can be scripted \for a variety o\f \foreseeable contingencies. Nevertheless, given the current results, we argue that teams trained to automatically \follow a standardized coordination procedure become rigid and slow to adapt to novel changes in highly dynamic task environments. Whereas proceduralization may be good \for unchanging and \foreseeable aspects o\f a task, in training adaptive teams, TABLE 4: Mean Interpositional \beamwork and \baskwork Knowledge by \braining Condition Measure Cross-\brained PerturbationPro\fedural \beamwork Se ssion 1 .73 (.12) .78 (.12).74 (.11) Se ssion 2 .87 (.07)* .79 (.06).77 (.09) \baskwork Se ssion 1 .47 (.04) .46 (.05).48 (.05) Se ssion 2 .47 (.01) .48 (.04).48 (.03) Note. Standard deviations in pa\lrentheses.

*p < .10. 0.5 0.6 0.7 0.8 0.91 Session 1Interpositional Te\gamwork Knowledge Cross-Tr\finedPert\brb\ftionProced\br\fl Session 2 Figure 4. Interpositional teamwork knowledge by training condition across knowledge measurement sessions (error bars represent 90% con\fidence intervals). at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 305 there should be a match between interaction variability and the changing dynamics o\f the task environment.We examined team per\formance under three critical situations: adaptation to a novel event (roadblock), a\fter a retention interval, and under high workload. This is not an exhaustive list o\f possibilities. There are many \forms o\f adaptation \for which a team could be trained (e.g., role struc\b ture adaptation; Lepine, 2005). The current results are intuitively plausible, however, given the nature o\f mechanisms o\f team adaptation currently \found in the adaptive team literature, and extend the idea o\f process\bbased adaptability training. The building and maintenance o\f shared mental models are thought to support team adaptation (Burke, Stagl, Salas, Pierce, & Kendall, 2006; Stout et al., 1999; Waller, Gupta, & Giambatista, 2004). Indeed, cross\btraining was success\ful in building shared teamwork knowledge, and cross\b trained teams exhibited potential \for high per\b \formance under novel conditions. Parallel to the motivation \for perturbation training, however, teams directly adapt via \flexible interaction pro\b cesses (Gorman et al., in press; Manser, Harrison, Gaba, & Howard, 2009; Stachowski et al., 2009; Waller, 1999). Kozlowski, Gully, Nason, and Smith (1999) suggested that teams adapt by selecting an appropriate \form o\f interaction \from a preexisting repertoire or by creating a new \form. Approaches like perturbation training have the potential to broaden a team’s interac\b tion repertoire not by prescribing preexisting \forms o\f coordination but by allowing teams to exercise bottom\bup organization o\f new coordi\b nation links. What are the practical implications \for train\b ing adaptive teams, and how can we apply perturbation training? Simulation\bbased team training (Dorsey et al., 2009) would allow \for the design o\f perturbations that \focus on speci\fic events, times, or interactions (Gorman, Cooke, & Duran, 2009). Simulation\bbased training can emphasize physical (equipment) \fidelity or cog\b nitive \fidelity (how well the simulation exercises psychological processes required \for that task; Goettle, Ashworth, & Chaiken, 2007). For per\b turbation training, cognitive \fidelity should be emphasized in order to exercise the team interaction processes needed \for the real\bworld task (Bowers & Jentsch, 2001). Another con\b cern is the speci\fics o\f introducing perturbations:

when, how many, what kind, and how o\ften?

Simulation\bbased training would be the ideal venue \for perturbation training, and although approaches such as crew resource management (see Salas, Wilson, Burke, & Wightman, 2006) may use simulators to train \for rare or novel events, our results suggest that more thought and research should go into identi\fying the types o\f team interaction experiences needed and the ideal timing o\f those experiences.

What are the implications o\f the varying the\b oretical training motives—shared knowledge, proceduralization, \flexible interactions—\for team cognition? Prevalent in the team cognition literature is a distinction between knowledge and process and which contributes most to team e\f\fectiveness (e.g., Cooke, Gorman, & Winner, 2007). We submit that cross\btraining most directly impacts knowledge, that perturbation training most directly impacts process, and that procedural training may have little impact on either. The current study is not an unequivocal test o\f knowledge versus process accounts o\f team cognition, nor is it an exhaustive sam\b pling o\f variations on procedural, perturbation, or cross\btraining in a variety o\f contexts.

Nonetheless, the results do suggest that training \focused on process may contribute something to team e\f\fectiveness that a knowledge\b\focused approach does not. CONCLUSION The details o\f team adaptation are not speci\b \fied at the outset o\f a novel event. The details accrue gradually, during the process o\f adapta\b tion, and there lies the problem \for training adaptive teams: They must be able to decide, plan, think, and act under conditions never expe\b rienced. Adaptation is the altering o\f structure in accordance with environmental change and, under many circumstances, is not a purely top\b down, knowledge\bdriven process. Teams should be provided opportunities to exercise adaptive competency using not only top\bdown (knowledge\b \focused) training but also bottom\bup (process\b oriented) training. Perturbing coordination as team members interact is one means o\f eliciting the bottom\bup, process\boriented \flexibility that at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from 306 April 2010 - Human Factors teams need in order to adapt. Future research should continue to explore mechanisms o\f \flex\b ible team interaction and how teams use them to adapt to the pressures o\f highly dynamic, high\bstakes work environments. ACKNOWLEDGMENT S This research was \funded by Air Force O\f\fice o\f Scienti\fic Research Grant FA9550\b04\b1\b0234 and Air Force Research Laboratory Grant FA8650\b04\b6442; additional support came \from National Science Foundation Grant BCS 0447039.

The authors would like to thank the \following individuals who contributed to this research:

Dee Andrews, Christy Caballero, Olena Connor, Jasmine Duran, Preston Kiekel, Harry Pedersen, Steven Shope, Amanda Taylor, and Jenni\fer Winner. Some team per\formance results were previously reported in a technical report (Cooke, Gorman, Pederson, et al., 2007) and a con\fer\b ence proceeding pa\lper (Gorman et al.,\l 2007). REFERENCES Bjork, R. A. (1994). Memory and metamemory considerations in the training o\f human beings. In J. Metcal\fe & A. Shimamura (Eds.), Metacognition: Knowing about knowing (pp. 185–205).

Cambridge, MA: MIT Press.

Blickensder\fer, E., Cannon\bBowers, J. A., & Salas, E. (1998). Cross\btraining and team per\formance. In J. A. Cannon\bBowers & E. Salas (Eds.), Making decisions under stress: Im\flications for individual and team training (pp. 299–311). Washington, DC:

American Psychologic\lal Association.

Bowers, C. A. & Jentsch, F. (2001). Use o\f commercial, o\f\f\bthe\b shel\f, simulations \for team research. In E. Salas (Ed.), Advances in human \ferformance and cognitive engineering research (Vol. 1., pp. 293–317). Amsterdam, Netherlands: Elsevier Science.

Burke, C. S., Stagl, K. C., Salas, E., Pierce, L., & Kendall, D. (2006). Understanding team adaptation: A conceptual analysis and model. Journal of A\f\flied Psychology, 9\b, 1189–1207.

Cannon\bBowers, J. A., Salas, E., Blickensder\fer, E., & Bowers, C. A. (1998). The impact o\f cross\btraining and workload on team \functioning: A replication and extension o\f initial \findings.

Human Factors, 40, 92–101.

Cannon\bBowers, J. A., Salas, E., & Converse, S. (1993). Shared mental models in expert team decision making. In N. J. Castellan (Ed.), Individual and grou\f decision making (pp. 221–246).

Hillsdale, NJ: Erlba\lum.

Cooke, N. J., Gorman, J. C., Duran, J. L., & Taylor, A. R. (2007). Team cognition in experienced command\band\bcontrol \l teams.

Journal of Ex\ferimen\btal Psychology: A\f\flied, \b3, 146–157.

Cooke, N. J., Gorman, J. C., Pedersen, H. K., Winner, J., Duran, J., Taylor, A., Amazeen P. G., Andrews, D., & Rowe, L. (2007).

Acquisition and retention of team coordination in command- and-control (Technical Report \for AFOSR Grant FA9550\b04\b 1\b0234 and AFRL Award No. FA8650\b04\b6442). Washington, DC: Department o\f D\le\fense. Cooke, N. J., Gorman, J. C., & Rowe, L. J. (2009). An ecological perspective on team cognition. In E. Salas, J. Goodwin, & C. S. Burke (Eds.), Team effectiveness in com\flex organizations:

Cross-disci\flinary \fers\fectives and a\f\froaches (pp. 157–182).

New York, NY: Taylor and Francis.

Cooke, N. J., Gorman, J. C., & Winner, J. L. (2007). Team cogni\b tion. In F. Durso, R. Nickerson, S. Dumais, S. Lewandowsky, & T. Per\fect (Eds.), Handbook of a\f\flied cognition (2nd ed., pp. 239–268). New York, NY: Wiley.

Cooke, N. J., Kiekel, P. A., Salas, E., Stout, R. J., Bowers, C., & Cannon\bBowers, J. (2003). Measuring team knowledge: A window to the cognitive underpinnings o\f team per\for\b mance. Grou\f Dynamics: Theory, Research and Practice, 7, 179–199.

Cooke, N. J., & Shope, S. M. (2005). Synthetic task environments \for teams: CERTT’s UAV\bSTE. In N. Stanton, A. Hedge, K.

Brookhuis, E. Salas & H. Hendrick (Eds.), Handbook of Human Factors and Ergonomics Methods (pp. 46\b41\b46\b46).

Boca Raton, FL: CRC\l Press.

Dorsey, D., Russell, S., Keil, C., Campbell, G., Van Buskirk, W., & Schuck, P. (2009). Measuring teams in action: Automated per\b \formance measurement and \feedback in simulation\bbased training. In E. Salas, G. F. Goodwin, & C. S. Burke (Eds.), Team effectiveness in com\flex organizations: Cross-disci\flinary \fers\fectives and a\f\froaches (pp. 351–381). New York, NY:

Taylor and Francis.

Entin, E. E. & Ser\faty, D. (1999). Adaptive team coordination. Human Factors, 4\b, 312\b325.

Fiore, S. M., Salas, E., & Cannon\bBowers, J. A. (2001). Group dynamics and shared mental model development. In M. London (Ed.), How \feo\fle evaluate others in organizations (pp. 309– 336). Mahwah, NJ: Erlbaum\l.

Ford, J. K., & Schmidt, A. M. (2000). Emergency response training: Strategies \for enhancing real\bworld per\formance. Journal of Hazardous Materials, 75, 195–215.

Gaddy, C. D., & Wachtel, J. A. (1992). Team skills training in nuclear power plant operations. In R. W. Swezey & E. Salas (Eds.), Teams: Their training and \ferformance (pp. 379–396).

Norwood, NJ: Ablex.

Gersick, C. J. G., & Hackman, J. R. (1990). Habitual routines in task\bper\forming groups. Organizational Behavior and Human Decision Processes, 47, 65–97.

Goettle, B. P., Ashworth, A. R. S., III, & Chaiken, S. R. (2007). Advanced distributed learning \for team training in command and control applications. In S. M. Fiore & E. Salas (Eds.), Toward a science of distributed learning (pp. 93–117).

Washington, DC: American Psychologic\lal Association.

Gorman, J. C., Amazeen, P. G., & Cooke, N. J. (in press). Team coordination dynamics. In Nonlinear dynamics, \fsychology and life sciences.

Gorman, J. C., Cooke, N. J., Amazeen, P. G., Winner, J. L., Duran, J. L., Pedersen, H. K., & Taylor, A. R. (2007).

Knowledge training versus process training: The e\f\fects o\f training protocol on team coordination and per\formance. In Proceedings of the Human Factors and Ergonomics Society 50th Annual Meeting (pp. 382–387). Santa Monica, CA:

Human Factors and \lErgonomics Society.

Gorman, J. C., Cooke, N. J., & Duran, J. L. (2009). Development o\f simulated team environments \for measuring team cognition and per\formance. In D. Schmorrow, J. Cohen, & D. Nicholson (Eds.), PSI handbook of virtual environments for training and education (pp. 347–361). Westport, CT: Praeger Security International. at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from Training a dapT ive Teams 307 Gorman, J. C., Cooke, N. J., Pedersen, H. K., Winner, J. L., Andrews, D., & Amazeen, P. G. (2006). Changes in team com\b position a\fter a break: Building adaptive command\band\bcontrol \l teams. In Proceedings of the Human Factors and Ergonomics Society 49th Annual Meeting (pp. 487–491). Santa Monica, CA: Human Factors an\ld Ergonomics Society.

Gorman, J. C., Cooke, N. J., & Winner, J. L. (2006). Measuring team situation awareness in decentralized command and con\b trol environments. \lErgonomics, 49, 1312–1325.

Grote, G., Kolbe, M., Zala\bMezö, E., Biene\feld\bSeall, N., & Künzle, B. (2010). Adaptive coordination and heed\fulness make better cockpi\lt crews. Ergonomics, 53, 211–228.

Hockey, G. R. J., Sauer, J., & Wastell, D. G. (2007). Adaptability o\f training in simulated process control: Knowledge\b versus rule\b based guidance under task changes and environmental stress.

Human Factors, 49, 158–174.

Kolb, D. A. (1984). Ex\feriential learning: Ex\ferience as the source of learning and develo\fment. Englewood Cli\f\fs, NJ:

Prentice Hall.

Kozlowski, S. W. J., Gully, S. M., Nason, E. R., & Smith, E. M. (1999). Developing adaptive teams: A theory o\f compilation and per\formance across levels and time. In D. R. Ilgen & E. D. Pulakos (Eds.), The changing nature of work and \ferfor- mance: Im\flications for staffing, \fersonnel actions, and devel- o\fment (pp. 240–292). San\l Francisco, CA: Joss\ley\bBass.

Lepine, J. A. (2005). Adaptation o\f teams in response to un\foreseen change: E\f\fects o\f goal di\f\ficulty and team composition in terms o\f cognitive ability and goal orientation. Journal of A\f\flied Psychology, 90, 1153–1167.

Luchins, A. S. (1942). Mechanization in problem solving. Psychological Monog\bra\fhs, 54(248).

Manser, T., Harrison, T. K., Gaba, D. M., & Howard, S. K. (2009). Coordination patterns related to high clinical per\formance in a simulated anesthetic crisis. International Anesthesia Research Society, \b08, 1606–1615.

Marks, M. A., Sabella, M. J., Burke, C. S., & Zaccaro, S. J. (2002). The impact o\f cross\btraining on team e\f\fectiveness. Journal of A\f\flied Psychology, 87, 3–13.

Marks, M. A., Zaccaro, S. J., & Mathieu, J. E. (2000). Per\formance implications o\f leader brie\fings and team\binteraction training \for team adaptation to novel environments. Journal of A\f\flied Psychology, 85, 971–986.

Sauer, J., Burkolter, D., Kluge, A., Ritzmann, S., & Schuler, K. (2008). The e\f\fects o\f heuristic rule training on operator per\formance in a simulated process control environment.

Ergonomics, 5\b, 953–967.

Salas, E., DiazGranados, D., Klein, C., Burke, C. S., Stagl, K. C., Goodwin, G. F., & Halpin, S. M. (2008). Does team training improve team per\formance? A meta\banalysis. Human Factors, 50, 903–933.

Salas, E., Nichols, D., & Driskell, J. E. (2007). Testing three team training strategies in intact teams: A meta\banalysis. Small Grou\f Research, 38, 471–488.

Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C. (2006). Does crew resource management training work? An update, an extension, and some critical needs. Human Factors, 48, 392–412.

Schmidt, R. A. (1975). A schema theory o\f discrete motor skill learning. Psychological Revie\bw, 82, 225–260. Schmidt, R. A., & Bjork, R. A. (1992). New conceptualizations \lo\f practice: Common principles in three paradigms suggest new concepts \for traini\lng. Psychological Scien\bce, 3, 207–217.

Schvaneveldt, R. W. (1990). Pathfinder associative networks:

Studies in knowled\bge organization. Norwood, NJ: Ablex.

Snook, S. A. (2002). Friendly fire: The accidental shootdown of U.S. black hawks over northern Iraq. Princeton, NJ: Princeton University Press.

Stachowski, A. A., Kaplan, S. A., & Waller, M. J. (2009). The ben\b e\fits o\f \flexible team interaction during crises. Journal of A\f\flied Psychology, 94, 1536–1543.

Stout, R. J., Cannon\bBowers, J. A., Salas, E., & Milanovich, D. M. (1999). Planning, shared mental models, and coordinated per\b \formance: An empirical link is established. Human Factors, 4\b, 61–71.

Turvey, M. T. (1990). Coordination. American Psychologist, 45, 938–953.

Vaughan, D. (1996). The Challenger launch decision: Risky tech- nology, culture, and deviance at NASA. Chicago, IL: University o\f Chicago Press.

Volpe, C. E., Cannon\bBowers, J. A., Salas, E., & Spector, P. E. (1996). The impact o\f cross\btraining on team \functioning: An empirical investiga\ltion. Human Factors, 38, 87\b100.

Waller, M. J. (1999). The timing o\f adaptive group responses to nonroutine events. Academy of Management Journal, 42, 127–137.

Waller, M. J., Gupta, N., & Giambatista, R. C. (2004). E\f\fects o\f adaptive behaviors and shared mental models on control crew per\formance. Management Science\b, 50, 1534–1544. Jamie C. Gorman received his PhD in cognitive psychology \from New Mexico State University, Las Cruces, in 2006 and is a postdoctoral research associate at Arizona State University–Polytech\lnic and the Cognitive Engineering Research Institute in Mesa, Arizona. Beginning August 2010, he will be an assistant pro\fessor in the Human Factors Program at Texas Tech University.

Nancy J. Cooke received her PhD in cognitive psychology \from New Mexico State University, Las Cruces, in 1987 and is a pro\fessor in applied psychology at Arizona State University–Polytech\lnic and science director o\f the Cognitive Engineering Research Institute\l in Mesa, Arizona.

Polemnia G. Amazeen received her PhD in experi\b mental psychology \from the University o\f Connecticut, Storrs, in 1996 and is an associate pro\fessor in psy\b chology at Arizona State University, Tempe, and a \fac\b ulty research associate at the Cognitive Engineering Research Institute in Mesa, Arizona.

Date received: May \b3, 20\b09 Date acce\fted: March 29, 20\b0 at LIBERTY UNIV LIBRARY on August 17, 2012 hfs.sagepub.com Downloaded from