test: uploading a file via git

20-F-1436-1-141

@@ -0,0 +1,141 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

36733527-1-22

@@ -0,0 +1,1274 @@
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      2Monterey, Ca.tact identification, intent, available responses andtheir consequences. For example, the close prox-
+imity of U.S. Navy forces and potential adversary
+forces makes interpreting the actions of an in-
+bound aircraft who does not respond to radio
+warnings much more difficult. Should the air-
+craft's behavior be interpreted as an attack profile,
+or does the pilot merely intend to harass, or does
+the aircraft in question not carry the equipment
+necessary to receive verbal warnings, leaving the
+pilot unable to  receive radio warnings directed
+toward him and unaware of his precarious posi-
+tion? In extreme cases there is no clear cut right or
+wrong answer about a decision. Rapidly unfolding
+events result in severe time pressure and severe
+(often catastro-phic) consequences for errors.
+While current real-time battle management sys-
+tems are well-suited to the demands of all-out
+conflicts, they may not be optimized for littoral
+situations where human intervention in decision-
+making is even more important (Office of Naval
+Technology [ONT], 1992). (Since 70 percent of
+the world's population lives within 200 miles of
+the sea, most future contingencies are likely to
+involve littoral warfare (Mundy, 1994).)Two unfortunate and highly publicized eventsfocused attention on the difficult types of deci-
+sions confronting naval commanders and provided
+the impetus for this research. In the case of the
+U.S.S. Stark, the commander made the decision to
+not engage an inbound aircraft which was be-
+lieved to not be a threat to his ship, and 27 U.S.
+naval personnel lost their lives as a result. In the
+case of the U.S.S. Vincennes, the commander
+made the decision to engage the inbound aircraft
+believing it was a threat to his ship—which turned
+out to be a commercial airliner—and all personnel
+aboard the airliner were killed as a result. In rec-
+ognition of the complex and difficult decisions
+required in these types of situations the Tactical
+Decision Making Under Stress (TADMUS) pro-
+gram was initiated to conduct research in the areas
+of human factors and training technology. The
+objective is to develop and apply principles that
+can help avoid these types of situations in the fu-
+ture. This paper, and a two companion papers
+(Hutchins, Kelly, & Morrison, 1996; Kelly,
+Hutchins, & Morrison, 1996), report on a multi-year, multi-experiment research effort conductedunder the TADMUS program to apply recent de-
+velopments in decision theory and human-system
+interaction technology to the design of a decision
+support system (DSS) for enhancing tactical deci-
+sion making under highly complex conditions.1.1 "Naturalistic" and Classical Decision-Making Para-digmsIn the same time frame that these tragic acci-dents occurred, a radical shift was occurring in the
+way psychologists viewed human decision mak-
+ing. Research was now focused on experienced
+decision makers performing their normal tasks in
+natural settings. "Naturalistic" decision-making
+research studies the decision strategies people ac-
+tually use in bringing their expertise to bear under
+challenging real-world conditions. Decision
+making milieus encompassed under the naturalis-
+tic paradigm include hospital emergency rooms,
+aircrew flight coordination, military command and
+control settings, process control systems, and po-
+lice and fire units.The naturalistic perspective, also known as"everyday cognition," is based on a belief that
+cognitive functions "elicited in natural settings
+(are) likely to differ, either quantitatively, or
+qualitatively, from those that occur in artificial or
+contrived situations, and results from sterile and
+contrived situations may not generalize to less
+constrained and more natural environments"
+(Salthouse, 1992, p. 982). There is a growing
+body of work that demonstrates that experienced,
+real-world decision makers rarely use traditional
+resource intensive strategies to make decisions in
+the face of dynamic, adverse conditions and time-
+pressure (Kaempf & Militelo, 1992; Klein, 1989;
+1993). Instead, experts rely on their abilities to
+recognize and appropriately classify situations:
+these abilities are based on having much experi-
+ence in the task domain. Once they know what
+they are facing they also tend to know what re-
+sponse option to apply, based on retrieval from
+memory of typical responses and outcomes that
+worked well in past similar situations. They use
+the limited time available to evaluate the feasibil-
+ity of that option before implementing it. Experi-
+enced decision makers recognize the situation or
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      3Monterey, Ca.scenario based on a comparison of the features ofthe current situation with stored memory represen-
+tations, or schemata. Schemata are highly inter-
+con-nected clusters of knowledge concerning cer-
+tain situations, or particular problem types, and
+associated actions or solution procedures (Fede-
+rico, 1995). Once the situation is recognized, so-
+lutions are stimulated by activation of these mem-
+ory representations.In contrast to the naturalistic perspective, ear-lier analytical methods applied in decision support
+systems were primarily used for option generation
+and evaluation, rather than for situation assess-
+ment. Traditional decision theorists argue that op-
+timum decision making involves thorough analy-
+sis of all the available data and the evaluation of
+all possible hypotheses; these approaches tend to
+rely on extensive calculations designed to arrive at
+optimal solutions. Processes for making decisions
+where the person weighs the pros and cons of
+various options and selects the one option that
+provides the most benefit are described in the de-
+cision making literature. The extensive time re-
+quirements and complicated mathematical calcu-
+lations involved (e.g., Multi-Attribute Utility
+Analysis), however, make these approaches unre-
+alistic for situations requiring rapid decision
+making. Generally, these analytical strategies in-
+volve the following steps (Kaempf & Militello,
+1992):  specify all relevant features of the task;
+  identify the full range of options;
+  identify the key evaluation dimensions;
+  identify weights for each dimension;
+  rate each option on each dimension;
+  tabulate the results, and
+  select the best option.These analytical strategies may be appropriate for
+inexperienced subjects making decisions about
+novel tasks, but not for experienced personnel
+making real-time decisions. In natural settings,
+time constraints and the difficulty in assigning
+weights and rating the dimensions involved render
+classical analysis techniques untenable. Typically
+in realistic settings experts employ recognition-
+based reasoning, not classical analytical ap-
+proaches. Experienced decision makers use their
+extensive knowledge to seek information, identifyand interpret the problem, understand the signifi-cance, derive the intention (where possible),
+model the situation (as time allows), select the
+action, evaluate the choice, and anticipate the con-
+sequence. "This decision cycle is distinctly differ-
+ent from classical models, which are based on the
+assumption that all options, outcomes, and prefer-
+ences are known and calculated in advance"
+(Federico, 1995, p. 106).Traditionally, research in decision making hasbeen directed largely toward situations in which
+(1) decision makers have sufficient time to gener-
+ate options, conduct option assessment, and select
+a course of action; (2) the consequences of an in-
+correct response are not immediately severe; (3)
+decisions are reached via consensus of a group;
+and (4) workload is manageable. Little research
+has been conducted into the development of tacti-
+cal decision support systems for use in naturalistic
+situations characterized by time pressure, high
+risk, uncertainty and information ambiguity, high
+workload, team coordination demands and task
+complexity (ONT, 1992).The central hypothesis for the research re-ported here is that presenting decision makers
+with decision support tools which were designed
+to parallel the cognitive strategies employed by
+experts, as observed in naturalistic settings, will
+reduce the number of decision making errors. This
+is accomplished by developing the architecture
+and algorithms to process information the same
+way research indicates humans do under similar
+circumstances.2   Tactical Decision Making TasksThe global tactical decision-making task involvesidentification of and responding to numerous
+contacts. When an aircraft (or a surface contact) is
+detected the CIC personnel work as a team to de-
+termine the identity and to try to determine
+whether or not the aircraft poses a threat. The high
+degree of inherent ambiguity associated with
+contact information can often make threat assess-
+ment a very difficult task. This is because many
+pieces of data fit multiple hypotheses regarding
+threat assessment. The global response choices
+(that is, engage, monitor, do nothing) are largely
+determined by the ship's orders and the current
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      4Monterey, Ca.geopolitical situation. Specific actions (such as,change course, issue verbal warnings, illuminate
+with radar, challenge with other sensors, etc.) de-
+pend on the local conditions and the relative posi-
+tions of the inbound contact of interest and own-
+ship. Determining which of these actions is likely
+to be effective depends on maintaining an accurate
+threat assessment which requires "continual up-
+dating in accordance with recurrent situation as-
+sessments" (Sarter & Woods, 1991, p. 52).This decision problem presents a highly chal-lenging cognitive task, that is, making inferences
+and deductions from incomplete and uncertain
+information derived from multiple sources and
+relating to several concurrent threats (or potential
+threats) under time-compressed conditions. The
+cognitive functions performed by the tactical deci-
+sion maker are both data and resource limited
+(Norman & Bobrow, 1975). Decisions are re-
+source limited by the mental resources of the deci-
+sion makers, who must maintain large amounts of
+information in memory under conditions of high
+workload and stress. The decisions are data lim-
+ited by the inability of the sensors to provide
+complete, error-free, unambiguous data to support
+the identification process. In particular, the ex-
+perimental scenarios were designed to follow the
+pattern of being set in an ambiguous situation
+where one or more threats of uncertain origin and
+uncertain intent approach either own ship or the
+ship being protected and may not respond to
+warnings. Scenarios were designed to be highly
+ambiguous, as this quality of uncertainty is in-
+dicative of the types of decisions to be made in
+current and future scenarios.2.1  Threat AssessmentIn the antiair warfare problem, threat assess-ment is particularly difficult because the available
+information is often incomplete or ambiguous.
+The ambiguity could be due to (a) the information
+transmission characteristics of the transmission
+medium, such as a radar transmission or a radio
+report that is only intercepted on an intermittent
+basis, (b) deliberate deceptive actions (such as ra-
+dar jamming) by the pilot flying the aircraft, or (c)
+the overlapping classification categories typical of
+many parameter measure-ments. For example, air-craft can typically fly at altitudes ranging between2,000 and 40,000 feet. Generally, an aircraft that
+is flying above 20,000 feet is not considered to be
+a threat. Conversely, an aircraft below 10,000 ft is
+considered to be more of a potential threat.  How-
+ever, aircraft flying in the middle range, (that is,
+below 20,000 ft. and above 10,000 ft.) can be
+much more difficult to categorize. Because many
+aircraft do fly in this middle range, other variables
+need to be considered in conjunction with altitude.
+This same situation of overlapping categorization
+categories exists for several other variables. These
+variables include radars that are found on both
+threat and non-threat platforms, country of origin,
+and measures of course and speed. In the case of
+speed, for example, when an aircraft flying at a
+low altitude decreases speed this could be viewed
+as indicative of a threat action (that is, slowing
+down in order to obtain better targeting informa-
+tion); however, at the same time, there could be
+other viable explanations for an aircraft's de-
+creasing speed.If the decision maker had access to all dataabout a contact approximately twelve variables
+would be used to determine identity and to infer
+intent. Two or three of these items, alone, do not
+provide definitive answers because, in many
+cases, these parameter values do not fall within
+clear-cut ranges for a particular assessment cate-
+gory (i.e., threat, non-threat). Thus, a single time
+slice of information provides an incomplete pic-
+ture of the situation. In the dynamic, ambiguous
+conditions characteristic of littoral operations, the
+rate and direction of change (data history) can
+help one better assess the threat and predict the
+future state of the situation. When the incoming
+information changes over time, the integration of
+information as it changes can help the user extract
+the message (Kirshenbaum, 1992). The DSS was
+designed to do precisely this: to facilitate the inte-
+gration process and present a synthesized picture
+of the situation to the user in a format that can be
+quickly assimilated. The variables, that are used to
+develop a threat assessment, can be divided into
+two classes: sensor information (raw or computer-
+processed information) and the contact's response,
+or lack of response, to actions taken by the team.
+These other actions, and the integration of the in-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      5Monterey, Ca.formation received via the contact's response orlack of response to them, are necessary to clarify
+the tactical picture.2.2  Situation AwarenessWhile recent increasing interest among re-searchers regarding the concept of situation
+awareness (SA) has generated a debate on the pre-
+cise definition of this term most researchers ac-
+knowledge the importance of the concept. In gen-
+eral, SA refers to the decision maker's moment-
+by-moment ability to monitor and understand the
+state of the complex system and its environment
+(Adams, Tenney, & Pew, 1995). These authors
+state the essential idea which is that when emer-
+gencies arise, the completeness and accuracy of
+the decision maker's SA are critical to the ability
+to make decisions, revise plans, and manage the
+system. Specific decision-making tasks included
+under SA include the ability to:  (1) maintain an
+accurate perception of the surrounding environ-
+ment (both internal and external to the ship); (2)
+identify problems and/or potential problems; (3)
+recognize a need for action; (4) note deviations in
+the mission; and (5) maintain awareness of tasks
+performed (Shrestha, Prince, Baker and Salas,
+1995). To maintain an accurate SA the decision
+maker should take into account both information
+that is available and that which can be activated
+from memory (Sarter and Woods, 1991).However, a difficulty arises as a result of theheavy workload imposed by this process. When
+the decision maker is faced with several concur-
+rent contacts of interest, all of which have numer-
+ous associated data items (i.e., as many as  a
+dozen), some or all of which may change over the
+course of the scenario (e.g., intelligence, active
+radar emitters, various kinematic parameters, etc.)
+memory load easily exceeds human capacity.
+Moreover, changing parameters may impart dif-
+ferent interpretations to what is occurring. In some
+cases the moment-to-moment attentional demands
+of a tactical situation are relentless and unforgiv-
+ing (such as, a terrorist aircraft directly inbound
+toward "own-ship" which can result in "task fixa-
+tion"), sometimes relevant background knowledge
+is unavoidably incomplete (such as, an unfamiliar
+aircraft), and sometimes the decision maker is al-ready thinking and working as hard as possible,even when there are no unanticipated events when
+there is a high contact density (Adams, et al,
+1995). These instances provide a few illustrations
+of situations that can degrade situation awareness.Complex information gathering and process-ing systems have been designed to aid the deci-
+sion-maker in the past. However, these systems
+often increase the decision-maker’s burden due to
+the inherent system complexity and the failure to
+design them in a way that they will fit the user's
+cognitive processing limitations. Often, these
+systems require operators to perform difficult
+cognitive tasks under heavy workloads. They must
+perceive, synthesize and determine the relevance
+of a continual stream of incoming information,
+often pertaining to several concurrent contacts,
+while projecting future anticipated events and
+making decisions regarding actions to be taken.
+Decision makers must assess, compare, and re-
+solve conflicting information, while making diffi-
+cult judgments, and remembering the status of
+critical contacts along with the contact's response
+to actions taken by the CIC team. These decision-
+making tasks are interleaved with other required
+tasks, such as keeping other team members in-
+formed (both on and off the ship). Furthermore,
+these complex tasks are performed under condi-
+tions where adverse environmental (noise, vibra-
+tion, temperature extremes, etc.) and internal
+stressors (boredom, fatigue, anxiety, and fear) are
+part of the environment.3  Decision Support PrinciplesA case has been made that previous generations ofdecision support systems, which focus primarily
+on solution optimization and base decision sup-
+port on normative models of human decision
+making, are less applicable than a DSS that par-
+allels the cognitive strategies used by domain ex-
+perts in situations characterized by time-
+constrained situations with uncertain and ambigu-
+ous data (Smith & Grossman, 1993).  These
+authors point out that rarely, if ever, were earlier
+tactical decision aids intended as psychological
+models of human cognitive behavior.  Instead,
+these aids performed "complex and burdensome
+calculations, reducing the work-load on personnel,
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      6Monterey, Ca.speeding up the dissemination of information, andproviding more time for command decision mak-
+ing" (Tolcott, 1991, p. 44).3.1  Feature Matching and Story GenerationWe have applied two of these new models ofhuman decision making—which parallel the cog-
+nitive strategies used by domain experts—to the
+design of a DSS for enhancing antiair warfare
+tactical decision making. These two models that
+people use in assessing a situation are feature
+matching and story generation. The feature
+matching model, described by Noble (1989), in-
+volves an organization of memory, or "schemas,"
+and information-processing where decision mak-
+ers use their previous experiences to assess a
+situation and identify promising actions. Incoming
+information is categorized, selected, edited, and
+organized on the basis of a person's general
+knowledge about a domain. Both story generation
+and feature matching occur under conditions
+where a large base of implication-rich, condition-
+ally dependent pieces of evidence must be evalu-
+ated before choosing an alternative from a set of
+prospective courses of action. The feature match-
+ing model applies a spatio-temporal dependence,
+whereas story generation is an example of causal
+dependence. According to the explanation-based
+model, decision makers construct a causal model
+to explain the available evidence (Pennington &
+Hastie, 1993). At the same time, the decision
+maker creates a set of alternatives from which an
+action will be chosen.  A decision is made when a
+story is successfully matched to an alternative in
+the choice set. Story generation occurs in complex
+situations where the decision maker may not have
+all the necessary information or when a series of
+facts may appear to contradict each other. The de-
+cision maker must then develop causal links be-
+tween these facts to produce a coherent picture of
+the situation (Klein, 1989; 1993).The explanation-based reasoning model isbased on research which found that jurors develop
+a narrative story to organize trial information
+where causal and intentional relations between
+events are central (Pennington & Hastie, 1992).
+Pennington & Hastie propose four certainty prin-
+ciples—coverage, coherence, uniqueness, andgoodness-of-fit—that govern (i) which story willbe accepted, (ii) which decision will be selected,and the (iii) confidence or degree of certainty withwhich a particular decision will be made. This or-
+ganization of the evidence by the decision maker
+is believed to facilitate evidence comprehension.
+A central component of this model is that the
+story the juror constructs determines the juror's
+decision.  This explanation-based decision process
+is employed when the body of evidence relevant
+to a decision is large, complex, and the implica-
+tions of its components are interdependent.Feature matching, also referred to as the rec-ognition-primed decision (RPD) model, "occurs
+when the decision maker recognizes the features
+of the present situation as similar or identical to
+those of a previous situation" (Kaempf & Militelo,
+1992, p. 6). An adequate match triggers recall of
+information learned about this type of situation:
+(a) plausible goals, (b) critical cues to be moni-
+tored, (c) expectations of what should happen, and
+(d) a course of action that worked in similar situa-
+tions. According to this recent approach, expert
+decision makers may rely on well-developed
+memory representations to guide decision making
+in new (but similar) situations. The RPD model of
+decision making fuses two processes—situation
+assessment and mental simulation (Klein, 1993).
+In the simplest case the situation is recognized as
+familiar or prototypical, using feature matching,
+and the obvious response is implemented. In a
+more complex case the decision maker performs a
+conscious evaluation of the response, using men-
+tal simulation to uncover problems prior to im-
+plementing the response. In the most complex
+case the evaluation reveals flaws requiring modi-
+fication, or the option is judged inadequate and
+rejected in favor of the next most typical reaction.3.2  Situation AssessmentIn general, the overall task of responding toantiair warfare scenarios consists of situation as-
+sessment ("what's going on") and course of action
+selection ("what to do about it"). Recent theories
+of decision making emphasize the importance of
+situation assessment for good decision making in
+naturalistic, event-driven situations. Moreover,
+they stress that decisions regarding actions to be
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      7Monterey, Ca.taken are a by-product of developing the situationawareness that precedes action selection. Klein
+(1989) has found that usually the situation itself
+either determines or constrains the response op-
+tions and that experienced decision makers make
+up to 90% of all decisions without considering
+alternatives. If the situation appears similar to one
+that the decision maker has previously experi-
+enced, the pattern will be recognized and the
+course of action is usually immediately obvious.
+On the other hand, if the situation does not seem
+familiar complex RPD will be involved where the
+decision maker adjusts the option after evaluating
+it.Additional evidence was found in the specifictask domain of interest to the TADMUS program
+which added support to these findings on the way
+real-world decision makers make decisions in the
+context of their normal jobs. Research was con-
+ducted to determine decision requirements for
+command-level decision makers in the combat
+information center (CIC) of an Aegis cruiser.
+Analysis of 14 incidents from actual problems re-
+vealed 183 decisions. Of these, 103 concerned
+situation assessments. Results obtained after ana-
+lysts coded these situation assessments indicated
+that decision makers arrived at approximately
+87% of their situation assessments through feature
+matching and the remaining 13% through story
+generation (Kaempf, Wolf, & Miller, 1993). The
+other eighty decisions that were identified, from
+analysis of the real-world incidents mentioned
+above, involved course of action selection. These
+course of action decisions served a variety of
+functions, although, relatively few were intended
+to end the incident. Twenty were intended as a
+final course of action decision; 14 were imple-
+mented to obtain more information, 22 to manage
+resources, and 24 to put themselves in a more fa-
+vorable tactical position. A recognition-based
+strategy was also used by decision makers to de-
+velop a course of action, accounting for 95% of
+the actions taken in the 14 incidents. The decision
+makers generated and compared multiple options
+in only 5% of the cases. In line with these find-
+ings, the TADMUS program has adopted the po-
+sition that decision aiding systems should assist in
+the decision making process, and focus on aidingthe situation assessment portion of the decision-making task.A DSS was developed to support decision-making processes which research has shown are
+used by decision makers in real-world settings
+(Hutchins, Kelly, & Morrison, 1996). Specifi-
+cally, the DSS parallels the strategies used by ex-
+perienced decision makers to perform situation
+assessment (Nobel, 1989; 1993). This approach to
+supporting the user's intuitive approach to dealing
+with dynamic decision-making situations should
+produce tools that are both more easily understood
+and used, and that more effectively "exploit the
+decision maker's knowledge and expertise that
+might facilitate adaptation to complex, novel
+situations" (Cohen, 1993, p. 265).4  Human-System Interaction PrinciplesThe vast majority of research on human-computerinteraction design has been devoted to character-
+istics of displays that impact human perception,
+such as symbol legibility or detectability, and on
+relatively simple cognitive functions such as
+memory tasks. Fewer efforts have been devoted to
+understanding the effects of the format and man-
+ner in which information is presented on more
+complex levels of human cognition such as deci-
+sion making. Consequently, principles that can be
+applied to the design of the interface between the
+user and a decision support system for the purpose
+of enhancing cognitive changing situations, are
+not available to any significant degree (ONT,
+1992).4.1  Graphic PresentationsSeveral advantages are offered by graphicpresentations over a text-based presentation for-
+mat (Larkin and Simon, 1987). Graphic presenta-
+tions should (1) reduce the amount of mental
+computation required to perform tasks; and (2)
+allow users to spend less time searching for
+needed information. Casner (1991) elaborated on
+these ideas and found that graphics allow users to
+substitute less demanding perceptual opera-tions
+for more complex logical operations. For exam-
+ple, determining a change in altitude (and the de-
+gree of change) is immediately apparent when the
+user glances at the track history module. (Note
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      8Monterey, Ca.that the words contact and track can be used inter-changeably. The reader is referred to Figure 1.)The objective for the track history module isto facilitate the contact identification process by
+providing information that is integrated in a way
+that supports a recognitional decision strategy.
+This module depicts a contact’s speed, altitude,
+course and range on a two-dimensional graphical
+display along with a geometric representation of
+both the contact’s weapon release envelope and
+own-ship’s weapons coverage. A large amount of
+parametric data is portrayed graphically for rapidassimilation by the user. The user can see, at aglance, a synthesized picture of the contact’s be-
+havior. Compare this rather simple perceptual op-
+eration with the more complex logical operation
+involved in current operational systems which re-
+quire the user to recall and subtract numerical val-
+ues for past and current altitudes.Graphics also allow users to omit steps thatare otherwise necessary when a task is performed
+without a graphic. An example of this advantage
+isFigure 1.  Decision Support System Display Modules.also illustrated in the track history module whichincludes templates indicating weapon's coverage
+for both the inbound contact and "own-ship." To
+determine whether the aircraft is within its
+weapon's launch range there is no need to recall
+the specific launch range values and then compare
+them with the aircraft's current range. Instead, the
+user can determine if the aircraft is within its
+launch range by a quick glance at the display.Graphics help users save time when searchingfor needed information when several related di-
+mensions of information are encoded in a single
+graphical object. This is accomplished by inte-
+grating the kinematic parameters of speed, course,
+altitude, bearing, and range for a contact. The user
+can see, at a glance, a synthesized picture of the
+contact's behavior.  Compare this process with
+reading, in a text-based format, the individual pa-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      9Monterey, Ca.rameters which need to be integrated by the userinto a coherent picture of the contact's behavior.5   Limited Cognitive Processing CapabilitiesSince there are limits to the cognitive processingcapability of humans, it is important for the sys-
+tem to provide the needed information in a format
+that best supports the user operating under dy-
+namic decision-making conditions. It may be the
+case that current systems are inadequate to support
+the cognitive processing demands required by
+certain littoral scenarios. For example, according
+to Gruner (1990, p. 41), the U.S.S. Vincennes of-
+ficers and system operators "could not make better
+decisions because they did not have time to con-
+firm or deny the information uncertainties pre-
+sented them." Gruner maintains that the rapid pace
+involved in these types of situations can exceed
+the capacity of the human to comprehend the
+rapid flow of information presented by complex
+systems. In the case of the Vincennes, the CIC
+team had three minutes and 40 seconds to make
+their decision. This includes the time required for
+the operators to perceive and interpret sensor data
+and for the commanding officer to make informed
+judgments from these data (Roberts & Dotterway,
+1995). The result of the human's limited cognitive
+processing capabilities is that the decision makers
+may fail to remember critical pieces of data,
+overlook stored information, draw hasty conclu-
+sions, and produce flawed answers.Evidence of the effects of limitations in mem-ory and shared attention capacity on human deci-
+sion making were found during baseline testing
+(Hutchins & Kowalski, 1994; Hutchins & Westra,
+1995) and during empirical evaluation of the DSS
+(Kelly, Hutchins, & Morrison, 1996; Kelly, Mor-
+rison, & Hutchins. 1996).Simon (1978, p. 273) states, "...the human in-formation processing system...operates almost en-
+tirely serially, one process at a time, rather than in
+parallel fashion. This seriality is reflected in the
+narrowness of its momentary focus of attention."
+However, the AAW problem forces the decision
+maker to operate in a parallel processing mode
+when several contacts demand attention at the
+same time. The requirement to monitor and
+maintain an accurate SA for these concurrentcontacts, over the course of the evolving situation,imposes an additional load of strategically man-
+aging the overall situation. Several researchers
+have argued that "managing the attentional and
+conceptual processes that permit cogent SA in-
+volves significant cognitive resources" (Adams, et
+al, 1995, p. 91; Endsley, 1988). The tasks of pri-
+oritizing contacts and the associated actions to be
+taken by the team, updating the status of critical
+contacts, responding to the other requisite tasks in
+the queue and, more generally, of strategically
+managing the workload of current multitask sys-
+tems under dynamically changing scenarios can
+place an unrealistic cognitive load on the decision
+maker.A major advantage offered by the experimen-tal DSS is that it should "buy time" for the user by
+(1) performing many of the cognitive processing
+tasks for the user and (2) by presenting informa-
+tion in graphic format. The DSS will synthesize
+much of the information used to develop situation
+awareness and present a coherent picture of the
+situation to the user. This integrated picture will
+be portrayed graphically—rather than in the cur-
+rent text-based format—which should further re-
+duce the amount of time required to assimilate
+this information. By performing several informa-
+tion processing steps for the decision maker the
+decision maker's limited cognitive resources can
+be used for the types of decisions which require
+human abilities (e.g., the decision on whether to
+engage).5.1  Working Memory RequirementsAn essential information processing step re-quired by this task—and one which levies a heavy
+load on working memory—involves integrating
+kinematic and sensor variables and maintaining an
+awareness of changes in these variables over time.
+Changes in a contact's behavior such as, decreas-
+ing altitude, increasing speed, changes in elec-
+tronic emissions, etc., can provide key indicators
+of possible hostile intent. With current systems,
+the decision maker receives numerous reports
+from CIC team members who provide various
+pieces of the overall tactical picture (such as,
+kinematic parameter values, active electronic
+emitter identifications, and behavioral responses
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      10Monterey, Ca.of the contact in response to queries by the team)regarding a particular contact. Some of this infor-
+mation is also displayed in a text-based format for
+the user when a contact is "hooked" (that is, se-
+lected for display) by the decision maker. How-
+ever, to recognize a change in certain variables,
+current systems require the user to retain parame-
+ter values in short-term memory in order to recog-
+nize a change in the parameter, such as altitude.When the decision maker is monitoring sev-eral concurrent contacts (such as, cycling through
+three or four contacts in a 1-minute period) human
+working memory capabilities may quickly be sur-
+passed. To detect a change in a critical parameter
+value, the decision maker must maintain the pa-
+rameter values for the contacts of interest in
+working memory as he or she cycles between sev-
+eral contacts. For example, the decision maker
+must be able to recall that contact 7022 was at
+14,000 ft. altitude one minute ago, and then sub-
+tract the current altitude value of 10,000 ft., which
+will then indicate the aircraft is in a rapid descent.
+The DSS was developed to aid the decision maker
+by performing several of these cognitive process-
+ing tasks, thus, reducing the cognitive load for the
+user. By presenting the synthesized picture of the
+contact's behavior over time, through the use of
+graphical displays,  critical changes should be
+immediately apparent to the user.A second memory-intensive task involvesmaintaining, in working memory, a current list of
+actions taken by team members, the contact's re-
+sponse to these actions taken by the CIC team,
+and pending actions. Research has established that
+"memory is limited and that list maintenance is
+effortful and fallible—more so if the list must be
+ordered and still more if the membership of the
+list must be dynamically reordered and modified
+during retention" (Bower, 1970, as cited in Ad-
+ams, et al, 1995, p. 91).  The DSS should reduce
+the cognitive effort required for distributing atten-
+tion among the many contacts to be attended to
+and actions that are required. Working memory
+requirements should be reduced by having the
+DSS act as an intelligent "assistant," reminding
+the user regarding what actions are to be taken and
+when the actions are to be taken.A third way the DSS will reduce memory andinformation processing requirements is by dis-
+playing templates depicting weapons' envelopes
+for both the inbound contact and "own-ship." This
+should facilitate critical comparisons and judg-
+ments regarding timing of actions. During a sce-
+nario decision makers have to either rely on mem-
+ory to recall the launch range for various weapons
+or query a team member for this information. Both
+of these methods waste limited resources. The
+high workload and high tempo characteristic of
+littoral scenarios produce a stress-ful decision-
+making environment. The phenom-enon that in-
+creasing stress leads to decreasing working mem-
+ory is well documented (e.g., Hockey, 1986). The
+latter method for obtaining the desired informa-
+tion wastes limited resources by increasing the
+communications load and requiring more time to
+wait for a team members' response to the queryUnder these high-tempo and high workloadconditions human memory and attentional re-
+sources can easily be surpassed. Several cogni-
+tively resource intensive information processing
+steps are eliminated for the human decision maker
+by having them performed by the DSS. We pre-
+dict that the decision support tools will reduce the
+cognitive workload imposed on the decision
+maker in the following three ways: (1) by reduc-
+ing the amount of information processing to be
+performed, (2) reducing working memory re-
+quirements, and (3) assisting the user in allocating
+limited attentional resources.5.2  Reducing Human ErrorThe study of human cognitive processes andrelated error mechanisms has gained rapidly in-
+creasing interest in the past decade. Rasmussen
+(1987) argues that the emphasis in attempting to
+understand human errors must shift from tasks to
+the human-task mismatch.  For example, Gruner
+(p. 39), in discussing the Vincennes incident,
+maintains that "the system was poorly suited for
+use by human beings during rapid military action."
+He ascribes this lack of suitability to a human-
+machine mismatch between the rate of data flow
+possible with modern computer systems that can
+process and display information at phenomenal
+data rates and the "comprehension capability of
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      11Monterey, Ca.users which has remained almost static for thou-sands of years." This causal approach to under-
+standing human error is based on the premise that
+errors are rarely random and can be traced to
+causes and contributing factors. Once these con-
+tributing factors are identified they can be miti-
+gated.The impact and vulnerability of systems andhuman interfaces, because of incompatibilities
+between the way people perceive, think, and act,
+are documented in the popular and technical lit-
+erature (Buck, 1989; Casey, 1993; Norman, 1988;
+Perrow, 1984; Wilson & Zarakas, 1978). Newly
+developed systems will succeed or fail based on
+our ability to minimize these incompat-ibilities
+between the characteristics of the things we create
+and the way we use them. There are many well-
+documented instances of critical systems or pa-
+rameter changes going unnoticed or unheeded be-
+cause the operating procedures, or the human ma-
+chine interface, provided no historical trace. For
+example, an unnoticed increase in altitude con-
+tributed to the shoot down of the Iranian airbus by
+a U.S. Navy ship—when the team mistakenly be-
+lieved the aircraft to be descending—because
+there was no historical trace to make the aircraft's
+actual increasing altitude apparent (Dotterway,1992).  Five personnel in the U.S.S. Vincennes's
+combat information center, all viewing separate
+displays, reported the aircraft as descending while
+the Aegis data tapes later revealed a flight pattern
+of ascent (Roberts & Dotterway, 1995). One of the
+official investiga-tions of this incident, the Fo-
+garty Report (1988, p. 45), states that "stress, task
+fixation, and an unconscious distortion of data
+may have played a major role in this incident." A
+panel of five psychologists from the American
+Psychological Association who testified before
+Congress concluded that there were "predictable
+failings of human judgment under intense stress
+compounded by complex technology [which]
+clearly contributed to the accidental shooting of
+Iranian airliner Flight 655" (APA, p. 4).It is generally accepted that between 60-80percent of the accidents and malfunctions in
+transportation, manufacturing, process control,
+weapon, and other systems are attributable to hu-
+man error (Senders & Moray, 1991; Van Cott,1993; Weiner, 1994). Reducing tactical decisionmaking errors is one goal of the TADMUS pro-
+gram.  The following section presents a brief re-
+view of an experiment conducted to develop a
+baseline on tactical decision making performance
+in response to fairly stressful scenarios. A com-
+panion paper (Hutchins, Kelly, & Morrison, 1996)
+describes the experimental DSS modules and the
+way they are hypothesized to enhance tactical de-
+cision making performance.6   TADMUS Baseline ExperimentEarly research involved data collection in the De-cision-Making Evaluation Facility for Tactical
+Teams (DEFTT) Laboratory using simulated ex-
+isting shipboard displays to establish a baseline on
+decision-making performance. The purpose of this
+effort was to document baseline decision-making
+performance for experienced naval officers. Dur-
+ing the baseline phase of testing, a detailed under-
+standing was developed of the cognitive processes
+underlying the various tasks involved in situation
+assessment—and where the bottlenecks occur.
+This understanding was then used to design the
+way the information is presented to the user in or-
+der to facilitate performance of the required tasks.6.1  SubjectsThis study focused on the command-level de-cision makers of an antiair warfare team on an
+Aegis cruiser—the commanding officer and the
+tactical action officer. Subjects in the study con-
+sisted of six commanding officer/tactical action
+officer teams drawn from twelve active duty Na-
+val personnel; some were from training com-
+mands while others were from operational com-
+mands aboard ship or assigned to group staffs.6.2  ProcedureData were collected in the  DEFTT Labora-tory, a six-station test-bed environment that
+simulates console positions in a Navy Aegis
+cruiser combat information center. (For a detailed
+description of the DEFTT Laboratory see Hutch-
+ins, 1996.) Four stations were filled by confeder-
+ates (active duty Navy personnel) who play antiair
+warfare support-team member roles. These roles
+included the antiair warfare coordinator, identifi-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      12Monterey, Ca.cation supervisor, tactical information coordina-tor, and electronic warfare supervisor. After ap-
+proximately 1 1/2 hours of orientation to the labo-
+ratory and training in the use of the computer con-
+soles the subjects engaged in four scenarios. The
+scenarios were each about 25 minutes in length
+and contained between 11 and 14 contacts of in-
+terest per scenario, in addition to numerous back-
+ground contacts.6.3  Treatment of DataTeam communications were recorded on amultichannel audio recorder; these included all
+intra-team exchanges, as well as all communica-
+tions with simulated off-ship personnel. Audio
+tapes were used to produce verbatim, time-
+stamped transcripts of all team communications.
+A modified version of the TapRoot® Incident In-
+vestigation System (Paradies, 1991; Paradies and
+Unger, 1991) was then applied to identify errors.
+The objective was to identify tactically significant
+errors committed during the scenario. Tactically
+significant errors were defined as those errors that
+may lead to loss of life or significant political em-
+barrassment. The following criteria were used for
+counting an error as tactically significant:  (1) loss
+of situation awareness, (2) failure to take defen-
+sive action when within the weapon's range of an
+approaching contact, or (3) a violation of rules of
+engagement (ROE). Video recordings were made
+of the commanding officer and tactical action of-
+ficer computer screens. Detailed analyses of all
+audio and video recordings were conducted. (For
+a more detailed coverage of the methodology and
+results see Hutchins and Westra, in preparation).6.4  ResultsThe complex, time-constrained, decision-making situations embodied in the experimental
+scenarios resulted in a large number of decision
+errors. The mean number of tactically significant
+errors documented across six teams and four sce-
+narios was 14; the number of errors ranged from
+nine to twenty-two. The standard deviation was
+3.7.  Subjects performed an average of 50% of the
+required behaviors as specified in the rules of en-
+gagement. The ordinal agreement between three
+raters (navy subject matter experts) on error countranks from TapRoot® analyses was computed.Results showed a high degree of agreement with
+the Kendall's W of .93 indicating that 93% of the
+possible rank variance is accounted for.6.4.1  Decision-Making ErrorsDetailed examinations of the informationprocessing sequences performed during tactical
+decision making have revealed a variety of errors.
+On average, subjects failed to take required ac-
+tions, about half of the time. Explanations based
+in the cognitive psychology literature have been
+pursued, as a major goal of the TADMUS pro-
+gram is to develop a DSS based on an under-
+standing of the way in which human decision
+makers actually process information under rapidly
+evolving situations.The majority of documented errors involvederrors of omission, that is, "failure to take defen-
+sive measures” and "failure to adhere to ROE.”
+Failure to take defensive measures included fail-
+ure to take actions to defend own-ship when an
+approaching aircraft had reached its weapon's re-
+lease range. An example involved a case where
+two contacts were within the specified ROE limit,
+yet no action had been taken. The types of actions
+included in the “failure to adhere to ROE” cate-
+gory include failure to take action regarding the
+items listed and defined below: (a) issuing warn-ings is part of the usual identification process andinvolves three levels of warnings with increasing
+levels of urgency; (b) establish friendly force cri-teria refers to establishing a plan with otherfriendly ships in the area to coordinate how they
+will respond to potential threats; (c) changes inkinematics/ identification friend or foe—subjectsare expected to notice significant kinematic
+changes and/or identification friend or foe pa-
+rameter changes; and (d) other identification pro-cedures includes actions such as illuminating withfire control radar.The other major category of tactically signifi-cant error involved “loss of SA.” Loss of SA er-
+rors were grouped under errors of commission and
+errors of omission and then further categorized
+into subgroups. Fifty-five percent of the loss of
+SA errors involved taking the wrong action (error
+of commission) while 45% of the errors involved
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      13Monterey, Ca.failing to take some required action (error ofomission). Error categories included under errors
+of commission involved incorrectly engaging a
+track (3%), incorrectly warning a track (29%),
+other incorrect actions (16%), and incorrect re-
+porting (7%). The two instances of incorrectly en-
+gaging an aircraft, which were F-1 Mirage aircraft,
+were considered errors because the decision
+maker failed to take certain actions prior to en-
+gaging—not necessarily because the aircraft
+should not have been engaged. The actions that
+the decision makers failed to take involved ascer-
+tain-ing the identification of the aircraft for one
+case and failure to warn and illuminate the aircraft
+prior to engaging for the second case. Most in-
+stances of incorrectly issuing warnings to the air-
+craft involved issuing the warning when the air-
+craft was within its territorial airspace (that is, in-
+side the 12 nautical mile limit which is interna-
+tionally recognized as under control of that nation)
+or issuing a warning at a level different from what
+was required. Other incorrect actions included il-
+luminating the aircraft, “locking up” with radar, or
+ordering the aircraft to divert when the aircraft
+was still within its territorial airspace. Incorrect
+reporting involved inaccurate reports on the status
+of the tactical situation (such as, indicating to the
+battle group commander that certain actions had
+been taken when they had not, misidentification of
+an aircraft, or omitting critical tracks from a re-
+port).Errors of omission categorized under the “lossof SA” category included: (a) failure to identify or
+attend to a contact; (b) failure to take action (e.g.,
+to issue “hold-fire” when a contact turned out-
+bound); (c) failure to recognize a threat (e.g.,
+designating an aircraft as a non-threat because it
+had passed its closet point-of-approach, yet it was
+still within missile-launch range); (d) instances of
+confusion or forgetting (e.g., forgetting or ignor-
+ing critical data, forgetting whether or not it had
+been warned, illuminated, or “locked-on,” or for-
+getting the aircraft's response, or lack of response
+to these actions, forgetting the status of a contact,
+and confusing contacts); (e) misperception of data
+(e.g., reporting a contact as turning outbound
+when it is still inbound); (f) unclear communica-
+tion (issuing vague orders regarding actions to betaken by team member, such as, failure to specifywhich weapon system is to be used or which con-
+tact is to be engaged).6.4.2  Cognitive explanationsThe cause of failures to take required actionsis, in many cases, attributed to the extremely high
+task demands levied on the decision maker by the
+scenario and the human decision-maker's limited
+attentional resources. Many cases are also attrib-
+uted to working memory limitations.  Maintaining
+an awareness of the status of each contact and the
+status of many actions to be taken by the antiair
+warfare team—which actions have been taken and
+what the contact's response to the action was—
+severely taxes the decision maker's working
+memory. The high workload entailed in the
+scenarios produces a highly time-compressed
+decision making situation. This time-compressed
+decision making situation—where attentional
+resources and working memory capacity are
+limited—do not allow the decision maker to
+maintain accurate SA for all tracks at any given
+time. We anticipate that the decision support
+modules in the DSS will mitigate these types of
+errors.Human information processing capabilities arenot well suited to dealing with a "multiplicity of
+simultaneous and disjointed tasks. Thoughtful at-
+tention is modular:  People can consciously think
+about only one thing at a time" (Adams, et al,
+1995, p. 92). As a result, they do not handle inter-
+ruptions very well. Research indicates that when
+an operator is faced with as few as two tasks that
+consist of merely the detection or recognition of
+simple signals, a cost may be incurred in terms of
+a significant loss in sensitivity or time that can be
+allocated to either by the requirement to divide or
+switch attention between them (Broadbent, 1957;
+Schneider and Detweiler, 1988; Swets, 1984).The memory demands of managing complex,multi-task situations can easily surpass human
+limitations. The decision maker must not forget
+any of the contacts or tasks requiring action. In
+addition to remembering all the tasks needing at-
+tention, however, are the complexities entailed in
+keeping track of the data and substeps associated
+with each contact and prior action. The aviation
+literature provides many examples of incidents
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      14Monterey, Ca.with explanations similar to the root causes forerrors that were found in the TADMUS program.
+One category includes the potentially disastrous
+effects of interruptions in the task for air traffic
+controllers and pilots. Similarly, in the AAW en-
+vironment, momentary intervening attention to
+another task or contact, or an interruption in a
+procedure can leave the procedure, or processing
+of a contact incomplete with potentially cata-
+strophic results.A fairly consistent pattern of tactical decision-making errors was documented from data col-
+lected during the baseline data collection period.
+The root causes of these errors were traced to
+cognitive mechanisms such as limited attentional
+resources and working memory limitations. By
+developing an understanding of the pattern and
+types of errors most frequently observed in this
+task domain we hope to provide a DSS which will
+mitigate these errors.7  DiscussionFailure to take appropriate actions may be ex-plained by the limited resource capacity of human
+memory. In these scenarios a large number of
+contacts are monitored for changes in any of sev-
+eral key parameters. Three modules in the DSS
+are hypothesized to assist with recognizing a
+problem and taking the appropriate actions: track
+history; response manager; and the track priority
+list and alerts.Features offered by the DSS to address errorsattributed to limited attentional resources include
+focusing attention on (1) high priority contacts
+(i.e., track priority list and alerts), as well as on (2)
+missing data (e.g., basis for assessment), and (3)
+enabling the decision maker to use more data than
+is typically used in current systems (e.g., track
+history, comparison to norms). Current systems
+require the user to retain previous contact data in
+memory to compare with current values for criti-
+cal parameters. Current systems also require the
+user to rely on recall of vast amounts of informa-
+tion from training and experience. Presenting all
+known data on a contact in a synthesized way
+should reduce working-memory requirements and
+facilitate recognition. Additional potential per-
+formance enhancement features, offered by theDSS, include displaying the complete kinematiccontact history, presenting graphic displays of lo-
+cation and trends, highlighting missing data, pro-
+viding alerts, and providing assessments of cur-
+rent contact identity that go beyond what existing
+systems currently present.Focusing the user's attention on trend andhistory data should decrease the cognitive work-
+load imposed by these scenarios where many
+contacts must be identified and responded to un-
+der severe time constraints. Similarly, delineating
+trend and history data can assist in the identifica-
+tion of a contact where noticing changes in critical
+parameters is essential. Presentation of trend and
+history data, as well as threat assessment and
+comparison to norms, should also mitigate cogni-
+tive "tunnel vision" effects where the decision
+maker attends to a smaller number of cues when
+under stress.The notion of time is an important character-istic of situation awareness (Harwood, Barnett,
+and Wickens, 1988). The past is critical to under-
+standing the present, and both past and present
+information must be used to predict future events
+(Shrestha, et al, 1995). Endsley (1988) referred to
+the "projection of their (perceived elements) status
+in the near future" when discussing situational
+awareness. However, Endsley also noted the task
+of attending to incoming information and subse-
+quently predicting future events places a heavy
+load on working memory. Several decision sup-
+port modules were developed to assist the user in
+remaining aware of the contact's history and
+changes over time. Remembering which actions
+are to be taken at what time levies an additional
+burden by placing a heavy load on working mem-
+ory. A secondary time savings should be achieved
+by the DSS acting as an intelligent "advisor," that
+is, by assisting the decision maker in knowing
+what actions to take, when to take them, and
+which actions have already been taken. A tertiary
+time savings can be achieved by including a tem-
+plate depicting the weapons' release ranges so the
+decision maker does not need to rely on fallible
+human memory or query a team member regard-
+ing weapons ranges. By graphically depicting a
+synthesized view of a contact's kinematic history,
+with the focus on changes in the contact's behav-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      15Monterey, Ca.ior over time, along with the contact's weaponsenvelope in relation to both own-ship's radar and
+weapon coverage, information processing time
+can be saved for the decision maker.8  ConclusionsThe research reported here focuses on developinga  DSS which reflects the natural decision-making
+strategies of humans. Presenting synthesized in-
+formation in the form of graphic presentations is
+expected to reduce the cognitive processing load
+for the decision maker when performing situation
+assessment. The intention is to aid the decision
+maker by providing information in a way that will
+minimize the need to maintain information in
+working memory, reduce information processing
+demands, help focus attentional resources on the
+highest priority contacts, remind the user of ac-
+tions which need to be taken, help make decisions
+under stress, and support higher levels of situation
+awareness.Decision support systems require that the hu-man's strengths be used in synergy with the ad-
+vantages offered by the DSS. Limitations associ-
+ated with the current generation of automated de-
+cision aids include the idea that (1) they cannot
+adequately capture the expertise developed by ex-
+perience over time and (2) since all contingencies
+cannot be anticipated, the expert's abilities to use
+intuition is indispensable (Mosier, in press). Mo-
+sier's review of the limitations of automated deci-
+sion systems delineates the characteristics of hu-
+man expertise that surpass the capabilities of
+automated systems. These include the human ca-
+pacity for creativity, adaptability, the ability to
+incorporate experience, the presence of a broad
+focus, analogical reasoning, and commonsense
+knowledge. The goal for the DSS is to capitalize
+on the strengths of the human along with the ad-
+vantages provided by the decision support system.9  Testing the DSSThe prototype DSS display modules are currentlybeing empirically evaluated in the simulated tacti-
+cal environment provided in the DEFTT Labora-
+tory. Experienced naval decision makers engage
+in experimental scenarios with and without access
+to the DSS display. The various decision supportmodules will be tested individually and in combi-nation in future experiments. Data on reduction of
+errors, improvements in users' situation awareness
+scores, changes in communication patterns, and
+subjective responses to the decision support sys-
+tem will be collected.10   Future ResearchWhile tools based on both the RPD and explana-tion-based reasoning models of decision making
+are included in the DSS there is no direct connec-
+tion between the two. Research is currently being
+conducted to extend schema theory to dynamic
+decision-making situations. This involves devel-
+oping and testing a hybrid model of cognitive be-
+havior in decision making to incorporate both
+types of knowledge, i.e., feature matching and
+story generation, as elements of the same schema
+model of naturalistic decision making (Smith &
+Marshall, in press). Schema theory as described
+by these authors, offers a context for integrating
+these two models which have typically been
+viewed as separate entities.AcknowledgmentsThe authors gratefully acknowledge the assistanceof Steve Francis, Brent Hardy, C.C. Johnson, Pat
+Kelly, Ron Moore, Connie O’Leary, Pat Marvel,
+Mike Quinn, and Will Rogers in data collection,
+interpretation, developing the DSS, and in con-
+ducting this research.ReferencesAdams, M. J., Tenney, Y. J., & Pew, R. W.(1995). Situation Awareness and the Cognitive
+Management of Complex Systems. HumanFactors, 37(1), 85-104.American Psychological Association, ScienceAgenda, Fall, 1988.Bower, G. H.  (1970).  Organizational Factors inMemory. Cognitive Psychology, 1, 18-46.Broadbent, D. E. (1957).  A mechanical model forhuman attention and immediate memory. Psy-chological Review, 64, 205-215.Buck, L.  (1989).  Human error at sea.  HumanFactors Bulletin, September, 32 (9), 12.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      16Monterey, Ca.Casey, S.  (1993).  Set Phasers on Stun and OtherTrue Tales of Design, Technology, and Hu-
+man Error.  Aegean Publishing Company,Santa Barbara, CA.Casner, S. (1991).  Task-Analytic Design ofGraphic Presentations. Technical Report  AIP-145.  Departments of Computer Sci-ence and Psychology, Carnegie Mellon Uni-
+versity and Learning Research and Devel-
+opment Center, University of Pittsburgh, PA.Cohen, M. S.  (1993).  The Bottom Line:  Natu-ralistic Decision Aiding.  In G. Klein, J. Ora-
+sanu, R. Calderwood, & C. E. Zsambok (Eds.)Decision Making in Action:  Models andMethods (pp. 138-147).  Ablex PublishingCorporation, New Jersey.Dawes, R. M.  (1971).  A case study of graduateadmissions:  Application of three principles of
+human decision making. American Psycholo-gist.  34, 571-582.Dotterway, K. A. (1992).  Systematic Analysis ofComplex Dynamic Systems:  The case of the
+USS Vincennes.  Unpublished master's thesis,Naval Postgraduate School, Monterey, CA.Endsley, M. R. (1988).  Design and evaluation forsituational awareness enhancement.  Pro-ceedings of the Human Factors 32nd Annual
+Meeting, 97-101.Federico, P. A.  (1995).  Expert and novice recog-nition of similar situations.  Human Factors,37(1), 105-122.Formal Investigation into the Circumstances Sur-rounding the Downing of a Commercial Air-
+line by the USS Vincennes, by Rear Admiral
+William M. Fogerty, USN, 28 July 1988, pp.
+4, 6.  Hereafter referred to as the Fogerty Re-
+port.Gruner, W. P. No Time For Decision Making.(1990, November). U.S. Naval Institute Pro-ceedings, 39-41.Harwood, K., Barnett, B., and Wickens, C.(1988). Situational awareness:  A conceptual
+and methodological framework. Proceedingsof the Symposium Psychology in the Depart-
+ment of Defense.Hockey, G. R. (1986).  Changes in Operator Effi-ciency as a Function of Environmental Stress,
+Fatigue, and Circadian Rhythms.  In:  K. R.Boff, L. Kaufman, J. P. & Thomas (Eds.):Handbook of Perception and Human Per-formance.  Wiley, New York.Hutchins, S. G. (in press). Decision-MakingEvaluation Facility for Tactical Teams. Naval
+Command, Control, and Ocean Surveillance
+Center, RDT&E Division Technical Report, in
+press, San Diego, CA.Hutchins, S. G., Kelly, R. T. and Morrison, J. G.(in press). Decision Support for Tactical Deci-
+sion Making Under Stress.  Proceedings of theSecond International Symposium on Com-
+mand and Control Research and Technology.June 1996, Monterey, CA.Hutchins, S. G. and Kowalski, J. T.  (1993).  Tac-tical Decision Making Under Stress:  Prelimi-
+nary Results and Lessons Learned.  Proceed-ings of the 10th Annual Conference on Com-
+mand and Control Decision Aids.  June 1993,Washington, D. C.Hutchins, S. G. and Rummel, B. K. (1995).  ADecision Support System for Tactical Deci-
+sion Making Under Stress.  Proceedings of the
+First International Conference on Command
+and Control Research and Technology.  June
+1995, Washington, D. C.Hutchins, S. G. and Westra, D. P. (1995). Patternsof Errors Shown by Experienced Navy Com-
+bat Information Center Teams.  Proceedingsof the 39th Annual Meeting of the Human
+Factors and Ergonomics Society, San Diego,CA. October 1995.Hutchins, S. G. and Westra, D. P. (in preparation).TADMUS Baseline Experimental Results.
+Naval Command, Control, and Ocean Sur-
+veillance Center, RDT&E Division Technical
+Report, in preparation, San Diego, CA.Kaempf, G. L. and Militelo, L. G., (1992). Deci-sion Making in Emergencies, First Offshore
+Installation Management Conference: Emer-
+gency Command Responsibilities Collected
+Papers, Aberdeen, Scotland.Kaempf, G. L. , Wolf, S. and Miller, T. E.  (1993).Decision Making in the Aegis Combat Infor-
+mation Center. In Proceedings of the Human
+Factors and Ergonomics Society 37th Annual
+Meeting (pp. 1107-1111).  Santa Monica, CA:
+Human Factors and Ergonomics Society.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      17Monterey, Ca.Kelly, R. T. , Hutchins, S. G., and Morrison, J. G.(1996). Decision Processes and Team  Com-
+munications with a Decision Support System.Proceedings of the Second International Sym-posium on Command and Control Research
+and Technology. June 1996, Monterey, CA.Kelly, R. T. , Morrison, J. G., and  Hutchins, S.G., (in press).  Impact of Naturalistic Decision
+Support on Tactical Situation Awareness.Proceedings of the 40th Annual Meeting of theHuman Factors and Ergonomics Society,Philadelphia, PA. September 1996.Kirshenbaum, S. S. (1992).  Influence of Experi-ence on Information-Gathering Strategies.Journal of Applied Psychology, 77, 343-352.Klein, G. A. (1989). Recognition-Primed Decisions. In W. R. Rouse (Ed.)  Advancesin Man-Machine Systems Research (pp. 47 - 92), Vol. 5. JAI Press, Inc.Klein, G. A. (1993).  A Recognition-Primed Deci-sion (RPD) Model of Rapid Decision Making.
+In G. A. Klein, J. Orasanu, R. Calderwood, &
+C. E. Zsambok (Eds.)  Decision Making inAction:  Models and Methods (pp. 138-147).Ablex Publishing Corporation, New Jersey.Larkin,  J. H. and Simon, H. A. (1987). Why adiagram is (sometimes) worth 10,000 words.Cognitive Science, 1, 65-99.Mosier, (in press). Myths associated with Auto-mated Decision Aids. In G. A. Klein, & C. E.
+Zsambok (Eds.) Advances in Naturalistic De-cision Making:  Research and Applications,Hillsdale, NJ:  Erlbaum.Mundy, C. E., Jr.  (1994).  Thunder and Light-ning:  Joint Littoral Warfare, Joint ForceQuarterly,  4, Spring, 45-50.Noble, D.  (1989).  Application of theory of cog-nition to situation assessment.  Vienna, VA:Engineering Research Associates.Noble, D.  (1993).  A Model to Support Devel-opment of Situation Assessment  Aids.  n G.
+A. Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.)  Decision Making in Action:Models and Methods (pp. 287-305). AblexPublishing Corporation, New Jersey.Norman, D. A.  (1988).  The Psychology of Eve-ryday Things.  Basic Books, Inc. New York.Norman, D. A. and Bobrow, D. G.  (1975).  Onthe data-limited and resources-limited proc-
+esses.  Cognitive Psychology, 7, 44-64.Office of Naval Technology.  (1992).  FY 1993Program Plan for Tactical Decision-Making
+Under Stress, Arlington, VA: July 1992.Paradies, M.  (1991).  Root Cause Analysis andHuman Factors. Human Factors Society Bul-letin, 34(8), 1-4.Paradies, M. & Unger, L. (1991).  TapRoot Inci-dent Investigation System Manual. Volumes1-7. System Improvements, Inc. Knoxville,
+TN.Pennington, N.  & Hastie, R (1992). Explainingthe Evidence: Tests of the Story Model of De-
+cision Making. Journal of Personality and So-cial Psychology, Vol. No. 2, 189-206.Pennington, N. & Hastie, R   (1993).  A theory ofExplanation-Based Decision Making.  In G. A.
+Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.) Decision Making in Action:Models and Methods (pp. 188-201).  AblexPublishing Corporation, New Jersey.Perrow, C.  (1984).  Normal Accidents:  Livingwith High Risk Technologies.  New York, Ba-sic Books, Inc.Rasmussen, J.  (1986). Information Processingand Human-Machine Interaction.  In A. P.
+Sage (Ed.) Series Volume 12, North-Holland,
+Amsterdam.Roberts, N. C. & Dotterway, K. A.  (1995). TheVincennes Incident:  Another Player on theStage?  Defense Analysis Vol 11, No. 1, pp.31-45.Salthouse, T. A. (1992).  Cognition and Context.Science, 257, 982-983.Sarter, N. B. and Woods,  D. D. (1991).  Situationawareness:  A critical but ill-defined phe-
+nomenon.  The International Journal of Avia-tion Psychology, 1(1), 45-57.Schneider, W., and Detweiler, M.  (1988). Therole of practice in dual-task performance:
+Toward workload modeling in a connectionist/
+control architecture. Human Factors, 30, 539-566.Senders, J.  W. & Moray, N. P.  (1991).  HumanError:  Cause, Prediction, and Reduction.Lawrence Erlbaum Associates, New Jersey.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      18Monterey, Ca.Shrestha, L. B., Prince, C., Baker, D. P. and Salas,E., (1995). Understanding Situation Aware-
+ness: Concepts, Methods, and Training. Hu-man/Technology Interaction in Complex Sys-
+tems.  Vol 7, W. B. Rouse (Ed.).  San Fran-cisco:  JAI Press.Simon, H. A.  (1978).   Information-processingtheory of human problem solving.  In W. K.
+Estes (Ed.), Handbook of Learning and Cog-nitive Processes, Vol 5, Human InformationProcessing, New York: Wiley.Smith, D. E. and Grossman, J. D. (1993).  Under-standing and Aiding Decision Making in
+Time-Constrained and Ambiguous Situations.Unpublished manuscript.Smith, D. E. and Marshall, S.  (in press).  Apply-ing Hybrid Models of Cognition in Decision
+Aids. In G. A. Klein, & C. E. Zsambok (Eds.)Advances in Naturalistic Decision Making:Research and Applications, Hillsdale, NJ:Erlbaum.Swets, J. A. (1984).  Mathematical models of at-tention. In R. Parasuraman and R. Davies
+(Eds.), Varieties of Attention (pp. 183-242).New York: Academic.Tolcott, M. A. (1991). Understanding and AidingMilitary Decisions. Office of Naval ResearchEuropean Office. 27th International Applied
+Psychology Symposium, Stockholm, Sweden,
+June 1991.Wilson, G. L. & Zarakas, P.  (1978).  Anatomy ofa blackout.  IEEE Spectrum, February, 339-346.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      19Monterey, Ca.The requirement to interleave a mul-tiplicity of tasks—although not necessary
+an ongoing characteristic of shipboard
+scenarios—represents the type of situation
+where providing decision support may
+make the critical difference in the outcome
+for a scenario. For example, during the
+experimental scenarios the decision mak-
+ers may have to perform the following:  monitor ship location (relative toother ships and objects in the vicinity)  monitor and apply rules of en-gagement to all applicable tracks in the
+local        operating area
+
+  receive and send radio messagesto the battle group commander and other        operating units in the area
+  monitor tracks on the Aegis dis-play system and maintain situation 
+                   awareness for all contacts of
+interest  monitor performance of actionstaken by team members to assess the       situation
+  monitor the tactical action offi-cer's/commanding officer's performance  maintain communications withCIC team members regarding their       assessment of tracks and vari-ous actions taken
+In broad perspective, although teammembers spend much of their time in rou-
+tine activities, a number of different atten-
+tionally demanding, knowledge-intensive,
+and procedurally complex tasks may de-
+mand attention at any moment. Each of
+these tasks is usually triggered by a
+stimulus event, such as a communication
+from a team member or an alert, and, in
+order to obtain proper interpretation, mayrequire additional information-seeking be-havior.  The cognitive challenge of select-
+ing and interpreting information to main-
+tain and revise one's SA is inherently
+complex. (Jager, Tenney, and Pew, 1995
+HF) Problems arise when, in the dynamic
+and multidimensional environments of
+some littoral antiair warfare scenarios, the
+situation-critical data become more time-
+compressed or ambiguous than humans
+can handle within the inherent time con-
+straints of the evolving scenario.Resources are such things as processing effort,the various forms of memory capacity, and
+communications channels (Bobrow & Norman,
+1975, in Rasmussen et al).Topics to be discussed include:  (1) a descriptionof the difficult tasks identified for analysis; (2) the
+general methodological approach; (3) develop-
+ment of the performance measures and issues
+related to their development; (4) discussion of the
+; and (5) discussion of the types of errors made by
+decision makers and interpretations for the cause
+of these errors based in the cognitive psychology
+literature.
+Paper presented to the Second International Command and Control Research and Technology Symposium,20Monterey, Ca.tactical operations require decision makingconditions of time pressure, stress, am-
+biguous, inaccurate and missing informa-
+tion, uncertain communications, and
+shifting conditions. These conditions make
+it difficult to perform careful analysis prior
+to making decisions.  Traditionally, most
+decision research and decision support
+system development have focused on well-
+defined tasks in carefully controlled envi-
+ronments.  Recent research has suggested
+that tactical decision makers use experi-
+ence to generate a likely course of action
+and then evaluate its feasibility using
+mental simulation.One way to prevent the same type of casualtyfrom being repeated is to thoroughly investigate
+and analyze root causes of actual mishaps, as well
+as data collected in a simulated tactical
+environment, and apply the findings in a concrete
+manner to improve tactical decision making.According to the recognition-primed decision-making (RPD) model, experienced decision
+makers can make rapid, high-quality decisions by
+associating a situation directly with the actions
+that normally work well in that kind of situation.
+Roberts, K. H., Stout, S. K., and Halpern, J. J.(1994)  Decision Dynamics in Two High
+Reliability Military Organizations.  Man-agement Science, Vol. 40, No. 5, May1994, 614-624.Tetlock, P. E., Accountability and the Persever-ance of First Impressions, Social Psycho-logical Quarterly, 26(1983), 285-292.Tetlock, P. E., Accountability:  The NeglectedSocial Context of Judgment and Choice, in
+L. L. Cummings and B. M. Straw (Eds.),
+Research in Organizational Behavior, Vol.7, JAI Press, Greenwich, CT, 1985, 297-332.The comparison to norms tool is based on acognitive model of human information processing
+which uses a feature matching strategy.  The
+model proposes a data-driven process.  As such,
+the comparison to norms tool is a knowledge-
+based tool with the knowledge represented as
+templates. Each template is a linked timeline
+associating individual features, through a series of
+feature matches, with expected actions. This
+allows the tool to make an assessment of the
+situation presented to the decision maker. A
+related module is the response manager. This
+assessment consists of a categorization of the
+situation (e.g., “hostile aircraft attacking”) and
+presentation of an associated template. The
+response manager module shows a template for an
+assessment of the current situation. This is a
+timeline display for template features, or events.
+The response manager module also provide inputs
+to the Prioritized track list, Alerts, and Responses
+and Tripwires.SABER is a model of another cognitivestrategy employed in making decisions. This
+strategy is known as explanation-based reasoning,
+or story generation.  In this approach, available
+data are assembled into explanatory structures,
+with one structure for each possible conclusion.
+Each of the explanations attempts to explain how
+every piece of data can be accounted for in
+support of each conclusion, even though some of
+the data items would naturally contradict reaching
+some conclusions. Contradictory data are
+explained through the of internal assumptions.  It
+is assumed that there are a fixed number of pre-
+defined possible conclusions and each data item
+points directly to one of those possible conclu-
+sions.Once the explanations are constructed,SABER evaluates them to determine which seems
+most plausible. Plausibility is based on three
+Paper presented to the Second International Command and Control Research and Technology Symposium,21Monterey, Ca.criteria:  simplicity, completeness, and impor-tance.  SABER provide data products which relate
+data to explanatory hypotheses for the current
+situation.  Hypotheses are presented in rank order,
+with evidence for and against each hypothesis and
+missing data is presented below the corresponding
+hypothesis. For a more detailed description of the
+DSS the reader is referred to Hair & Pickslay,
+1992; Hair, Pickslay, & Chow, 1992; Hutchins
+and Rummel, 1994.Hair, D. C. and Pickslay, K. (1992). Explanation-Based Reasoning in Decision Support
+Systems, Proceedings of the 9th AnnualConference on Command and Control De-cision Aids, June 1992, Monterey, CA.Hair, D. C.,  Pickslay, K. & Chow, S.  (1992)Explanation-Based Decision Support in
+Real Time Situations.  Proceedings of the1992 IEEE International Conference onTools with AI.  Nov. 1992, Arlington, VA.Various studies have indicated that as much as 90
+percent of industrial and system failures are
+produced by human error (Senders & Moray,
+1991).
+At the same time, a shift was occurring in U.S.
+Navy doctrine, away from a "blue-water" strategy
+to a doctrine of littoral operations aimed at
+potentially hostile regional powers.Tactical decision makers in today's oper-ating environment are required to perform
+complex tasks in a highly dynamic envi-
+ronment. Numerous interactive surface
+units and aircraft whose parameters are in
+flux and which must be continually
+sensed, processed, their future status pro-
+jected (development of hypotheses re-
+garding their future behavior) and actions
+taken to assure the successful outcome.Operating close to land presents additionalchallenges to the tactical decision maker.
+In littoral (i.e., near-land) settings, which
+most likely represent the majority of future
+anticipated naval conflicts, the decisions to
+be made are even more complex than they
+would be in full-scale warfare. When ci-
+vilian and neutral nation resources are in
+the conflict area incoming information
+carries an added element of uncertainty. In
+these situations, interpretation of the rules
+of engagement, contact identification, de-
+termining the capability and possible in-
+tent of the potential threat, and the
+shoot/no-shoot decision often pose ex-
+tremely difficult decision problems. (Since
+70 percent of the world's population lives
+within 200 miles of the sea, most future
+contingencies are likely to involve littoral
+warfare (Mundy, 1994).)as opposed to replacing "the user's approach tothe problem" (emphasis in original, Cohen, 1993)with tools based on decision analysis or
+mathematical optimization,
+ A considerably smaller number are attributable to
+other causes such as mechanical, electrical and
+materials failure (Meshkati, 1993). The purpose for this phase of the TADMUSprogram is to empirically evaluate the effective-
+ness of a DSS based on these recent approaches to
+decision support, to the extent possible,"This focus on why errors occuris...different from...the typical study of
+human errors which solely emphasize what
+occurs, a point of view which has received
+considerable criticism." (Rouse & Rouse,
+1983, p. 539)

36733527-1-23

@@ -0,0 +1,1319 @@
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      2Monterey, Ca.tact identification, intent, available responses andtheir consequences. For example, the close prox-
+imity of U.S. Navy forces and potential adversary
+forces makes interpreting the actions of an in-
+bound aircraft who does not respond to radio
+warnings much more difficult. Should the air-
+craft's behavior be interpreted as an attack profile,
+or does the pilot merely intend to harass, or does
+the aircraft in question not carry the equipment
+necessary to receive verbal warnings, leaving the
+pilot unable to  receive radio warnings directed
+toward him and unaware of his precarious posi-
+tion? In extreme cases there is no clear cut right or
+wrong answer about a decision. Rapidly unfolding
+events result in severe time pressure and severe
+(often catastro-phic) consequences for errors.
+While current real-time battle management sys-
+tems are well-suited to the demands of all-out
+conflicts, they may not be optimized for littoral
+situations where human intervention in decision-
+making is even more important (Office of Naval
+Technology [ONT], 1992). (Since 70 percent of
+the world's population lives within 200 miles of
+the sea, most future contingencies are likely to
+involve littoral warfare (Mundy, 1994).)Two unfortunate and highly publicized eventsfocused attention on the difficult types of deci-
+sions confronting naval commanders and provided
+the impetus for this research. In the case of the
+U.S.S. Stark, the commander made the decision to
+not engage an inbound aircraft which was be-
+lieved to not be a threat to his ship, and 27 U.S.
+naval personnel lost their lives as a result. In the
+case of the U.S.S. Vincennes, the commander
+made the decision to engage the inbound aircraft
+believing it was a threat to his ship—which turned
+out to be a commercial airliner—and all personnel
+aboard the airliner were killed as a result. In rec-
+ognition of the complex and difficult decisions
+required in these types of situations the Tactical
+Decision Making Under Stress (TADMUS) pro-
+gram was initiated to conduct research in the areas
+of human factors and training technology. The
+objective is to develop and apply principles that
+can help avoid these types of situations in the fu-
+ture. This paper, and a two companion papers
+(Hutchins, Kelly, & Morrison, 1996; Kelly,
+Hutchins, & Morrison, 1996), report on a multi-year, multi-experiment research effort conductedunder the TADMUS program to apply recent de-
+velopments in decision theory and human-system
+interaction technology to the design of a decision
+support system (DSS) for enhancing tactical deci-
+sion making under highly complex conditions.1.1 "Naturalistic" and Classical Decision-Making Para-digmsIn the same time frame that these tragic acci-dents occurred, a radical shift was occurring in the
+way psychologists viewed human decision mak-
+ing. Research was now focused on experienced
+decision makers performing their normal tasks in
+natural settings. "Naturalistic" decision-making
+research studies the decision strategies people ac-
+tually use in bringing their expertise to bear under
+challenging real-world conditions. Decision
+making milieus encompassed under the naturalis-
+tic paradigm include hospital emergency rooms,
+aircrew flight coordination, military command and
+control settings, process control systems, and po-
+lice and fire units.The naturalistic perspective, also known as"everyday cognition," is based on a belief that
+cognitive functions "elicited in natural settings
+(are) likely to differ, either quantitatively, or
+qualitatively, from those that occur in artificial or
+contrived situations, and results from sterile and
+contrived situations may not generalize to less
+constrained and more natural environments"
+(Salthouse, 1992, p. 982). There is a growing
+body of work that demonstrates that experienced,
+real-world decision makers rarely use traditional
+resource intensive strategies to make decisions in
+the face of dynamic, adverse conditions and time-
+pressure (Kaempf & Militelo, 1992; Klein, 1989;
+1993). Instead, experts rely on their abilities to
+recognize and appropriately classify situations:
+these abilities are based on having much experi-
+ence in the task domain. Once they know what
+they are facing they also tend to know what re-
+sponse option to apply, based on retrieval from
+memory of typical responses and outcomes that
+worked well in past similar situations. They use
+the limited time available to evaluate the feasibil-
+ity of that option before implementing it. Experi-
+enced decision makers recognize the situation or
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      3Monterey, Ca.scenario based on a comparison of the features ofthe current situation with stored memory represen-
+tations, or schemata. Schemata are highly inter-
+con-nected clusters of knowledge concerning cer-
+tain situations, or particular problem types, and
+associated actions or solution procedures (Fede-
+rico, 1995). Once the situation is recognized, so-
+lutions are stimulated by activation of these mem-
+ory representations.In contrast to the naturalistic perspective, ear-lier analytical methods applied in decision support
+systems were primarily used for option generation
+and evaluation, rather than for situation assess-
+ment. Traditional decision theorists argue that op-
+timum decision making involves thorough analy-
+sis of all the available data and the evaluation of
+all possible hypotheses; these approaches tend to
+rely on extensive calculations designed to arrive at
+optimal solutions. Processes for making decisions
+where the person weighs the pros and cons of
+various options and selects the one option that
+provides the most benefit are described in the de-
+cision making literature. The extensive time re-
+quirements and complicated mathematical calcu-
+lations involved (e.g., Multi-Attribute Utility
+Analysis), however, make these approaches unre-
+alistic for situations requiring rapid decision
+making. Generally, these analytical strategies in-
+volve the following steps (Kaempf & Militello,
+1992):  specify all relevant features of the task;
+  identify the full range of options;
+  identify the key evaluation dimensions;
+  identify weights for each dimension;
+  rate each option on each dimension;
+  tabulate the results, and
+  select the best option.These analytical strategies may be appropriate for
+inexperienced subjects making decisions about
+novel tasks, but not for experienced personnel
+making real-time decisions. In natural settings,
+time constraints and the difficulty in assigning
+weights and rating the dimensions involved render
+classical analysis techniques untenable. Typically
+in realistic settings experts employ recognition-
+based reasoning, not classical analytical ap-
+proaches. Experienced decision makers use their
+extensive knowledge to seek information, identifyand interpret the problem, understand the signifi-cance, derive the intention (where possible),
+model the situation (as time allows), select the
+action, evaluate the choice, and anticipate the con-
+sequence. "This decision cycle is distinctly differ-
+ent from classical models, which are based on the
+assumption that all options, outcomes, and prefer-
+ences are known and calculated in advance"
+(Federico, 1995, p. 106).Traditionally, research in decision making hasbeen directed largely toward situations in which
+(1) decision makers have sufficient time to gener-
+ate options, conduct option assessment, and select
+a course of action; (2) the consequences of an in-
+correct response are not immediately severe; (3)
+decisions are reached via consensus of a group;
+and (4) workload is manageable. Little research
+has been conducted into the development of tacti-
+cal decision support systems for use in naturalistic
+situations characterized by time pressure, high
+risk, uncertainty and information ambiguity, high
+workload, team coordination demands and task
+complexity (ONT, 1992).The central hypothesis for the research re-ported here is that presenting decision makers
+with decision support tools which were designed
+to parallel the cognitive strategies employed by
+experts, as observed in naturalistic settings, will
+reduce the number of decision making errors. This
+is accomplished by developing the architecture
+and algorithms to process information the same
+way research indicates humans do under similar
+circumstances.2   Tactical Decision Making TasksThe global tactical decision-making task involvesidentification of and responding to numerous
+contacts. When an aircraft (or a surface contact) is
+detected the CIC personnel work as a team to de-
+termine the identity and to try to determine
+whether or not the aircraft poses a threat. The high
+degree of inherent ambiguity associated with
+contact information can often make threat assess-
+ment a very difficult task. This is because many
+pieces of data fit multiple hypotheses regarding
+threat assessment. The global response choices
+(that is, engage, monitor, do nothing) are largely
+determined by the ship's orders and the current
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      4Monterey, Ca.geopolitical situation. Specific actions (such as,change course, issue verbal warnings, illuminate
+with radar, challenge with other sensors, etc.) de-
+pend on the local conditions and the relative posi-
+tions of the inbound contact of interest and own-
+ship. Determining which of these actions is likely
+to be effective depends on maintaining an accurate
+threat assessment which requires "continual up-
+dating in accordance with recurrent situation as-
+sessments" (Sarter & Woods, 1991, p. 52).This decision problem presents a highly chal-lenging cognitive task, that is, making inferences
+and deductions from incomplete and uncertain
+information derived from multiple sources and
+relating to several concurrent threats (or potential
+threats) under time-compressed conditions. The
+cognitive functions performed by the tactical deci-
+sion maker are both data and resource limited
+(Norman & Bobrow, 1975). Decisions are re-
+source limited by the mental resources of the deci-
+sion makers, who must maintain large amounts of
+information in memory under conditions of high
+workload and stress. The decisions are data lim-
+ited by the inability of the sensors to provide
+complete, error-free, unambiguous data to support
+the identification process. In particular, the ex-
+perimental scenarios were designed to follow the
+pattern of being set in an ambiguous situation
+where one or more threats of uncertain origin and
+uncertain intent approach either own ship or the
+ship being protected and may not respond to
+warnings. Scenarios were designed to be highly
+ambiguous, as this quality of uncertainty is in-
+dicative of the types of decisions to be made in
+current and future scenarios.2.1  Threat AssessmentIn the antiair warfare problem, threat assess-ment is particularly difficult because the available
+information is often incomplete or ambiguous.
+The ambiguity could be due to (a) the information
+transmission characteristics of the transmission
+medium, such as a radar transmission or a radio
+report that is only intercepted on an intermittent
+basis, (b) deliberate deceptive actions (such as ra-
+dar jamming) by the pilot flying the aircraft, or (c)
+the overlapping classification categories typical of
+many parameter measure-ments. For example, air-craft can typically fly at altitudes ranging between2,000 and 40,000 feet. Generally, an aircraft that
+is flying above 20,000 feet is not considered to be
+a threat. Conversely, an aircraft below 10,000 ft is
+considered to be more of a potential threat.  How-
+ever, aircraft flying in the middle range, (that is,
+below 20,000 ft. and above 10,000 ft.) can be
+much more difficult to categorize. Because many
+aircraft do fly in this middle range, other variables
+need to be considered in conjunction with altitude.
+This same situation of overlapping categorization
+categories exists for several other variables. These
+variables include radars that are found on both
+threat and non-threat platforms, country of origin,
+and measures of course and speed. In the case of
+speed, for example, when an aircraft flying at a
+low altitude decreases speed this could be viewed
+as indicative of a threat action (that is, slowing
+down in order to obtain better targeting informa-
+tion); however, at the same time, there could be
+other viable explanations for an aircraft's de-
+creasing speed.If the decision maker had access to all dataabout a contact approximately twelve variables
+would be used to determine identity and to infer
+intent. Two or three of these items, alone, do not
+provide definitive answers because, in many
+cases, these parameter values do not fall within
+clear-cut ranges for a particular assessment cate-
+gory (i.e., threat, non-threat). Thus, a single time
+slice of information provides an incomplete pic-
+ture of the situation. In the dynamic, ambiguous
+conditions characteristic of littoral operations, the
+rate and direction of change (data history) can
+help one better assess the threat and predict the
+future state of the situation. When the incoming
+information changes over time, the integration of
+information as it changes can help the user extract
+the message (Kirshenbaum, 1992). The DSS was
+designed to do precisely this: to facilitate the inte-
+gration process and present a synthesized picture
+of the situation to the user in a format that can be
+quickly assimilated. The variables, that are used to
+develop a threat assessment, can be divided into
+two classes: sensor information (raw or computer-
+processed information) and the contact's response,
+or lack of response, to actions taken by the team.
+These other actions, and the integration of the in-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      5Monterey, Ca.formation received via the contact's response orlack of response to them, are necessary to clarify
+the tactical picture.2.2  Situation AwarenessWhile recent increasing interest among re-searchers regarding the concept of situation
+awareness (SA) has generated a debate on the pre-
+cise definition of this term most researchers ac-
+knowledge the importance of the concept. In gen-
+eral, SA refers to the decision maker's moment-
+by-moment ability to monitor and understand the
+state of the complex system and its environment
+(Adams, Tenney, & Pew, 1995). These authors
+state the essential idea which is that when emer-
+gencies arise, the completeness and accuracy of
+the decision maker's SA are critical to the ability
+to make decisions, revise plans, and manage the
+system. Specific decision-making tasks included
+under SA include the ability to:  (1) maintain an
+accurate perception of the surrounding environ-
+ment (both internal and external to the ship); (2)
+identify problems and/or potential problems; (3)
+recognize a need for action; (4) note deviations in
+the mission; and (5) maintain awareness of tasks
+performed (Shrestha, Prince, Baker and Salas,
+1995). To maintain an accurate SA the decision
+maker should take into account both information
+that is available and that which can be activated
+from memory (Sarter and Woods, 1991).However, a difficulty arises as a result of theheavy workload imposed by this process. When
+the decision maker is faced with several concur-
+rent contacts of interest, all of which have numer-
+ous associated data items (i.e., as many as  a
+dozen), some or all of which may change over the
+course of the scenario (e.g., intelligence, active
+radar emitters, various kinematic parameters, etc.)
+memory load easily exceeds human capacity.
+Moreover, changing parameters may impart dif-
+ferent interpretations to what is occurring. In some
+cases the moment-to-moment attentional demands
+of a tactical situation are relentless and unforgiv-
+ing (such as, a terrorist aircraft directly inbound
+toward "own-ship" which can result in "task fixa-
+tion"), sometimes relevant background knowledge
+is unavoidably incomplete (such as, an unfamiliar
+aircraft), and sometimes the decision maker is al-ready thinking and working as hard as possible,even when there are no unanticipated events when
+there is a high contact density (Adams, et al,
+1995). These instances provide a few illustrations
+of situations that can degrade situation awareness.Complex information gathering and process-ing systems have been designed to aid the deci-
+sion-maker in the past. However, these systems
+often increase the decision-maker’s burden due to
+the inherent system complexity and the failure to
+design them in a way that they will fit the user's
+cognitive processing limitations. Often, these
+systems require operators to perform difficult
+cognitive tasks under heavy workloads. They must
+perceive, synthesize and determine the relevance
+of a continual stream of incoming information,
+often pertaining to several concurrent contacts,
+while projecting future anticipated events and
+making decisions regarding actions to be taken.
+Decision makers must assess, compare, and re-
+solve conflicting information, while making diffi-
+cult judgments, and remembering the status of
+critical contacts along with the contact's response
+to actions taken by the CIC team. These decision-
+making tasks are interleaved with other required
+tasks, such as keeping other team members in-
+formed (both on and off the ship). Furthermore,
+these complex tasks are performed under condi-
+tions where adverse environmental (noise, vibra-
+tion, temperature extremes, etc.) and internal
+stressors (boredom, fatigue, anxiety, and fear) are
+part of the environment.3  Decision Support PrinciplesA case has been made that previous generations ofdecision support systems, which focus primarily
+on solution optimization and base decision sup-
+port on normative models of human decision
+making, are less applicable than a DSS that par-
+allels the cognitive strategies used by domain ex-
+perts in situations characterized by time-
+constrained situations with uncertain and ambigu-
+ous data (Smith & Grossman, 1993).  These
+authors point out that rarely, if ever, were earlier
+tactical decision aids intended as psychological
+models of human cognitive behavior.  Instead,
+these aids performed "complex and burdensome
+calculations, reducing the work-load on personnel,
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      6Monterey, Ca.speeding up the dissemination of information, andproviding more time for command decision mak-
+ing" (Tolcott, 1991, p. 44).3.1  Feature Matching and Story GenerationWe have applied two of these new models ofhuman decision making—which parallel the cog-
+nitive strategies used by domain experts—to the
+design of a DSS for enhancing antiair warfare
+tactical decision making. These two models that
+people use in assessing a situation are feature
+matching and story generation. The feature
+matching model, described by Noble (1989), in-
+volves an organization of memory, or "schemas,"
+and information-processing where decision mak-
+ers use their previous experiences to assess a
+situation and identify promising actions. Incoming
+information is categorized, selected, edited, and
+organized on the basis of a person's general
+knowledge about a domain. Both story generation
+and feature matching occur under conditions
+where a large base of implication-rich, condition-
+ally dependent pieces of evidence must be evalu-
+ated before choosing an alternative from a set of
+prospective courses of action. The feature match-
+ing model applies a spatio-temporal dependence,
+whereas story generation is an example of causal
+dependence. According to the explanation-based
+model, decision makers construct a causal model
+to explain the available evidence (Pennington &
+Hastie, 1993). At the same time, the decision
+maker creates a set of alternatives from which an
+action will be chosen.  A decision is made when a
+story is successfully matched to an alternative in
+the choice set. Story generation occurs in complex
+situations where the decision maker may not have
+all the necessary information or when a series of
+facts may appear to contradict each other. The de-
+cision maker must then develop causal links be-
+tween these facts to produce a coherent picture of
+the situation (Klein, 1989; 1993).The explanation-based reasoning model isbased on research which found that jurors develop
+a narrative story to organize trial information
+where causal and intentional relations between
+events are central (Pennington & Hastie, 1992).
+Pennington & Hastie propose four certainty prin-
+ciples—coverage, coherence, uniqueness, andgoodness-of-fit—that govern (i) which story willbe accepted, (ii) which decision will be selected,and the (iii) confidence or degree of certainty withwhich a particular decision will be made. This or-
+ganization of the evidence by the decision maker
+is believed to facilitate evidence comprehension.
+A central component of this model is that the
+story the juror constructs determines the juror's
+decision.  This explanation-based decision process
+is employed when the body of evidence relevant
+to a decision is large, complex, and the implica-
+tions of its components are interdependent.Feature matching, also referred to as the rec-ognition-primed decision (RPD) model, "occurs
+when the decision maker recognizes the features
+of the present situation as similar or identical to
+those of a previous situation" (Kaempf & Militelo,
+1992, p. 6). An adequate match triggers recall of
+information learned about this type of situation:
+(a) plausible goals, (b) critical cues to be moni-
+tored, (c) expectations of what should happen, and
+(d) a course of action that worked in similar situa-
+tions. According to this recent approach, expert
+decision makers may rely on well-developed
+memory representations to guide decision making
+in new (but similar) situations. The RPD model of
+decision making fuses two processes—situation
+assessment and mental simulation (Klein, 1993).
+In the simplest case the situation is recognized as
+familiar or prototypical, using feature matching,
+and the obvious response is implemented. In a
+more complex case the decision maker performs a
+conscious evaluation of the response, using men-
+tal simulation to uncover problems prior to im-
+plementing the response. In the most complex
+case the evaluation reveals flaws requiring modi-
+fication, or the option is judged inadequate and
+rejected in favor of the next most typical reaction.3.2  Situation AssessmentIn general, the overall task of responding toantiair warfare scenarios consists of situation as-
+sessment ("what's going on") and course of action
+selection ("what to do about it"). Recent theories
+of decision making emphasize the importance of
+situation assessment for good decision making in
+naturalistic, event-driven situations. Moreover,
+they stress that decisions regarding actions to be
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      7Monterey, Ca.taken are a by-product of developing the situationawareness that precedes action selection. Klein
+(1989) has found that usually the situation itself
+either determines or constrains the response op-
+tions and that experienced decision makers make
+up to 90% of all decisions without considering
+alternatives. If the situation appears similar to one
+that the decision maker has previously experi-
+enced, the pattern will be recognized and the
+course of action is usually immediately obvious.
+On the other hand, if the situation does not seem
+familiar complex RPD will be involved where the
+decision maker adjusts the option after evaluating
+it.Additional evidence was found in the specifictask domain of interest to the TADMUS program
+which added support to these findings on the way
+real-world decision makers make decisions in the
+context of their normal jobs. Research was con-
+ducted to determine decision requirements for
+command-level decision makers in the combat
+information center (CIC) of an Aegis cruiser.
+Analysis of 14 incidents from actual problems re-
+vealed 183 decisions. Of these, 103 concerned
+situation assessments. Results obtained after ana-
+lysts coded these situation assessments indicated
+that decision makers arrived at approximately
+87% of their situation assessments through feature
+matching and the remaining 13% through story
+generation (Kaempf, Wolf, & Miller, 1993). The
+other eighty decisions that were identified, from
+analysis of the real-world incidents mentioned
+above, involved course of action selection. These
+course of action decisions served a variety of
+functions, although, relatively few were intended
+to end the incident. Twenty were intended as a
+final course of action decision; 14 were imple-
+mented to obtain more information, 22 to manage
+resources, and 24 to put themselves in a more fa-
+vorable tactical position. A recognition-based
+strategy was also used by decision makers to de-
+velop a course of action, accounting for 95% of
+the actions taken in the 14 incidents. The decision
+makers generated and compared multiple options
+in only 5% of the cases. In line with these find-
+ings, the TADMUS program has adopted the po-
+sition that decision aiding systems should assist in
+the decision making process, and focus on aidingthe situation assessment portion of the decision-making task.A DSS was developed to support decision-making processes which research has shown are
+used by decision makers in real-world settings
+(Hutchins, Kelly, & Morrison, 1996). Specifi-
+cally, the DSS parallels the strategies used by ex-
+perienced decision makers to perform situation
+assessment (Nobel, 1989; 1993). This approach to
+supporting the user's intuitive approach to dealing
+with dynamic decision-making situations should
+produce tools that are both more easily understood
+and used, and that more effectively "exploit the
+decision maker's knowledge and expertise that
+might facilitate adaptation to complex, novel
+situations" (Cohen, 1993, p. 265).4  Human-System Interaction PrinciplesThe vast majority of research on human-computerinteraction design has been devoted to character-
+istics of displays that impact human perception,
+such as symbol legibility or detectability, and on
+relatively simple cognitive functions such as
+memory tasks. Fewer efforts have been devoted to
+understanding the effects of the format and man-
+ner in which information is presented on more
+complex levels of human cognition such as deci-
+sion making. Consequently, principles that can be
+applied to the design of the interface between the
+user and a decision support system for the purpose
+of enhancing cognitive changing situations, are
+not available to any significant degree (ONT,
+1992).4.1  Graphic PresentationsSeveral advantages are offered by graphicpresentations over a text-based presentation for-
+mat (Larkin and Simon, 1987). Graphic presenta-
+tions should (1) reduce the amount of mental
+computation required to perform tasks; and (2)
+allow users to spend less time searching for
+needed information. Casner (1991) elaborated on
+these ideas and found that graphics allow users to
+substitute less demanding perceptual opera-tions
+for more complex logical operations. For exam-
+ple, determining a change in altitude (and the de-
+gree of change) is immediately apparent when the
+user glances at the track history module. (Note
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      8Monterey, Ca.that the words contact and track can be used inter-changeably. The reader is referred to Figure 1.)The objective for the track history module isto facilitate the contact identification process by
+providing information that is integrated in a way
+that supports a recognitional decision strategy.
+This module depicts a contact’s speed, altitude,
+course and range on a two-dimensional graphical
+display along with a geometric representation of
+both the contact’s weapon release envelope and
+own-ship’s weapons coverage. A large amount of
+parametric data is portrayed graphically for rapidassimilation by the user. The user can see, at aglance, a synthesized picture of the contact’s be-
+havior. Compare this rather simple perceptual op-
+eration with the more complex logical operation
+involved in current operational systems which re-
+quire the user to recall and subtract numerical val-
+ues for past and current altitudes.Graphics also allow users to omit steps thatare otherwise necessary when a task is performed
+without a graphic. An example of this advantage
+isFigure 1.  Decision Support System Display Modules.also illustrated in the track history module whichincludes templates indicating weapon's coverage
+for both the inbound contact and "own-ship." To
+determine whether the aircraft is within its
+weapon's launch range there is no need to recall
+the specific launch range values and then compare
+them with the aircraft's current range. Instead, the
+user can determine if the aircraft is within its
+launch range by a quick glance at the display.Graphics help users save time when searchingfor needed information when several related di-
+mensions of information are encoded in a single
+graphical object. This is accomplished by inte-
+grating the kinematic parameters of speed, course,
+altitude, bearing, and range for a contact. The user
+can see, at a glance, a synthesized picture of the
+contact's behavior.  Compare this process with
+reading, in a text-based format, the individual pa-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      9Monterey, Ca.rameters which need to be integrated by the userinto a coherent picture of the contact's behavior.5   Limited Cognitive Processing CapabilitiesSince there are limits to the cognitive processingcapability of humans, it is important for the sys-
+tem to provide the needed information in a format
+that best supports the user operating under dy-
+namic decision-making conditions. It may be the
+case that current systems are inadequate to support
+the cognitive processing demands required by
+certain littoral scenarios. For example, according
+to Gruner (1990, p. 41), the U.S.S. Vincennes of-
+ficers and system operators "could not make better
+decisions because they did not have time to con-
+firm or deny the information uncertainties pre-
+sented them." Gruner maintains that the rapid pace
+involved in these types of situations can exceed
+the capacity of the human to comprehend the
+rapid flow of information presented by complex
+systems. In the case of the Vincennes, the CIC
+team had three minutes and 40 seconds to make
+their decision. This includes the time required for
+the operators to perceive and interpret sensor data
+and for the commanding officer to make informed
+judgments from these data (Roberts & Dotterway,
+1995). The result of the human's limited cognitive
+processing capabilities is that the decision makers
+may fail to remember critical pieces of data,
+overlook stored information, draw hasty conclu-
+sions, and produce flawed answers.Evidence of the effects of limitations in mem-ory and shared attention capacity on human deci-
+sion making were found during baseline testing
+(Hutchins & Kowalski, 1994; Hutchins & Westra,
+1995) and during empirical evaluation of the DSS
+(Kelly, Hutchins, & Morrison, 1996; Kelly, Mor-
+rison, & Hutchins. 1996).Simon (1978, p. 273) states, "...the human in-formation processing system...operates almost en-
+tirely serially, one process at a time, rather than in
+parallel fashion. This seriality is reflected in the
+narrowness of its momentary focus of attention."
+However, the AAW problem forces the decision
+maker to operate in a parallel processing mode
+when several contacts demand attention at the
+same time. The requirement to monitor and
+maintain an accurate SA for these concurrentcontacts, over the course of the evolving situation,imposes an additional load of strategically man-
+aging the overall situation. Several researchers
+have argued that "managing the attentional and
+conceptual processes that permit cogent SA in-
+volves significant cognitive resources" (Adams, et
+al, 1995, p. 91; Endsley, 1988). The tasks of pri-
+oritizing contacts and the associated actions to be
+taken by the team, updating the status of critical
+contacts, responding to the other requisite tasks in
+the queue and, more generally, of strategically
+managing the workload of current multitask sys-
+tems under dynamically changing scenarios can
+place an unrealistic cognitive load on the decision
+maker.A major advantage offered by the experimen-tal DSS is that it should "buy time" for the user by
+(1) performing many of the cognitive processing
+tasks for the user and (2) by presenting informa-
+tion in graphic format. The DSS will synthesize
+much of the information used to develop situation
+awareness and present a coherent picture of the
+situation to the user. This integrated picture will
+be portrayed graphically—rather than in the cur-
+rent text-based format—which should further re-
+duce the amount of time required to assimilate
+this information. By performing several informa-
+tion processing steps for the decision maker the
+decision maker's limited cognitive resources can
+be used for the types of decisions which require
+human abilities (e.g., the decision on whether to
+engage).5.1  Working Memory RequirementsAn essential information processing step re-quired by this task—and one which levies a heavy
+load on working memory—involves integrating
+kinematic and sensor variables and maintaining an
+awareness of changes in these variables over time.
+Changes in a contact's behavior such as, decreas-
+ing altitude, increasing speed, changes in elec-
+tronic emissions, etc., can provide key indicators
+of possible hostile intent. With current systems,
+the decision maker receives numerous reports
+from CIC team members who provide various
+pieces of the overall tactical picture (such as,
+kinematic parameter values, active electronic
+emitter identifications, and behavioral responses
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      10Monterey, Ca.of the contact in response to queries by the team)regarding a particular contact. Some of this infor-
+mation is also displayed in a text-based format for
+the user when a contact is "hooked" (that is, se-
+lected for display) by the decision maker. How-
+ever, to recognize a change in certain variables,
+current systems require the user to retain parame-
+ter values in short-term memory in order to recog-
+nize a change in the parameter, such as altitude.When the decision maker is monitoring sev-eral concurrent contacts (such as, cycling through
+three or four contacts in a 1-minute period) human
+working memory capabilities may quickly be sur-
+passed. To detect a change in a critical parameter
+value, the decision maker must maintain the pa-
+rameter values for the contacts of interest in
+working memory as he or she cycles between sev-
+eral contacts. For example, the decision maker
+must be able to recall that contact 7022 was at
+14,000 ft. altitude one minute ago, and then sub-
+tract the current altitude value of 10,000 ft., which
+will then indicate the aircraft is in a rapid descent.
+The DSS was developed to aid the decision maker
+by performing several of these cognitive process-
+ing tasks, thus, reducing the cognitive load for the
+user. By presenting the synthesized picture of the
+contact's behavior over time, through the use of
+graphical displays,  critical changes should be
+immediately apparent to the user.A second memory-intensive task involvesmaintaining, in working memory, a current list of
+actions taken by team members, the contact's re-
+sponse to these actions taken by the CIC team,
+and pending actions. Research has established that
+"memory is limited and that list maintenance is
+effortful and fallible—more so if the list must be
+ordered and still more if the membership of the
+list must be dynamically reordered and modified
+during retention" (Bower, 1970, as cited in Ad-
+ams, et al, 1995, p. 91).  The DSS should reduce
+the cognitive effort required for distributing atten-
+tion among the many contacts to be attended to
+and actions that are required. Working memory
+requirements should be reduced by having the
+DSS act as an intelligent "assistant," reminding
+the user regarding what actions are to be taken and
+when the actions are to be taken.A third way the DSS will reduce memory andinformation processing requirements is by dis-
+playing templates depicting weapons' envelopes
+for both the inbound contact and "own-ship." This
+should facilitate critical comparisons and judg-
+ments regarding timing of actions. During a sce-
+nario decision makers have to either rely on mem-
+ory to recall the launch range for various weapons
+or query a team member for this information. Both
+of these methods waste limited resources. The
+high workload and high tempo characteristic of
+littoral scenarios produce a stress-ful decision-
+making environment. The phenom-enon that in-
+creasing stress leads to decreasing working mem-
+ory is well documented (e.g., Hockey, 1986). The
+latter method for obtaining the desired informa-
+tion wastes limited resources by increasing the
+communications load and requiring more time to
+wait for a team members' response to the queryUnder these high-tempo and high workloadconditions human memory and attentional re-
+sources can easily be surpassed. Several cogni-
+tively resource intensive information processing
+steps are eliminated for the human decision maker
+by having them performed by the DSS. We pre-
+dict that the decision support tools will reduce the
+cognitive workload imposed on the decision
+maker in the following three ways: (1) by reduc-
+ing the amount of information processing to be
+performed, (2) reducing working memory re-
+quirements, and (3) assisting the user in allocating
+limited attentional resources.5.2  Reducing Human ErrorThe study of human cognitive processes andrelated error mechanisms has gained rapidly in-
+creasing interest in the past decade. Rasmussen
+(1987) argues that the emphasis in attempting to
+understand human errors must shift from tasks to
+the human-task mismatch.  For example, Gruner
+(p. 39), in discussing the Vincennes incident,
+maintains that "the system was poorly suited for
+use by human beings during rapid military action."
+He ascribes this lack of suitability to a human-
+machine mismatch between the rate of data flow
+possible with modern computer systems that can
+process and display information at phenomenal
+data rates and the "comprehension capability of
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      11Monterey, Ca.users which has remained almost static for thou-sands of years." This causal approach to under-
+standing human error is based on the premise that
+errors are rarely random and can be traced to
+causes and contributing factors. Once these con-
+tributing factors are identified they can be miti-
+gated.The impact and vulnerability of systems andhuman interfaces, because of incompatibilities
+between the way people perceive, think, and act,
+are documented in the popular and technical lit-
+erature (Buck, 1989; Casey, 1993; Norman, 1988;
+Perrow, 1984; Wilson & Zarakas, 1978). Newly
+developed systems will succeed or fail based on
+our ability to minimize these incompat-ibilities
+between the characteristics of the things we create
+and the way we use them. There are many well-
+documented instances of critical systems or pa-
+rameter changes going unnoticed or unheeded be-
+cause the operating procedures, or the human ma-
+chine interface, provided no historical trace. For
+example, an unnoticed increase in altitude con-
+tributed to the shoot down of the Iranian airbus by
+a U.S. Navy ship—when the team mistakenly be-
+lieved the aircraft to be descending—because
+there was no historical trace to make the aircraft's
+actual increasing altitude apparent (Dotterway,1992).  Five personnel in the U.S.S. Vincennes's
+combat information center, all viewing separate
+displays, reported the aircraft as descending while
+the Aegis data tapes later revealed a flight pattern
+of ascent (Roberts & Dotterway, 1995). One of the
+official investiga-tions of this incident, the Fo-
+garty Report (1988, p. 45), states that "stress, task
+fixation, and an unconscious distortion of data
+may have played a major role in this incident." A
+panel of five psychologists from the American
+Psychological Association who testified before
+Congress concluded that there were "predictable
+failings of human judgment under intense stress
+compounded by complex technology [which]
+clearly contributed to the accidental shooting of
+Iranian airliner Flight 655" (APA, p. 4).It is generally accepted that between 60-80percent of the accidents and malfunctions in
+transportation, manufacturing, process control,
+weapon, and other systems are attributable to hu-
+man error (Senders & Moray, 1991; Van Cott,1993; Weiner, 1994). Reducing tactical decisionmaking errors is one goal of the TADMUS pro-
+gram.  The following section presents a brief re-
+view of an experiment conducted to develop a
+baseline on tactical decision making performance
+in response to fairly stressful scenarios. A com-
+panion paper (Hutchins, Kelly, & Morrison, 1996)
+describes the experimental DSS modules and the
+way they are hypothesized to enhance tactical de-
+cision making performance.6   TADMUS Baseline ExperimentEarly research involved data collection in the De-cision-Making Evaluation Facility for Tactical
+Teams (DEFTT) Laboratory using simulated ex-
+isting shipboard displays to establish a baseline on
+decision-making performance. The purpose of this
+effort was to document baseline decision-making
+performance for experienced naval officers. Dur-
+ing the baseline phase of testing, a detailed under-
+standing was developed of the cognitive processes
+underlying the various tasks involved in situation
+assessment—and where the bottlenecks occur.
+This understanding was then used to design the
+way the information is presented to the user in or-
+der to facilitate performance of the required tasks.6.1  SubjectsThis study focused on the command-level de-cision makers of an antiair warfare team on an
+Aegis cruiser—the commanding officer and the
+tactical action officer. Subjects in the study con-
+sisted of six commanding officer/tactical action
+officer teams drawn from twelve active duty Na-
+val personnel; some were from training com-
+mands while others were from operational com-
+mands aboard ship or assigned to group staffs.6.2  ProcedureData were collected in the  DEFTT Labora-tory, a six-station test-bed environment that
+simulates console positions in a Navy Aegis
+cruiser combat information center. (For a detailed
+description of the DEFTT Laboratory see Hutch-
+ins, 1996.) Four stations were filled by confeder-
+ates (active duty Navy personnel) who play antiair
+warfare support-team member roles. These roles
+included the antiair warfare coordinator, identifi-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      12Monterey, Ca.cation supervisor, tactical information coordina-tor, and electronic warfare supervisor. After ap-
+proximately 1 1/2 hours of orientation to the labo-
+ratory and training in the use of the computer con-
+soles the subjects engaged in four scenarios. The
+scenarios were each about 25 minutes in length
+and contained between 11 and 14 contacts of in-
+terest per scenario, in addition to numerous back-
+ground contacts.6.3  Treatment of DataTeam communications were recorded on amultichannel audio recorder; these included all
+intra-team exchanges, as well as all communica-
+tions with simulated off-ship personnel. Audio
+tapes were used to produce verbatim, time-
+stamped transcripts of all team communications.
+A modified version of the TapRoot® Incident In-
+vestigation System (Paradies, 1991; Paradies and
+Unger, 1991) was then applied to identify errors.
+The objective was to identify tactically significant
+errors committed during the scenario. Tactically
+significant errors were defined as those errors that
+may lead to loss of life or significant political em-
+barrassment. The following criteria were used for
+counting an error as tactically significant:  (1) loss
+of situation awareness, (2) failure to take defen-
+sive action when within the weapon's range of an
+approaching contact, or (3) a violation of rules of
+engagement (ROE). Video recordings were made
+of the commanding officer and tactical action of-
+ficer computer screens. Detailed analyses of all
+audio and video recordings were conducted. (For
+a more detailed coverage of the methodology and
+results see Hutchins and Westra, in preparation).6.4  ResultsThe complex, time-constrained, decision-making situations embodied in the experimental
+scenarios resulted in a large number of decision
+errors. The mean number of tactically significant
+errors documented across six teams and four sce-
+narios was 14; the number of errors ranged from
+nine to twenty-two. The standard deviation was
+3.7.  Subjects performed an average of 50% of the
+required behaviors as specified in the rules of en-
+gagement. The ordinal agreement between three
+raters (navy subject matter experts) on error countranks from TapRoot® analyses was computed.Results showed a high degree of agreement with
+the Kendall's W of .93 indicating that 93% of the
+possible rank variance is accounted for.6.4.1  Decision-Making ErrorsDetailed examinations of the informationprocessing sequences performed during tactical
+decision making have revealed a variety of errors.
+On average, subjects failed to take required ac-
+tions, about half of the time. Explanations based
+in the cognitive psychology literature have been
+pursued, as a major goal of the TADMUS pro-
+gram is to develop a DSS based on an under-
+standing of the way in which human decision
+makers actually process information under rapidly
+evolving situations.The majority of documented errors involvederrors of omission, that is, "failure to take defen-
+sive measures” and "failure to adhere to ROE.”
+Failure to take defensive measures included fail-
+ure to take actions to defend own-ship when an
+approaching aircraft had reached its weapon's re-
+lease range. An example involved a case where
+two contacts were within the specified ROE limit,
+yet no action had been taken. The types of actions
+included in the “failure to adhere to ROE” cate-
+gory include failure to take action regarding the
+items listed and defined below: (a) issuing warn-ings is part of the usual identification process andinvolves three levels of warnings with increasing
+levels of urgency; (b) establish friendly force cri-teria refers to establishing a plan with otherfriendly ships in the area to coordinate how they
+will respond to potential threats; (c) changes inkinematics/ identification friend or foe—subjectsare expected to notice significant kinematic
+changes and/or identification friend or foe pa-
+rameter changes; and (d) other identification pro-cedures includes actions such as illuminating withfire control radar.The other major category of tactically signifi-cant error involved “loss of SA.” Loss of SA er-
+rors were grouped under errors of commission and
+errors of omission and then further categorized
+into subgroups. Fifty-five percent of the loss of
+SA errors involved taking the wrong action (error
+of commission) while 45% of the errors involved
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      13Monterey, Ca.failing to take some required action (error ofomission). Error categories included under errors
+of commission involved incorrectly engaging a
+track (3%), incorrectly warning a track (29%),
+other incorrect actions (16%), and incorrect re-
+porting (7%). The two instances of incorrectly en-
+gaging an aircraft, which were F-1 Mirage aircraft,
+were considered errors because the decision
+maker failed to take certain actions prior to en-
+gaging—not necessarily because the aircraft
+should not have been engaged. The actions that
+the decision makers failed to take involved ascer-
+tain-ing the identification of the aircraft for one
+case and failure to warn and illuminate the aircraft
+prior to engaging for the second case. Most in-
+stances of incorrectly issuing warnings to the air-
+craft involved issuing the warning when the air-
+craft was within its territorial airspace (that is, in-
+side the 12 nautical mile limit which is interna-
+tionally recognized as under control of that nation)
+or issuing a warning at a level different from what
+was required. Other incorrect actions included il-
+luminating the aircraft, “locking up” with radar, or
+ordering the aircraft to divert when the aircraft
+was still within its territorial airspace. Incorrect
+reporting involved inaccurate reports on the status
+of the tactical situation (such as, indicating to the
+battle group commander that certain actions had
+been taken when they had not, misidentification of
+an aircraft, or omitting critical tracks from a re-
+port).Errors of omission categorized under the “lossof SA” category included: (a) failure to identify or
+attend to a contact; (b) failure to take action (e.g.,
+to issue “hold-fire” when a contact turned out-
+bound); (c) failure to recognize a threat (e.g.,
+designating an aircraft as a non-threat because it
+had passed its closet point-of-approach, yet it was
+still within missile-launch range); (d) instances of
+confusion or forgetting (e.g., forgetting or ignor-
+ing critical data, forgetting whether or not it had
+been warned, illuminated, or “locked-on,” or for-
+getting the aircraft's response, or lack of response
+to these actions, forgetting the status of a contact,
+and confusing contacts); (e) misperception of data
+(e.g., reporting a contact as turning outbound
+when it is still inbound); (f) unclear communica-
+tion (issuing vague orders regarding actions to betaken by team member, such as, failure to specifywhich weapon system is to be used or which con-
+tact is to be engaged).6.4.2  Cognitive explanationsThe cause of failures to take required actionsis, in many cases, attributed to the extremely high
+task demands levied on the decision maker by the
+scenario and the human decision-maker's limited
+attentional resources. Many cases are also attrib-
+uted to working memory limitations.  Maintaining
+an awareness of the status of each contact and the
+status of many actions to be taken by the antiair
+warfare team—which actions have been taken and
+what the contact's response to the action was—
+severely taxes the decision maker's working
+memory. The high workload entailed in the
+scenarios produces a highly time-compressed
+decision making situation. This time-compressed
+decision making situation—where attentional
+resources and working memory capacity are
+limited—do not allow the decision maker to
+maintain accurate SA for all tracks at any given
+time. We anticipate that the decision support
+modules in the DSS will mitigate these types of
+errors.Human information processing capabilities arenot well suited to dealing with a "multiplicity of
+simultaneous and disjointed tasks. Thoughtful at-
+tention is modular:  People can consciously think
+about only one thing at a time" (Adams, et al,
+1995, p. 92). As a result, they do not handle inter-
+ruptions very well. Research indicates that when
+an operator is faced with as few as two tasks that
+consist of merely the detection or recognition of
+simple signals, a cost may be incurred in terms of
+a significant loss in sensitivity or time that can be
+allocated to either by the requirement to divide or
+switch attention between them (Broadbent, 1957;
+Schneider and Detweiler, 1988; Swets, 1984).The memory demands of managing complex,multi-task situations can easily surpass human
+limitations. The decision maker must not forget
+any of the contacts or tasks requiring action. In
+addition to remembering all the tasks needing at-
+tention, however, are the complexities entailed in
+keeping track of the data and substeps associated
+with each contact and prior action. The aviation
+literature provides many examples of incidents
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      14Monterey, Ca.with explanations similar to the root causes forerrors that were found in the TADMUS program.
+One category includes the potentially disastrous
+effects of interruptions in the task for air traffic
+controllers and pilots. Similarly, in the AAW en-
+vironment, momentary intervening attention to
+another task or contact, or an interruption in a
+procedure can leave the procedure, or processing
+of a contact incomplete with potentially cata-
+strophic results.A fairly consistent pattern of tactical decision-making errors was documented from data col-
+lected during the baseline data collection period.
+The root causes of these errors were traced to
+cognitive mechanisms such as limited attentional
+resources and working memory limitations. By
+developing an understanding of the pattern and
+types of errors most frequently observed in this
+task domain we hope to provide a DSS which will
+mitigate these errors.7  DiscussionFailure to take appropriate actions may be ex-plained by the limited resource capacity of human
+memory. In these scenarios a large number of
+contacts are monitored for changes in any of sev-
+eral key parameters. Three modules in the DSS
+are hypothesized to assist with recognizing a
+problem and taking the appropriate actions: track
+history; response manager; and the track priority
+list and alerts.Features offered by the DSS to address errorsattributed to limited attentional resources include
+focusing attention on (1) high priority contacts
+(i.e., track priority list and alerts), as well as on (2)
+missing data (e.g., basis for assessment), and (3)
+enabling the decision maker to use more data than
+is typically used in current systems (e.g., track
+history, comparison to norms). Current systems
+require the user to retain previous contact data in
+memory to compare with current values for criti-
+cal parameters. Current systems also require the
+user to rely on recall of vast amounts of informa-
+tion from training and experience. Presenting all
+known data on a contact in a synthesized way
+should reduce working-memory requirements and
+facilitate recognition. Additional potential per-
+formance enhancement features, offered by theDSS, include displaying the complete kinematiccontact history, presenting graphic displays of lo-
+cation and trends, highlighting missing data, pro-
+viding alerts, and providing assessments of cur-
+rent contact identity that go beyond what existing
+systems currently present.Focusing the user's attention on trend andhistory data should decrease the cognitive work-
+load imposed by these scenarios where many
+contacts must be identified and responded to un-
+der severe time constraints. Similarly, delineating
+trend and history data can assist in the identifica-
+tion of a contact where noticing changes in critical
+parameters is essential. Presentation of trend and
+history data, as well as threat assessment and
+comparison to norms, should also mitigate cogni-
+tive "tunnel vision" effects where the decision
+maker attends to a smaller number of cues when
+under stress.The notion of time is an important character-istic of situation awareness (Harwood, Barnett,
+and Wickens, 1988). The past is critical to under-
+standing the present, and both past and present
+information must be used to predict future events
+(Shrestha, et al, 1995). Endsley (1988) referred to
+the "projection of their (perceived elements) status
+in the near future" when discussing situational
+awareness. However, Endsley also noted the task
+of attending to incoming information and subse-
+quently predicting future events places a heavy
+load on working memory. Several decision sup-
+port modules were developed to assist the user in
+remaining aware of the contact's history and
+changes over time. Remembering which actions
+are to be taken at what time levies an additional
+burden by placing a heavy load on working mem-
+ory. A secondary time savings should be achieved
+by the DSS acting as an intelligent "advisor," that
+is, by assisting the decision maker in knowing
+what actions to take, when to take them, and
+which actions have already been taken. A tertiary
+time savings can be achieved by including a tem-
+plate depicting the weapons' release ranges so the
+decision maker does not need to rely on fallible
+human memory or query a team member regard-
+ing weapons ranges. By graphically depicting a
+synthesized view of a contact's kinematic history,
+with the focus on changes in the contact's behav-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      15Monterey, Ca.ior over time, along with the contact's weaponsenvelope in relation to both own-ship's radar and
+weapon coverage, information processing time
+can be saved for the decision maker.8  ConclusionsThe research reported here focuses on developinga  DSS which reflects the natural decision-making
+strategies of humans. Presenting synthesized in-
+formation in the form of graphic presentations is
+expected to reduce the cognitive processing load
+for the decision maker when performing situation
+assessment. The intention is to aid the decision
+maker by providing information in a way that will
+minimize the need to maintain information in
+working memory, reduce information processing
+demands, help focus attentional resources on the
+highest priority contacts, remind the user of ac-
+tions which need to be taken, help make decisions
+under stress, and support higher levels of situation
+awareness.Decision support systems require that the hu-man's strengths be used in synergy with the ad-
+vantages offered by the DSS. Limitations associ-
+ated with the current generation of automated de-
+cision aids include the idea that (1) they cannot
+adequately capture the expertise developed by ex-
+perience over time and (2) since all contingencies
+cannot be anticipated, the expert's abilities to use
+intuition is indispensable (Mosier, in press). Mo-
+sier's review of the limitations of automated deci-
+sion systems delineates the characteristics of hu-
+man expertise that surpass the capabilities of
+automated systems. These include the human ca-
+pacity for creativity, adaptability, the ability to
+incorporate experience, the presence of a broad
+focus, analogical reasoning, and commonsense
+knowledge. The goal for the DSS is to capitalize
+on the strengths of the human along with the ad-
+vantages provided by the decision support system.9  Testing the DSSThe prototype DSS display modules are currentlybeing empirically evaluated in the simulated tacti-
+cal environment provided in the DEFTT Labora-
+tory. Experienced naval decision makers engage
+in experimental scenarios with and without access
+to the DSS display. The various decision supportmodules will be tested individually and in combi-nation in future experiments. Data on reduction of
+errors, improvements in users' situation awareness
+scores, changes in communication patterns, and
+subjective responses to the decision support sys-
+tem will be collected.10   Future ResearchWhile tools based on both the RPD and explana-tion-based reasoning models of decision making
+are included in the DSS there is no direct connec-
+tion between the two. Research is currently being
+conducted to extend schema theory to dynamic
+decision-making situations. This involves devel-
+oping and testing a hybrid model of cognitive be-
+havior in decision making to incorporate both
+types of knowledge, i.e., feature matching and
+story generation, as elements of the same schema
+model of naturalistic decision making (Smith &
+Marshall, in press). Schema theory as described
+by these authors, offers a context for integrating
+these two models which have typically been
+viewed as separate entities.AcknowledgmentsThe authors gratefully acknowledge the assistanceof Steve Francis, Brent Hardy, C.C. Johnson, Pat
+Kelly, Ron Moore, Connie O’Leary, Pat Marvel,
+Mike Quinn, and Will Rogers in data collection,
+interpretation, developing the DSS, and in con-
+ducting this research.ReferencesAdams, M. J., Tenney, Y. J., & Pew, R. W.(1995). Situation Awareness and the Cognitive
+Management of Complex Systems. HumanFactors, 37(1), 85-104.American Psychological Association, ScienceAgenda, Fall, 1988.Bower, G. H.  (1970).  Organizational Factors inMemory. Cognitive Psychology, 1, 18-46.Broadbent, D. E. (1957).  A mechanical model forhuman attention and immediate memory. Psy-chological Review, 64, 205-215.Buck, L.  (1989).  Human error at sea.  HumanFactors Bulletin, September, 32 (9), 12.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      16Monterey, Ca.Casey, S.  (1993).  Set Phasers on Stun and OtherTrue Tales of Design, Technology, and Hu-
+man Error.  Aegean Publishing Company,Santa Barbara, CA.Casner, S. (1991).  Task-Analytic Design ofGraphic Presentations. Technical Report  AIP-145.  Departments of Computer Sci-ence and Psychology, Carnegie Mellon Uni-
+versity and Learning Research and Devel-
+opment Center, University of Pittsburgh, PA.Cohen, M. S.  (1993).  The Bottom Line:  Natu-ralistic Decision Aiding.  In G. Klein, J. Ora-
+sanu, R. Calderwood, & C. E. Zsambok (Eds.)Decision Making in Action:  Models andMethods (pp. 138-147).  Ablex PublishingCorporation, New Jersey.Dawes, R. M.  (1971).  A case study of graduateadmissions:  Application of three principles of
+human decision making. American Psycholo-gist.  34, 571-582.Dotterway, K. A. (1992).  Systematic Analysis ofComplex Dynamic Systems:  The case of the
+USS Vincennes.  Unpublished master's thesis,Naval Postgraduate School, Monterey, CA.Endsley, M. R. (1988).  Design and evaluation forsituational awareness enhancement.  Pro-ceedings of the Human Factors 32nd Annual
+Meeting, 97-101.Federico, P. A.  (1995).  Expert and novice recog-nition of similar situations.  Human Factors,37(1), 105-122.Formal Investigation into the Circumstances Sur-rounding the Downing of a Commercial Air-
+line by the USS Vincennes, by Rear Admiral
+William M. Fogerty, USN, 28 July 1988, pp.
+4, 6.  Hereafter referred to as the Fogerty Re-
+port.Gruner, W. P. No Time For Decision Making.(1990, November). U.S. Naval Institute Pro-ceedings, 39-41.Harwood, K., Barnett, B., and Wickens, C.(1988). Situational awareness:  A conceptual
+and methodological framework. Proceedingsof the Symposium Psychology in the Depart-
+ment of Defense.Hockey, G. R. (1986).  Changes in Operator Effi-ciency as a Function of Environmental Stress,
+Fatigue, and Circadian Rhythms.  In:  K. R.Boff, L. Kaufman, J. P. & Thomas (Eds.):Handbook of Perception and Human Per-formance.  Wiley, New York.Hutchins, S. G. (in press). Decision-MakingEvaluation Facility for Tactical Teams. Naval
+Command, Control, and Ocean Surveillance
+Center, RDT&E Division Technical Report, in
+press, San Diego, CA.Hutchins, S. G., Kelly, R. T. and Morrison, J. G.(in press). Decision Support for Tactical Deci-
+sion Making Under Stress.  Proceedings of theSecond International Symposium on Com-
+mand and Control Research and Technology.June 1996, Monterey, CA.Hutchins, S. G. and Kowalski, J. T.  (1993).  Tac-tical Decision Making Under Stress:  Prelimi-
+nary Results and Lessons Learned.  Proceed-ings of the 10th Annual Conference on Com-
+mand and Control Decision Aids.  June 1993,Washington, D. C.Hutchins, S. G. and Rummel, B. K. (1995).  ADecision Support System for Tactical Deci-
+sion Making Under Stress.  Proceedings of the
+First International Conference on Command
+and Control Research and Technology.  June
+1995, Washington, D. C.Hutchins, S. G. and Westra, D. P. (1995). Patternsof Errors Shown by Experienced Navy Com-
+bat Information Center Teams.  Proceedingsof the 39th Annual Meeting of the Human
+Factors and Ergonomics Society, San Diego,CA. October 1995.Hutchins, S. G. and Westra, D. P. (in preparation).TADMUS Baseline Experimental Results.
+Naval Command, Control, and Ocean Sur-
+veillance Center, RDT&E Division Technical
+Report, in preparation, San Diego, CA.Kaempf, G. L. and Militelo, L. G., (1992). Deci-sion Making in Emergencies, First Offshore
+Installation Management Conference: Emer-
+gency Command Responsibilities Collected
+Papers, Aberdeen, Scotland.Kaempf, G. L. , Wolf, S. and Miller, T. E.  (1993).Decision Making in the Aegis Combat Infor-
+mation Center. In Proceedings of the Human
+Factors and Ergonomics Society 37th Annual
+Meeting (pp. 1107-1111).  Santa Monica, CA:
+Human Factors and Ergonomics Society.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      17Monterey, Ca.Kelly, R. T. , Hutchins, S. G., and Morrison, J. G.(1996). Decision Processes and Team  Com-
+munications with a Decision Support System.Proceedings of the Second International Sym-posium on Command and Control Research
+and Technology. June 1996, Monterey, CA.Kelly, R. T. , Morrison, J. G., and  Hutchins, S.G., (in press).  Impact of Naturalistic Decision
+Support on Tactical Situation Awareness.Proceedings of the 40th Annual Meeting of theHuman Factors and Ergonomics Society,Philadelphia, PA. September 1996.Kirshenbaum, S. S. (1992).  Influence of Experi-ence on Information-Gathering Strategies.Journal of Applied Psychology, 77, 343-352.Klein, G. A. (1989). Recognition-Primed Decisions. In W. R. Rouse (Ed.)  Advancesin Man-Machine Systems Research (pp. 47 - 92), Vol. 5. JAI Press, Inc.Klein, G. A. (1993).  A Recognition-Primed Deci-sion (RPD) Model of Rapid Decision Making.
+In G. A. Klein, J. Orasanu, R. Calderwood, &
+C. E. Zsambok (Eds.)  Decision Making inAction:  Models and Methods (pp. 138-147).Ablex Publishing Corporation, New Jersey.Larkin,  J. H. and Simon, H. A. (1987). Why adiagram is (sometimes) worth 10,000 words.Cognitive Science, 1, 65-99.Mosier, (in press). Myths associated with Auto-mated Decision Aids. In G. A. Klein, & C. E.
+Zsambok (Eds.) Advances in Naturalistic De-cision Making:  Research and Applications,Hillsdale, NJ:  Erlbaum.Mundy, C. E., Jr.  (1994).  Thunder and Light-ning:  Joint Littoral Warfare, Joint ForceQuarterly,  4, Spring, 45-50.Noble, D.  (1989).  Application of theory of cog-nition to situation assessment.  Vienna, VA:Engineering Research Associates.Noble, D.  (1993).  A Model to Support Devel-opment of Situation Assessment  Aids.  n G.
+A. Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.)  Decision Making in Action:Models and Methods (pp. 287-305). AblexPublishing Corporation, New Jersey.Norman, D. A.  (1988).  The Psychology of Eve-ryday Things.  Basic Books, Inc. New York.Norman, D. A. and Bobrow, D. G.  (1975).  Onthe data-limited and resources-limited proc-
+esses.  Cognitive Psychology, 7, 44-64.Office of Naval Technology.  (1992).  FY 1993Program Plan for Tactical Decision-Making
+Under Stress, Arlington, VA: July 1992.Paradies, M.  (1991).  Root Cause Analysis andHuman Factors. Human Factors Society Bul-letin, 34(8), 1-4.Paradies, M. & Unger, L. (1991).  TapRoot Inci-dent Investigation System Manual. Volumes1-7. System Improvements, Inc. Knoxville,
+TN.Pennington, N.  & Hastie, R (1992). Explainingthe Evidence: Tests of the Story Model of De-
+cision Making. Journal of Personality and So-cial Psychology, Vol. No. 2, 189-206.Pennington, N. & Hastie, R   (1993).  A theory ofExplanation-Based Decision Making.  In G. A.
+Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.) Decision Making in Action:Models and Methods (pp. 188-201).  AblexPublishing Corporation, New Jersey.Perrow, C.  (1984).  Normal Accidents:  Livingwith High Risk Technologies.  New York, Ba-sic Books, Inc.Rasmussen, J.  (1986). Information Processingand Human-Machine Interaction.  In A. P.
+Sage (Ed.) Series Volume 12, North-Holland,
+Amsterdam.Roberts, N. C. & Dotterway, K. A.  (1995). TheVincennes Incident:  Another Player on theStage?  Defense Analysis Vol 11, No. 1, pp.31-45.Salthouse, T. A. (1992).  Cognition and Context.Science, 257, 982-983.Sarter, N. B. and Woods,  D. D. (1991).  Situationawareness:  A critical but ill-defined phe-
+nomenon.  The International Journal of Avia-tion Psychology, 1(1), 45-57.Schneider, W., and Detweiler, M.  (1988). Therole of practice in dual-task performance:
+Toward workload modeling in a connectionist/
+control architecture. Human Factors, 30, 539-566.Senders, J.  W. & Moray, N. P.  (1991).  HumanError:  Cause, Prediction, and Reduction.Lawrence Erlbaum Associates, New Jersey.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      18Monterey, Ca.Shrestha, L. B., Prince, C., Baker, D. P. and Salas,E., (1995). Understanding Situation Aware-
+ness: Concepts, Methods, and Training. Hu-man/Technology Interaction in Complex Sys-
+tems.  Vol 7, W. B. Rouse (Ed.).  San Fran-cisco:  JAI Press.Simon, H. A.  (1978).   Information-processingtheory of human problem solving.  In W. K.
+Estes (Ed.), Handbook of Learning and Cog-nitive Processes, Vol 5, Human InformationProcessing, New York: Wiley.Smith, D. E. and Grossman, J. D. (1993).  Under-standing and Aiding Decision Making in
+Time-Constrained and Ambiguous Situations.Unpublished manuscript.Smith, D. E. and Marshall, S.  (in press).  Apply-ing Hybrid Models of Cognition in Decision
+Aids. In G. A. Klein, & C. E. Zsambok (Eds.)Advances in Naturalistic Decision Making:Research and Applications, Hillsdale, NJ:Erlbaum.Swets, J. A. (1984).  Mathematical models of at-tention. In R. Parasuraman and R. Davies
+(Eds.), Varieties of Attention (pp. 183-242).New York: Academic.Tolcott, M. A. (1991). Understanding and AidingMilitary Decisions. Office of Naval ResearchEuropean Office. 27th International Applied
+Psychology Symposium, Stockholm, Sweden,
+June 1991.Wilson, G. L. & Zarakas, P.  (1978).  Anatomy ofa blackout.  IEEE Spectrum, February, 339-346.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      19Monterey, Ca.The requirement to interleave a mul-tiplicity of tasks—although not necessary
+an ongoing characteristic of shipboard
+scenarios—represents the type of situation
+where providing decision support may
+make the critical difference in the outcome
+for a scenario. For example, during the
+experimental scenarios the decision mak-
+ers may have to perform the following:  monitor ship location (relative toother ships and objects in the vicinity)  monitor and apply rules of en-gagement to all applicable tracks in the
+local        operating area
+
+  receive and send radio messagesto the battle group commander and other        operating units in the area
+  monitor tracks on the Aegis dis-play system and maintain situation 
+                   awareness for all contacts of
+interest  monitor performance of actionstaken by team members to assess the       situation
+  monitor the tactical action offi-cer's/commanding officer's performance  maintain communications withCIC team members regarding their       assessment of tracks and vari-ous actions taken
+In broad perspective, although teammembers spend much of their time in rou-
+tine activities, a number of different atten-
+tionally demanding, knowledge-intensive,
+and procedurally complex tasks may de-
+mand attention at any moment. Each of
+these tasks is usually triggered by a
+stimulus event, such as a communication
+from a team member or an alert, and, in
+order to obtain proper interpretation, mayrequire additional information-seeking be-havior.  The cognitive challenge of select-
+ing and interpreting information to main-
+tain and revise one's SA is inherently
+complex. (Jager, Tenney, and Pew, 1995
+HF) Problems arise when, in the dynamic
+and multidimensional environments of
+some littoral antiair warfare scenarios, the
+situation-critical data become more time-
+compressed or ambiguous than humans
+can handle within the inherent time con-
+straints of the evolving scenario.Resources are such things as processing effort,the various forms of memory capacity, and
+communications channels (Bobrow & Norman,
+1975, in Rasmussen et al).Topics to be discussed include:  (1) a descriptionof the difficult tasks identified for analysis; (2) the
+general methodological approach; (3) develop-
+ment of the performance measures and issues
+related to their development; (4) discussion of the
+; and (5) discussion of the types of errors made by
+decision makers and interpretations for the cause
+of these errors based in the cognitive psychology
+literature.
+Paper presented to the Second International Command and Control Research and Technology Symposium,20Monterey, Ca.tactical operations require decision makingconditions of time pressure, stress, am-
+biguous, inaccurate and missing informa-
+tion, uncertain communications, and
+shifting conditions. These conditions make
+it difficult to perform careful analysis prior
+to making decisions.  Traditionally, most
+decision research and decision support
+system development have focused on well-
+defined tasks in carefully controlled envi-
+ronments.  Recent research has suggested
+that tactical decision makers use experi-
+ence to generate a likely course of action
+and then evaluate its feasibility using
+mental simulation.One way to prevent the same type of casualtyfrom being repeated is to thoroughly investigate
+and analyze root causes of actual mishaps, as well
+as data collected in a simulated tactical
+environment, and apply the findings in a concrete
+manner to improve tactical decision making.According to the recognition-primed decision-making (RPD) model, experienced decision
+makers can make rapid, high-quality decisions by
+associating a situation directly with the actions
+that normally work well in that kind of situation.
+Roberts, K. H., Stout, S. K., and Halpern, J. J.(1994)  Decision Dynamics in Two High
+Reliability Military Organizations.  Man-agement Science, Vol. 40, No. 5, May1994, 614-624.Tetlock, P. E., Accountability and the Persever-ance of First Impressions, Social Psycho-logical Quarterly, 26(1983), 285-292.Tetlock, P. E., Accountability:  The NeglectedSocial Context of Judgment and Choice, in
+L. L. Cummings and B. M. Straw (Eds.),
+Research in Organizational Behavior, Vol.7, JAI Press, Greenwich, CT, 1985, 297-332.The comparison to norms tool is based on acognitive model of human information processing
+which uses a feature matching strategy.  The
+model proposes a data-driven process.  As such,
+the comparison to norms tool is a knowledge-
+based tool with the knowledge represented as
+templates. Each template is a linked timeline
+associating individual features, through a series of
+feature matches, with expected actions. This
+allows the tool to make an assessment of the
+situation presented to the decision maker. A
+related module is the response manager. This
+assessment consists of a categorization of the
+situation (e.g., “hostile aircraft attacking”) and
+presentation of an associated template. The
+response manager module shows a template for an
+assessment of the current situation. This is a
+timeline display for template features, or events.
+The response manager module also provide inputs
+to the Prioritized track list, Alerts, and Responses
+and Tripwires.SABER is a model of another cognitivestrategy employed in making decisions. This
+strategy is known as explanation-based reasoning,
+or story generation.  In this approach, available
+data are assembled into explanatory structures,
+with one structure for each possible conclusion.
+Each of the explanations attempts to explain how
+every piece of data can be accounted for in
+support of each conclusion, even though some of
+the data items would naturally contradict reaching
+some conclusions. Contradictory data are
+explained through the of internal assumptions.  It
+is assumed that there are a fixed number of pre-
+defined possible conclusions and each data item
+points directly to one of those possible conclu-
+sions.Once the explanations are constructed,SABER evaluates them to determine which seems
+most plausible. Plausibility is based on three
+Paper presented to the Second International Command and Control Research and Technology Symposium,21Monterey, Ca.criteria:  simplicity, completeness, and impor-tance.  SABER provide data products which relate
+data to explanatory hypotheses for the current
+situation.  Hypotheses are presented in rank order,
+with evidence for and against each hypothesis and
+missing data is presented below the corresponding
+hypothesis. For a more detailed description of the
+DSS the reader is referred to Hair & Pickslay,
+1992; Hair, Pickslay, & Chow, 1992; Hutchins
+and Rummel, 1994.Hair, D. C. and Pickslay, K. (1992). Explanation-Based Reasoning in Decision Support
+Systems, Proceedings of the 9th AnnualConference on Command and Control De-cision Aids, June 1992, Monterey, CA.Hair, D. C.,  Pickslay, K. & Chow, S.  (1992)Explanation-Based Decision Support in
+Real Time Situations.  Proceedings of the1992 IEEE International Conference onTools with AI.  Nov. 1992, Arlington, VA.Various studies have indicated that as much as 90
+percent of industrial and system failures are
+produced by human error (Senders & Moray,
+1991).
+At the same time, a shift was occurring in U.S.
+Navy doctrine, away from a "blue-water" strategy
+to a doctrine of littoral operations aimed at
+potentially hostile regional powers.Tactical decision makers in today's oper-ating environment are required to perform
+complex tasks in a highly dynamic envi-
+ronment. Numerous interactive surface
+units and aircraft whose parameters are in
+flux and which must be continually
+sensed, processed, their future status pro-
+jected (development of hypotheses re-
+garding their future behavior) and actions
+taken to assure the successful outcome.Operating close to land presents additionalchallenges to the tactical decision maker.
+In littoral (i.e., near-land) settings, which
+most likely represent the majority of future
+anticipated naval conflicts, the decisions to
+be made are even more complex than they
+would be in full-scale warfare. When ci-
+vilian and neutral nation resources are in
+the conflict area incoming information
+carries an added element of uncertainty. In
+these situations, interpretation of the rules
+of engagement, contact identification, de-
+termining the capability and possible in-
+tent of the potential threat, and the
+shoot/no-shoot decision often pose ex-
+tremely difficult decision problems. (Since
+70 percent of the world's population lives
+within 200 miles of the sea, most future
+contingencies are likely to involve littoral
+warfare (Mundy, 1994).)as opposed to replacing "the user's approach tothe problem" (emphasis in original, Cohen, 1993)with tools based on decision analysis or
+mathematical optimization,
+ A considerably smaller number are attributable to
+other causes such as mechanical, electrical and
+materials failure (Meshkati, 1993). The purpose for this phase of the TADMUSprogram is to empirically evaluate the effective-
+ness of a DSS based on these recent approaches to
+decision support, to the extent possible,"This focus on why errors occuris...different from...the typical study of
+human errors which solely emphasize what
+occurs, a point of view which has received
+considerable criticism." (Rouse & Rouse,
+1983, p. 539)
+Paper presented to the Second International Command and Control Research and Technology Symposium,22Monterey, Ca. DSS which will mitigate typical types oferrors. The objective during this phase of
+the research was to develop an under-
+standing of the decision-making problems
+presented by current and future Navy sce-
+narios in order to identify the types and
+forms of information that are likely to fa-
+cilitate performance of these activities.
+Rouse, W. B. & Valusek, J. (1993). Evo-lutionary Design of Systems to Support
+Decision Making.  In G. A. Klein, J.
+Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.)  Decision Making inAction:  Models and Methods  (pp.270-286).  Ablex Publishing Corpora-
+tion, New Jersey.The Naturalistic Decision Making (NDM)model (Klein, 1989, 1993) seems more
+applicable than traditional decision-
+making models to the types of decisions
+involved in tactical decision-making.
+Early work conducted under the TADMUS
+program to determine the cognitive strate-
+gies employed by Navy AAW decision-
+makers found that when performing situa-
+tion assessment 87% of the time a recog-
+nitional strategy was used and 13% of the
+time story generation was used (Kaempf,
+Wolf, and Miller, 1993). We have applied
+these new models of human decision
+making—which parallel the cognitive
+strategies used by domain experts—to the
+design of a DSS for enhancing antiair war-
+fare tactical decision making. These two
+models for situation assessment are feature
+matching and story generation.
+(The scenarios were intentionally devel-
+oped to have many tracks with a high de-
+gree of uncertainty associated where it is
+not always clear whether a particular track
+should be engaged. Our interest in theTADMUS program was in gaining insightinto and aiding the decision process. The
+reader is referred to Hutchins, 1995, for a
+detailed coverage of performance meas-
+urement issues and scenario development
+issues.)
+These conditions include dynamic, fluid
+situations, time pressure, high risk, multi-
+ple decision makers, shifting and compet-
+ing goals, action-feedback loops, and
+situations with uncertain and incomplete
+data (Orasanu & Connolly, 1993).

PRINs_C2-1-22

@@ -0,0 +1,1319 @@
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      2Monterey, Ca.tact identification, intent, available responses andtheir consequences. For example, the close prox-
+imity of U.S. Navy forces and potential adversary
+forces makes interpreting the actions of an in-
+bound aircraft who does not respond to radio
+warnings much more difficult. Should the air-
+craft's behavior be interpreted as an attack profile,
+or does the pilot merely intend to harass, or does
+the aircraft in question not carry the equipment
+necessary to receive verbal warnings, leaving the
+pilot unable to  receive radio warnings directed
+toward him and unaware of his precarious posi-
+tion? In extreme cases there is no clear cut right or
+wrong answer about a decision. Rapidly unfolding
+events result in severe time pressure and severe
+(often catastro-phic) consequences for errors.
+While current real-time battle management sys-
+tems are well-suited to the demands of all-out
+conflicts, they may not be optimized for littoral
+situations where human intervention in decision-
+making is even more important (Office of Naval
+Technology [ONT], 1992). (Since 70 percent of
+the world's population lives within 200 miles of
+the sea, most future contingencies are likely to
+involve littoral warfare (Mundy, 1994).)Two unfortunate and highly publicized eventsfocused attention on the difficult types of deci-
+sions confronting naval commanders and provided
+the impetus for this research. In the case of the
+U.S.S. Stark, the commander made the decision to
+not engage an inbound aircraft which was be-
+lieved to not be a threat to his ship, and 27 U.S.
+naval personnel lost their lives as a result. In the
+case of the U.S.S. Vincennes, the commander
+made the decision to engage the inbound aircraft
+believing it was a threat to his ship—which turned
+out to be a commercial airliner—and all personnel
+aboard the airliner were killed as a result. In rec-
+ognition of the complex and difficult decisions
+required in these types of situations the Tactical
+Decision Making Under Stress (TADMUS) pro-
+gram was initiated to conduct research in the areas
+of human factors and training technology. The
+objective is to develop and apply principles that
+can help avoid these types of situations in the fu-
+ture. This paper, and a two companion papers
+(Hutchins, Kelly, & Morrison, 1996; Kelly,
+Hutchins, & Morrison, 1996), report on a multi-year, multi-experiment research effort conductedunder the TADMUS program to apply recent de-
+velopments in decision theory and human-system
+interaction technology to the design of a decision
+support system (DSS) for enhancing tactical deci-
+sion making under highly complex conditions.1.1 "Naturalistic" and Classical Decision-Making Para-digmsIn the same time frame that these tragic acci-dents occurred, a radical shift was occurring in the
+way psychologists viewed human decision mak-
+ing. Research was now focused on experienced
+decision makers performing their normal tasks in
+natural settings. "Naturalistic" decision-making
+research studies the decision strategies people ac-
+tually use in bringing their expertise to bear under
+challenging real-world conditions. Decision
+making milieus encompassed under the naturalis-
+tic paradigm include hospital emergency rooms,
+aircrew flight coordination, military command and
+control settings, process control systems, and po-
+lice and fire units.The naturalistic perspective, also known as"everyday cognition," is based on a belief that
+cognitive functions "elicited in natural settings
+(are) likely to differ, either quantitatively, or
+qualitatively, from those that occur in artificial or
+contrived situations, and results from sterile and
+contrived situations may not generalize to less
+constrained and more natural environments"
+(Salthouse, 1992, p. 982). There is a growing
+body of work that demonstrates that experienced,
+real-world decision makers rarely use traditional
+resource intensive strategies to make decisions in
+the face of dynamic, adverse conditions and time-
+pressure (Kaempf & Militelo, 1992; Klein, 1989;
+1993). Instead, experts rely on their abilities to
+recognize and appropriately classify situations:
+these abilities are based on having much experi-
+ence in the task domain. Once they know what
+they are facing they also tend to know what re-
+sponse option to apply, based on retrieval from
+memory of typical responses and outcomes that
+worked well in past similar situations. They use
+the limited time available to evaluate the feasibil-
+ity of that option before implementing it. Experi-
+enced decision makers recognize the situation or
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      3Monterey, Ca.scenario based on a comparison of the features ofthe current situation with stored memory represen-
+tations, or schemata. Schemata are highly inter-
+con-nected clusters of knowledge concerning cer-
+tain situations, or particular problem types, and
+associated actions or solution procedures (Fede-
+rico, 1995). Once the situation is recognized, so-
+lutions are stimulated by activation of these mem-
+ory representations.In contrast to the naturalistic perspective, ear-lier analytical methods applied in decision support
+systems were primarily used for option generation
+and evaluation, rather than for situation assess-
+ment. Traditional decision theorists argue that op-
+timum decision making involves thorough analy-
+sis of all the available data and the evaluation of
+all possible hypotheses; these approaches tend to
+rely on extensive calculations designed to arrive at
+optimal solutions. Processes for making decisions
+where the person weighs the pros and cons of
+various options and selects the one option that
+provides the most benefit are described in the de-
+cision making literature. The extensive time re-
+quirements and complicated mathematical calcu-
+lations involved (e.g., Multi-Attribute Utility
+Analysis), however, make these approaches unre-
+alistic for situations requiring rapid decision
+making. Generally, these analytical strategies in-
+volve the following steps (Kaempf & Militello,
+1992):  specify all relevant features of the task;
+  identify the full range of options;
+  identify the key evaluation dimensions;
+  identify weights for each dimension;
+  rate each option on each dimension;
+  tabulate the results, and
+  select the best option.These analytical strategies may be appropriate for
+inexperienced subjects making decisions about
+novel tasks, but not for experienced personnel
+making real-time decisions. In natural settings,
+time constraints and the difficulty in assigning
+weights and rating the dimensions involved render
+classical analysis techniques untenable. Typically
+in realistic settings experts employ recognition-
+based reasoning, not classical analytical ap-
+proaches. Experienced decision makers use their
+extensive knowledge to seek information, identifyand interpret the problem, understand the signifi-cance, derive the intention (where possible),
+model the situation (as time allows), select the
+action, evaluate the choice, and anticipate the con-
+sequence. "This decision cycle is distinctly differ-
+ent from classical models, which are based on the
+assumption that all options, outcomes, and prefer-
+ences are known and calculated in advance"
+(Federico, 1995, p. 106).Traditionally, research in decision making hasbeen directed largely toward situations in which
+(1) decision makers have sufficient time to gener-
+ate options, conduct option assessment, and select
+a course of action; (2) the consequences of an in-
+correct response are not immediately severe; (3)
+decisions are reached via consensus of a group;
+and (4) workload is manageable. Little research
+has been conducted into the development of tacti-
+cal decision support systems for use in naturalistic
+situations characterized by time pressure, high
+risk, uncertainty and information ambiguity, high
+workload, team coordination demands and task
+complexity (ONT, 1992).The central hypothesis for the research re-ported here is that presenting decision makers
+with decision support tools which were designed
+to parallel the cognitive strategies employed by
+experts, as observed in naturalistic settings, will
+reduce the number of decision making errors. This
+is accomplished by developing the architecture
+and algorithms to process information the same
+way research indicates humans do under similar
+circumstances.2   Tactical Decision Making TasksThe global tactical decision-making task involvesidentification of and responding to numerous
+contacts. When an aircraft (or a surface contact) is
+detected the CIC personnel work as a team to de-
+termine the identity and to try to determine
+whether or not the aircraft poses a threat. The high
+degree of inherent ambiguity associated with
+contact information can often make threat assess-
+ment a very difficult task. This is because many
+pieces of data fit multiple hypotheses regarding
+threat assessment. The global response choices
+(that is, engage, monitor, do nothing) are largely
+determined by the ship's orders and the current
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      4Monterey, Ca.geopolitical situation. Specific actions (such as,change course, issue verbal warnings, illuminate
+with radar, challenge with other sensors, etc.) de-
+pend on the local conditions and the relative posi-
+tions of the inbound contact of interest and own-
+ship. Determining which of these actions is likely
+to be effective depends on maintaining an accurate
+threat assessment which requires "continual up-
+dating in accordance with recurrent situation as-
+sessments" (Sarter & Woods, 1991, p. 52).This decision problem presents a highly chal-lenging cognitive task, that is, making inferences
+and deductions from incomplete and uncertain
+information derived from multiple sources and
+relating to several concurrent threats (or potential
+threats) under time-compressed conditions. The
+cognitive functions performed by the tactical deci-
+sion maker are both data and resource limited
+(Norman & Bobrow, 1975). Decisions are re-
+source limited by the mental resources of the deci-
+sion makers, who must maintain large amounts of
+information in memory under conditions of high
+workload and stress. The decisions are data lim-
+ited by the inability of the sensors to provide
+complete, error-free, unambiguous data to support
+the identification process. In particular, the ex-
+perimental scenarios were designed to follow the
+pattern of being set in an ambiguous situation
+where one or more threats of uncertain origin and
+uncertain intent approach either own ship or the
+ship being protected and may not respond to
+warnings. Scenarios were designed to be highly
+ambiguous, as this quality of uncertainty is in-
+dicative of the types of decisions to be made in
+current and future scenarios.2.1  Threat AssessmentIn the antiair warfare problem, threat assess-ment is particularly difficult because the available
+information is often incomplete or ambiguous.
+The ambiguity could be due to (a) the information
+transmission characteristics of the transmission
+medium, such as a radar transmission or a radio
+report that is only intercepted on an intermittent
+basis, (b) deliberate deceptive actions (such as ra-
+dar jamming) by the pilot flying the aircraft, or (c)
+the overlapping classification categories typical of
+many parameter measure-ments. For example, air-craft can typically fly at altitudes ranging between2,000 and 40,000 feet. Generally, an aircraft that
+is flying above 20,000 feet is not considered to be
+a threat. Conversely, an aircraft below 10,000 ft is
+considered to be more of a potential threat.  How-
+ever, aircraft flying in the middle range, (that is,
+below 20,000 ft. and above 10,000 ft.) can be
+much more difficult to categorize. Because many
+aircraft do fly in this middle range, other variables
+need to be considered in conjunction with altitude.
+This same situation of overlapping categorization
+categories exists for several other variables. These
+variables include radars that are found on both
+threat and non-threat platforms, country of origin,
+and measures of course and speed. In the case of
+speed, for example, when an aircraft flying at a
+low altitude decreases speed this could be viewed
+as indicative of a threat action (that is, slowing
+down in order to obtain better targeting informa-
+tion); however, at the same time, there could be
+other viable explanations for an aircraft's de-
+creasing speed.If the decision maker had access to all dataabout a contact approximately twelve variables
+would be used to determine identity and to infer
+intent. Two or three of these items, alone, do not
+provide definitive answers because, in many
+cases, these parameter values do not fall within
+clear-cut ranges for a particular assessment cate-
+gory (i.e., threat, non-threat). Thus, a single time
+slice of information provides an incomplete pic-
+ture of the situation. In the dynamic, ambiguous
+conditions characteristic of littoral operations, the
+rate and direction of change (data history) can
+help one better assess the threat and predict the
+future state of the situation. When the incoming
+information changes over time, the integration of
+information as it changes can help the user extract
+the message (Kirshenbaum, 1992). The DSS was
+designed to do precisely this: to facilitate the inte-
+gration process and present a synthesized picture
+of the situation to the user in a format that can be
+quickly assimilated. The variables, that are used to
+develop a threat assessment, can be divided into
+two classes: sensor information (raw or computer-
+processed information) and the contact's response,
+or lack of response, to actions taken by the team.
+These other actions, and the integration of the in-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      5Monterey, Ca.formation received via the contact's response orlack of response to them, are necessary to clarify
+the tactical picture.2.2  Situation AwarenessWhile recent increasing interest among re-searchers regarding the concept of situation
+awareness (SA) has generated a debate on the pre-
+cise definition of this term most researchers ac-
+knowledge the importance of the concept. In gen-
+eral, SA refers to the decision maker's moment-
+by-moment ability to monitor and understand the
+state of the complex system and its environment
+(Adams, Tenney, & Pew, 1995). These authors
+state the essential idea which is that when emer-
+gencies arise, the completeness and accuracy of
+the decision maker's SA are critical to the ability
+to make decisions, revise plans, and manage the
+system. Specific decision-making tasks included
+under SA include the ability to:  (1) maintain an
+accurate perception of the surrounding environ-
+ment (both internal and external to the ship); (2)
+identify problems and/or potential problems; (3)
+recognize a need for action; (4) note deviations in
+the mission; and (5) maintain awareness of tasks
+performed (Shrestha, Prince, Baker and Salas,
+1995). To maintain an accurate SA the decision
+maker should take into account both information
+that is available and that which can be activated
+from memory (Sarter and Woods, 1991).However, a difficulty arises as a result of theheavy workload imposed by this process. When
+the decision maker is faced with several concur-
+rent contacts of interest, all of which have numer-
+ous associated data items (i.e., as many as  a
+dozen), some or all of which may change over the
+course of the scenario (e.g., intelligence, active
+radar emitters, various kinematic parameters, etc.)
+memory load easily exceeds human capacity.
+Moreover, changing parameters may impart dif-
+ferent interpretations to what is occurring. In some
+cases the moment-to-moment attentional demands
+of a tactical situation are relentless and unforgiv-
+ing (such as, a terrorist aircraft directly inbound
+toward "own-ship" which can result in "task fixa-
+tion"), sometimes relevant background knowledge
+is unavoidably incomplete (such as, an unfamiliar
+aircraft), and sometimes the decision maker is al-ready thinking and working as hard as possible,even when there are no unanticipated events when
+there is a high contact density (Adams, et al,
+1995). These instances provide a few illustrations
+of situations that can degrade situation awareness.Complex information gathering and process-ing systems have been designed to aid the deci-
+sion-maker in the past. However, these systems
+often increase the decision-maker’s burden due to
+the inherent system complexity and the failure to
+design them in a way that they will fit the user's
+cognitive processing limitations. Often, these
+systems require operators to perform difficult
+cognitive tasks under heavy workloads. They must
+perceive, synthesize and determine the relevance
+of a continual stream of incoming information,
+often pertaining to several concurrent contacts,
+while projecting future anticipated events and
+making decisions regarding actions to be taken.
+Decision makers must assess, compare, and re-
+solve conflicting information, while making diffi-
+cult judgments, and remembering the status of
+critical contacts along with the contact's response
+to actions taken by the CIC team. These decision-
+making tasks are interleaved with other required
+tasks, such as keeping other team members in-
+formed (both on and off the ship). Furthermore,
+these complex tasks are performed under condi-
+tions where adverse environmental (noise, vibra-
+tion, temperature extremes, etc.) and internal
+stressors (boredom, fatigue, anxiety, and fear) are
+part of the environment.3  Decision Support PrinciplesA case has been made that previous generations ofdecision support systems, which focus primarily
+on solution optimization and base decision sup-
+port on normative models of human decision
+making, are less applicable than a DSS that par-
+allels the cognitive strategies used by domain ex-
+perts in situations characterized by time-
+constrained situations with uncertain and ambigu-
+ous data (Smith & Grossman, 1993).  These
+authors point out that rarely, if ever, were earlier
+tactical decision aids intended as psychological
+models of human cognitive behavior.  Instead,
+these aids performed "complex and burdensome
+calculations, reducing the work-load on personnel,
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      6Monterey, Ca.speeding up the dissemination of information, andproviding more time for command decision mak-
+ing" (Tolcott, 1991, p. 44).3.1  Feature Matching and Story GenerationWe have applied two of these new models ofhuman decision making—which parallel the cog-
+nitive strategies used by domain experts—to the
+design of a DSS for enhancing antiair warfare
+tactical decision making. These two models that
+people use in assessing a situation are feature
+matching and story generation. The feature
+matching model, described by Noble (1989), in-
+volves an organization of memory, or "schemas,"
+and information-processing where decision mak-
+ers use their previous experiences to assess a
+situation and identify promising actions. Incoming
+information is categorized, selected, edited, and
+organized on the basis of a person's general
+knowledge about a domain. Both story generation
+and feature matching occur under conditions
+where a large base of implication-rich, condition-
+ally dependent pieces of evidence must be evalu-
+ated before choosing an alternative from a set of
+prospective courses of action. The feature match-
+ing model applies a spatio-temporal dependence,
+whereas story generation is an example of causal
+dependence. According to the explanation-based
+model, decision makers construct a causal model
+to explain the available evidence (Pennington &
+Hastie, 1993). At the same time, the decision
+maker creates a set of alternatives from which an
+action will be chosen.  A decision is made when a
+story is successfully matched to an alternative in
+the choice set. Story generation occurs in complex
+situations where the decision maker may not have
+all the necessary information or when a series of
+facts may appear to contradict each other. The de-
+cision maker must then develop causal links be-
+tween these facts to produce a coherent picture of
+the situation (Klein, 1989; 1993).The explanation-based reasoning model isbased on research which found that jurors develop
+a narrative story to organize trial information
+where causal and intentional relations between
+events are central (Pennington & Hastie, 1992).
+Pennington & Hastie propose four certainty prin-
+ciples—coverage, coherence, uniqueness, andgoodness-of-fit—that govern (i) which story willbe accepted, (ii) which decision will be selected,and the (iii) confidence or degree of certainty withwhich a particular decision will be made. This or-
+ganization of the evidence by the decision maker
+is believed to facilitate evidence comprehension.
+A central component of this model is that the
+story the juror constructs determines the juror's
+decision.  This explanation-based decision process
+is employed when the body of evidence relevant
+to a decision is large, complex, and the implica-
+tions of its components are interdependent.Feature matching, also referred to as the rec-ognition-primed decision (RPD) model, "occurs
+when the decision maker recognizes the features
+of the present situation as similar or identical to
+those of a previous situation" (Kaempf & Militelo,
+1992, p. 6). An adequate match triggers recall of
+information learned about this type of situation:
+(a) plausible goals, (b) critical cues to be moni-
+tored, (c) expectations of what should happen, and
+(d) a course of action that worked in similar situa-
+tions. According to this recent approach, expert
+decision makers may rely on well-developed
+memory representations to guide decision making
+in new (but similar) situations. The RPD model of
+decision making fuses two processes—situation
+assessment and mental simulation (Klein, 1993).
+In the simplest case the situation is recognized as
+familiar or prototypical, using feature matching,
+and the obvious response is implemented. In a
+more complex case the decision maker performs a
+conscious evaluation of the response, using men-
+tal simulation to uncover problems prior to im-
+plementing the response. In the most complex
+case the evaluation reveals flaws requiring modi-
+fication, or the option is judged inadequate and
+rejected in favor of the next most typical reaction.3.2  Situation AssessmentIn general, the overall task of responding toantiair warfare scenarios consists of situation as-
+sessment ("what's going on") and course of action
+selection ("what to do about it"). Recent theories
+of decision making emphasize the importance of
+situation assessment for good decision making in
+naturalistic, event-driven situations. Moreover,
+they stress that decisions regarding actions to be
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      7Monterey, Ca.taken are a by-product of developing the situationawareness that precedes action selection. Klein
+(1989) has found that usually the situation itself
+either determines or constrains the response op-
+tions and that experienced decision makers make
+up to 90% of all decisions without considering
+alternatives. If the situation appears similar to one
+that the decision maker has previously experi-
+enced, the pattern will be recognized and the
+course of action is usually immediately obvious.
+On the other hand, if the situation does not seem
+familiar complex RPD will be involved where the
+decision maker adjusts the option after evaluating
+it.Additional evidence was found in the specifictask domain of interest to the TADMUS program
+which added support to these findings on the way
+real-world decision makers make decisions in the
+context of their normal jobs. Research was con-
+ducted to determine decision requirements for
+command-level decision makers in the combat
+information center (CIC) of an Aegis cruiser.
+Analysis of 14 incidents from actual problems re-
+vealed 183 decisions. Of these, 103 concerned
+situation assessments. Results obtained after ana-
+lysts coded these situation assessments indicated
+that decision makers arrived at approximately
+87% of their situation assessments through feature
+matching and the remaining 13% through story
+generation (Kaempf, Wolf, & Miller, 1993). The
+other eighty decisions that were identified, from
+analysis of the real-world incidents mentioned
+above, involved course of action selection. These
+course of action decisions served a variety of
+functions, although, relatively few were intended
+to end the incident. Twenty were intended as a
+final course of action decision; 14 were imple-
+mented to obtain more information, 22 to manage
+resources, and 24 to put themselves in a more fa-
+vorable tactical position. A recognition-based
+strategy was also used by decision makers to de-
+velop a course of action, accounting for 95% of
+the actions taken in the 14 incidents. The decision
+makers generated and compared multiple options
+in only 5% of the cases. In line with these find-
+ings, the TADMUS program has adopted the po-
+sition that decision aiding systems should assist in
+the decision making process, and focus on aidingthe situation assessment portion of the decision-making task.A DSS was developed to support decision-making processes which research has shown are
+used by decision makers in real-world settings
+(Hutchins, Kelly, & Morrison, 1996). Specifi-
+cally, the DSS parallels the strategies used by ex-
+perienced decision makers to perform situation
+assessment (Nobel, 1989; 1993). This approach to
+supporting the user's intuitive approach to dealing
+with dynamic decision-making situations should
+produce tools that are both more easily understood
+and used, and that more effectively "exploit the
+decision maker's knowledge and expertise that
+might facilitate adaptation to complex, novel
+situations" (Cohen, 1993, p. 265).4  Human-System Interaction PrinciplesThe vast majority of research on human-computerinteraction design has been devoted to character-
+istics of displays that impact human perception,
+such as symbol legibility or detectability, and on
+relatively simple cognitive functions such as
+memory tasks. Fewer efforts have been devoted to
+understanding the effects of the format and man-
+ner in which information is presented on more
+complex levels of human cognition such as deci-
+sion making. Consequently, principles that can be
+applied to the design of the interface between the
+user and a decision support system for the purpose
+of enhancing cognitive changing situations, are
+not available to any significant degree (ONT,
+1992).4.1  Graphic PresentationsSeveral advantages are offered by graphicpresentations over a text-based presentation for-
+mat (Larkin and Simon, 1987). Graphic presenta-
+tions should (1) reduce the amount of mental
+computation required to perform tasks; and (2)
+allow users to spend less time searching for
+needed information. Casner (1991) elaborated on
+these ideas and found that graphics allow users to
+substitute less demanding perceptual opera-tions
+for more complex logical operations. For exam-
+ple, determining a change in altitude (and the de-
+gree of change) is immediately apparent when the
+user glances at the track history module. (Note
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      8Monterey, Ca.that the words contact and track can be used inter-changeably. The reader is referred to Figure 1.)The objective for the track history module isto facilitate the contact identification process by
+providing information that is integrated in a way
+that supports a recognitional decision strategy.
+This module depicts a contact’s speed, altitude,
+course and range on a two-dimensional graphical
+display along with a geometric representation of
+both the contact’s weapon release envelope and
+own-ship’s weapons coverage. A large amount of
+parametric data is portrayed graphically for rapidassimilation by the user. The user can see, at aglance, a synthesized picture of the contact’s be-
+havior. Compare this rather simple perceptual op-
+eration with the more complex logical operation
+involved in current operational systems which re-
+quire the user to recall and subtract numerical val-
+ues for past and current altitudes.Graphics also allow users to omit steps thatare otherwise necessary when a task is performed
+without a graphic. An example of this advantage
+isFigure 1.  Decision Support System Display Modules.also illustrated in the track history module whichincludes templates indicating weapon's coverage
+for both the inbound contact and "own-ship." To
+determine whether the aircraft is within its
+weapon's launch range there is no need to recall
+the specific launch range values and then compare
+them with the aircraft's current range. Instead, the
+user can determine if the aircraft is within its
+launch range by a quick glance at the display.Graphics help users save time when searchingfor needed information when several related di-
+mensions of information are encoded in a single
+graphical object. This is accomplished by inte-
+grating the kinematic parameters of speed, course,
+altitude, bearing, and range for a contact. The user
+can see, at a glance, a synthesized picture of the
+contact's behavior.  Compare this process with
+reading, in a text-based format, the individual pa-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      9Monterey, Ca.rameters which need to be integrated by the userinto a coherent picture of the contact's behavior.5   Limited Cognitive Processing CapabilitiesSince there are limits to the cognitive processingcapability of humans, it is important for the sys-
+tem to provide the needed information in a format
+that best supports the user operating under dy-
+namic decision-making conditions. It may be the
+case that current systems are inadequate to support
+the cognitive processing demands required by
+certain littoral scenarios. For example, according
+to Gruner (1990, p. 41), the U.S.S. Vincennes of-
+ficers and system operators "could not make better
+decisions because they did not have time to con-
+firm or deny the information uncertainties pre-
+sented them." Gruner maintains that the rapid pace
+involved in these types of situations can exceed
+the capacity of the human to comprehend the
+rapid flow of information presented by complex
+systems. In the case of the Vincennes, the CIC
+team had three minutes and 40 seconds to make
+their decision. This includes the time required for
+the operators to perceive and interpret sensor data
+and for the commanding officer to make informed
+judgments from these data (Roberts & Dotterway,
+1995). The result of the human's limited cognitive
+processing capabilities is that the decision makers
+may fail to remember critical pieces of data,
+overlook stored information, draw hasty conclu-
+sions, and produce flawed answers.Evidence of the effects of limitations in mem-ory and shared attention capacity on human deci-
+sion making were found during baseline testing
+(Hutchins & Kowalski, 1994; Hutchins & Westra,
+1995) and during empirical evaluation of the DSS
+(Kelly, Hutchins, & Morrison, 1996; Kelly, Mor-
+rison, & Hutchins. 1996).Simon (1978, p. 273) states, "...the human in-formation processing system...operates almost en-
+tirely serially, one process at a time, rather than in
+parallel fashion. This seriality is reflected in the
+narrowness of its momentary focus of attention."
+However, the AAW problem forces the decision
+maker to operate in a parallel processing mode
+when several contacts demand attention at the
+same time. The requirement to monitor and
+maintain an accurate SA for these concurrentcontacts, over the course of the evolving situation,imposes an additional load of strategically man-
+aging the overall situation. Several researchers
+have argued that "managing the attentional and
+conceptual processes that permit cogent SA in-
+volves significant cognitive resources" (Adams, et
+al, 1995, p. 91; Endsley, 1988). The tasks of pri-
+oritizing contacts and the associated actions to be
+taken by the team, updating the status of critical
+contacts, responding to the other requisite tasks in
+the queue and, more generally, of strategically
+managing the workload of current multitask sys-
+tems under dynamically changing scenarios can
+place an unrealistic cognitive load on the decision
+maker.A major advantage offered by the experimen-tal DSS is that it should "buy time" for the user by
+(1) performing many of the cognitive processing
+tasks for the user and (2) by presenting informa-
+tion in graphic format. The DSS will synthesize
+much of the information used to develop situation
+awareness and present a coherent picture of the
+situation to the user. This integrated picture will
+be portrayed graphically—rather than in the cur-
+rent text-based format—which should further re-
+duce the amount of time required to assimilate
+this information. By performing several informa-
+tion processing steps for the decision maker the
+decision maker's limited cognitive resources can
+be used for the types of decisions which require
+human abilities (e.g., the decision on whether to
+engage).5.1  Working Memory RequirementsAn essential information processing step re-quired by this task—and one which levies a heavy
+load on working memory—involves integrating
+kinematic and sensor variables and maintaining an
+awareness of changes in these variables over time.
+Changes in a contact's behavior such as, decreas-
+ing altitude, increasing speed, changes in elec-
+tronic emissions, etc., can provide key indicators
+of possible hostile intent. With current systems,
+the decision maker receives numerous reports
+from CIC team members who provide various
+pieces of the overall tactical picture (such as,
+kinematic parameter values, active electronic
+emitter identifications, and behavioral responses
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      10Monterey, Ca.of the contact in response to queries by the team)regarding a particular contact. Some of this infor-
+mation is also displayed in a text-based format for
+the user when a contact is "hooked" (that is, se-
+lected for display) by the decision maker. How-
+ever, to recognize a change in certain variables,
+current systems require the user to retain parame-
+ter values in short-term memory in order to recog-
+nize a change in the parameter, such as altitude.When the decision maker is monitoring sev-eral concurrent contacts (such as, cycling through
+three or four contacts in a 1-minute period) human
+working memory capabilities may quickly be sur-
+passed. To detect a change in a critical parameter
+value, the decision maker must maintain the pa-
+rameter values for the contacts of interest in
+working memory as he or she cycles between sev-
+eral contacts. For example, the decision maker
+must be able to recall that contact 7022 was at
+14,000 ft. altitude one minute ago, and then sub-
+tract the current altitude value of 10,000 ft., which
+will then indicate the aircraft is in a rapid descent.
+The DSS was developed to aid the decision maker
+by performing several of these cognitive process-
+ing tasks, thus, reducing the cognitive load for the
+user. By presenting the synthesized picture of the
+contact's behavior over time, through the use of
+graphical displays,  critical changes should be
+immediately apparent to the user.A second memory-intensive task involvesmaintaining, in working memory, a current list of
+actions taken by team members, the contact's re-
+sponse to these actions taken by the CIC team,
+and pending actions. Research has established that
+"memory is limited and that list maintenance is
+effortful and fallible—more so if the list must be
+ordered and still more if the membership of the
+list must be dynamically reordered and modified
+during retention" (Bower, 1970, as cited in Ad-
+ams, et al, 1995, p. 91).  The DSS should reduce
+the cognitive effort required for distributing atten-
+tion among the many contacts to be attended to
+and actions that are required. Working memory
+requirements should be reduced by having the
+DSS act as an intelligent "assistant," reminding
+the user regarding what actions are to be taken and
+when the actions are to be taken.A third way the DSS will reduce memory andinformation processing requirements is by dis-
+playing templates depicting weapons' envelopes
+for both the inbound contact and "own-ship." This
+should facilitate critical comparisons and judg-
+ments regarding timing of actions. During a sce-
+nario decision makers have to either rely on mem-
+ory to recall the launch range for various weapons
+or query a team member for this information. Both
+of these methods waste limited resources. The
+high workload and high tempo characteristic of
+littoral scenarios produce a stress-ful decision-
+making environment. The phenom-enon that in-
+creasing stress leads to decreasing working mem-
+ory is well documented (e.g., Hockey, 1986). The
+latter method for obtaining the desired informa-
+tion wastes limited resources by increasing the
+communications load and requiring more time to
+wait for a team members' response to the queryUnder these high-tempo and high workloadconditions human memory and attentional re-
+sources can easily be surpassed. Several cogni-
+tively resource intensive information processing
+steps are eliminated for the human decision maker
+by having them performed by the DSS. We pre-
+dict that the decision support tools will reduce the
+cognitive workload imposed on the decision
+maker in the following three ways: (1) by reduc-
+ing the amount of information processing to be
+performed, (2) reducing working memory re-
+quirements, and (3) assisting the user in allocating
+limited attentional resources.5.2  Reducing Human ErrorThe study of human cognitive processes andrelated error mechanisms has gained rapidly in-
+creasing interest in the past decade. Rasmussen
+(1987) argues that the emphasis in attempting to
+understand human errors must shift from tasks to
+the human-task mismatch.  For example, Gruner
+(p. 39), in discussing the Vincennes incident,
+maintains that "the system was poorly suited for
+use by human beings during rapid military action."
+He ascribes this lack of suitability to a human-
+machine mismatch between the rate of data flow
+possible with modern computer systems that can
+process and display information at phenomenal
+data rates and the "comprehension capability of
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      11Monterey, Ca.users which has remained almost static for thou-sands of years." This causal approach to under-
+standing human error is based on the premise that
+errors are rarely random and can be traced to
+causes and contributing factors. Once these con-
+tributing factors are identified they can be miti-
+gated.The impact and vulnerability of systems andhuman interfaces, because of incompatibilities
+between the way people perceive, think, and act,
+are documented in the popular and technical lit-
+erature (Buck, 1989; Casey, 1993; Norman, 1988;
+Perrow, 1984; Wilson & Zarakas, 1978). Newly
+developed systems will succeed or fail based on
+our ability to minimize these incompat-ibilities
+between the characteristics of the things we create
+and the way we use them. There are many well-
+documented instances of critical systems or pa-
+rameter changes going unnoticed or unheeded be-
+cause the operating procedures, or the human ma-
+chine interface, provided no historical trace. For
+example, an unnoticed increase in altitude con-
+tributed to the shoot down of the Iranian airbus by
+a U.S. Navy ship—when the team mistakenly be-
+lieved the aircraft to be descending—because
+there was no historical trace to make the aircraft's
+actual increasing altitude apparent (Dotterway,1992).  Five personnel in the U.S.S. Vincennes's
+combat information center, all viewing separate
+displays, reported the aircraft as descending while
+the Aegis data tapes later revealed a flight pattern
+of ascent (Roberts & Dotterway, 1995). One of the
+official investiga-tions of this incident, the Fo-
+garty Report (1988, p. 45), states that "stress, task
+fixation, and an unconscious distortion of data
+may have played a major role in this incident." A
+panel of five psychologists from the American
+Psychological Association who testified before
+Congress concluded that there were "predictable
+failings of human judgment under intense stress
+compounded by complex technology [which]
+clearly contributed to the accidental shooting of
+Iranian airliner Flight 655" (APA, p. 4).It is generally accepted that between 60-80percent of the accidents and malfunctions in
+transportation, manufacturing, process control,
+weapon, and other systems are attributable to hu-
+man error (Senders & Moray, 1991; Van Cott,1993; Weiner, 1994). Reducing tactical decisionmaking errors is one goal of the TADMUS pro-
+gram.  The following section presents a brief re-
+view of an experiment conducted to develop a
+baseline on tactical decision making performance
+in response to fairly stressful scenarios. A com-
+panion paper (Hutchins, Kelly, & Morrison, 1996)
+describes the experimental DSS modules and the
+way they are hypothesized to enhance tactical de-
+cision making performance.6   TADMUS Baseline ExperimentEarly research involved data collection in the De-cision-Making Evaluation Facility for Tactical
+Teams (DEFTT) Laboratory using simulated ex-
+isting shipboard displays to establish a baseline on
+decision-making performance. The purpose of this
+effort was to document baseline decision-making
+performance for experienced naval officers. Dur-
+ing the baseline phase of testing, a detailed under-
+standing was developed of the cognitive processes
+underlying the various tasks involved in situation
+assessment—and where the bottlenecks occur.
+This understanding was then used to design the
+way the information is presented to the user in or-
+der to facilitate performance of the required tasks.6.1  SubjectsThis study focused on the command-level de-cision makers of an antiair warfare team on an
+Aegis cruiser—the commanding officer and the
+tactical action officer. Subjects in the study con-
+sisted of six commanding officer/tactical action
+officer teams drawn from twelve active duty Na-
+val personnel; some were from training com-
+mands while others were from operational com-
+mands aboard ship or assigned to group staffs.6.2  ProcedureData were collected in the  DEFTT Labora-tory, a six-station test-bed environment that
+simulates console positions in a Navy Aegis
+cruiser combat information center. (For a detailed
+description of the DEFTT Laboratory see Hutch-
+ins, 1996.) Four stations were filled by confeder-
+ates (active duty Navy personnel) who play antiair
+warfare support-team member roles. These roles
+included the antiair warfare coordinator, identifi-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      12Monterey, Ca.cation supervisor, tactical information coordina-tor, and electronic warfare supervisor. After ap-
+proximately 1 1/2 hours of orientation to the labo-
+ratory and training in the use of the computer con-
+soles the subjects engaged in four scenarios. The
+scenarios were each about 25 minutes in length
+and contained between 11 and 14 contacts of in-
+terest per scenario, in addition to numerous back-
+ground contacts.6.3  Treatment of DataTeam communications were recorded on amultichannel audio recorder; these included all
+intra-team exchanges, as well as all communica-
+tions with simulated off-ship personnel. Audio
+tapes were used to produce verbatim, time-
+stamped transcripts of all team communications.
+A modified version of the TapRoot® Incident In-
+vestigation System (Paradies, 1991; Paradies and
+Unger, 1991) was then applied to identify errors.
+The objective was to identify tactically significant
+errors committed during the scenario. Tactically
+significant errors were defined as those errors that
+may lead to loss of life or significant political em-
+barrassment. The following criteria were used for
+counting an error as tactically significant:  (1) loss
+of situation awareness, (2) failure to take defen-
+sive action when within the weapon's range of an
+approaching contact, or (3) a violation of rules of
+engagement (ROE). Video recordings were made
+of the commanding officer and tactical action of-
+ficer computer screens. Detailed analyses of all
+audio and video recordings were conducted. (For
+a more detailed coverage of the methodology and
+results see Hutchins and Westra, in preparation).6.4  ResultsThe complex, time-constrained, decision-making situations embodied in the experimental
+scenarios resulted in a large number of decision
+errors. The mean number of tactically significant
+errors documented across six teams and four sce-
+narios was 14; the number of errors ranged from
+nine to twenty-two. The standard deviation was
+3.7.  Subjects performed an average of 50% of the
+required behaviors as specified in the rules of en-
+gagement. The ordinal agreement between three
+raters (navy subject matter experts) on error countranks from TapRoot® analyses was computed.Results showed a high degree of agreement with
+the Kendall's W of .93 indicating that 93% of the
+possible rank variance is accounted for.6.4.1  Decision-Making ErrorsDetailed examinations of the informationprocessing sequences performed during tactical
+decision making have revealed a variety of errors.
+On average, subjects failed to take required ac-
+tions, about half of the time. Explanations based
+in the cognitive psychology literature have been
+pursued, as a major goal of the TADMUS pro-
+gram is to develop a DSS based on an under-
+standing of the way in which human decision
+makers actually process information under rapidly
+evolving situations.The majority of documented errors involvederrors of omission, that is, "failure to take defen-
+sive measures” and "failure to adhere to ROE.”
+Failure to take defensive measures included fail-
+ure to take actions to defend own-ship when an
+approaching aircraft had reached its weapon's re-
+lease range. An example involved a case where
+two contacts were within the specified ROE limit,
+yet no action had been taken. The types of actions
+included in the “failure to adhere to ROE” cate-
+gory include failure to take action regarding the
+items listed and defined below: (a) issuing warn-ings is part of the usual identification process andinvolves three levels of warnings with increasing
+levels of urgency; (b) establish friendly force cri-teria refers to establishing a plan with otherfriendly ships in the area to coordinate how they
+will respond to potential threats; (c) changes inkinematics/ identification friend or foe—subjectsare expected to notice significant kinematic
+changes and/or identification friend or foe pa-
+rameter changes; and (d) other identification pro-cedures includes actions such as illuminating withfire control radar.The other major category of tactically signifi-cant error involved “loss of SA.” Loss of SA er-
+rors were grouped under errors of commission and
+errors of omission and then further categorized
+into subgroups. Fifty-five percent of the loss of
+SA errors involved taking the wrong action (error
+of commission) while 45% of the errors involved
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      13Monterey, Ca.failing to take some required action (error ofomission). Error categories included under errors
+of commission involved incorrectly engaging a
+track (3%), incorrectly warning a track (29%),
+other incorrect actions (16%), and incorrect re-
+porting (7%). The two instances of incorrectly en-
+gaging an aircraft, which were F-1 Mirage aircraft,
+were considered errors because the decision
+maker failed to take certain actions prior to en-
+gaging—not necessarily because the aircraft
+should not have been engaged. The actions that
+the decision makers failed to take involved ascer-
+tain-ing the identification of the aircraft for one
+case and failure to warn and illuminate the aircraft
+prior to engaging for the second case. Most in-
+stances of incorrectly issuing warnings to the air-
+craft involved issuing the warning when the air-
+craft was within its territorial airspace (that is, in-
+side the 12 nautical mile limit which is interna-
+tionally recognized as under control of that nation)
+or issuing a warning at a level different from what
+was required. Other incorrect actions included il-
+luminating the aircraft, “locking up” with radar, or
+ordering the aircraft to divert when the aircraft
+was still within its territorial airspace. Incorrect
+reporting involved inaccurate reports on the status
+of the tactical situation (such as, indicating to the
+battle group commander that certain actions had
+been taken when they had not, misidentification of
+an aircraft, or omitting critical tracks from a re-
+port).Errors of omission categorized under the “lossof SA” category included: (a) failure to identify or
+attend to a contact; (b) failure to take action (e.g.,
+to issue “hold-fire” when a contact turned out-
+bound); (c) failure to recognize a threat (e.g.,
+designating an aircraft as a non-threat because it
+had passed its closet point-of-approach, yet it was
+still within missile-launch range); (d) instances of
+confusion or forgetting (e.g., forgetting or ignor-
+ing critical data, forgetting whether or not it had
+been warned, illuminated, or “locked-on,” or for-
+getting the aircraft's response, or lack of response
+to these actions, forgetting the status of a contact,
+and confusing contacts); (e) misperception of data
+(e.g., reporting a contact as turning outbound
+when it is still inbound); (f) unclear communica-
+tion (issuing vague orders regarding actions to betaken by team member, such as, failure to specifywhich weapon system is to be used or which con-
+tact is to be engaged).6.4.2  Cognitive explanationsThe cause of failures to take required actionsis, in many cases, attributed to the extremely high
+task demands levied on the decision maker by the
+scenario and the human decision-maker's limited
+attentional resources. Many cases are also attrib-
+uted to working memory limitations.  Maintaining
+an awareness of the status of each contact and the
+status of many actions to be taken by the antiair
+warfare team—which actions have been taken and
+what the contact's response to the action was—
+severely taxes the decision maker's working
+memory. The high workload entailed in the
+scenarios produces a highly time-compressed
+decision making situation. This time-compressed
+decision making situation—where attentional
+resources and working memory capacity are
+limited—do not allow the decision maker to
+maintain accurate SA for all tracks at any given
+time. We anticipate that the decision support
+modules in the DSS will mitigate these types of
+errors.Human information processing capabilities arenot well suited to dealing with a "multiplicity of
+simultaneous and disjointed tasks. Thoughtful at-
+tention is modular:  People can consciously think
+about only one thing at a time" (Adams, et al,
+1995, p. 92). As a result, they do not handle inter-
+ruptions very well. Research indicates that when
+an operator is faced with as few as two tasks that
+consist of merely the detection or recognition of
+simple signals, a cost may be incurred in terms of
+a significant loss in sensitivity or time that can be
+allocated to either by the requirement to divide or
+switch attention between them (Broadbent, 1957;
+Schneider and Detweiler, 1988; Swets, 1984).The memory demands of managing complex,multi-task situations can easily surpass human
+limitations. The decision maker must not forget
+any of the contacts or tasks requiring action. In
+addition to remembering all the tasks needing at-
+tention, however, are the complexities entailed in
+keeping track of the data and substeps associated
+with each contact and prior action. The aviation
+literature provides many examples of incidents
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      14Monterey, Ca.with explanations similar to the root causes forerrors that were found in the TADMUS program.
+One category includes the potentially disastrous
+effects of interruptions in the task for air traffic
+controllers and pilots. Similarly, in the AAW en-
+vironment, momentary intervening attention to
+another task or contact, or an interruption in a
+procedure can leave the procedure, or processing
+of a contact incomplete with potentially cata-
+strophic results.A fairly consistent pattern of tactical decision-making errors was documented from data col-
+lected during the baseline data collection period.
+The root causes of these errors were traced to
+cognitive mechanisms such as limited attentional
+resources and working memory limitations. By
+developing an understanding of the pattern and
+types of errors most frequently observed in this
+task domain we hope to provide a DSS which will
+mitigate these errors.7  DiscussionFailure to take appropriate actions may be ex-plained by the limited resource capacity of human
+memory. In these scenarios a large number of
+contacts are monitored for changes in any of sev-
+eral key parameters. Three modules in the DSS
+are hypothesized to assist with recognizing a
+problem and taking the appropriate actions: track
+history; response manager; and the track priority
+list and alerts.Features offered by the DSS to address errorsattributed to limited attentional resources include
+focusing attention on (1) high priority contacts
+(i.e., track priority list and alerts), as well as on (2)
+missing data (e.g., basis for assessment), and (3)
+enabling the decision maker to use more data than
+is typically used in current systems (e.g., track
+history, comparison to norms). Current systems
+require the user to retain previous contact data in
+memory to compare with current values for criti-
+cal parameters. Current systems also require the
+user to rely on recall of vast amounts of informa-
+tion from training and experience. Presenting all
+known data on a contact in a synthesized way
+should reduce working-memory requirements and
+facilitate recognition. Additional potential per-
+formance enhancement features, offered by theDSS, include displaying the complete kinematiccontact history, presenting graphic displays of lo-
+cation and trends, highlighting missing data, pro-
+viding alerts, and providing assessments of cur-
+rent contact identity that go beyond what existing
+systems currently present.Focusing the user's attention on trend andhistory data should decrease the cognitive work-
+load imposed by these scenarios where many
+contacts must be identified and responded to un-
+der severe time constraints. Similarly, delineating
+trend and history data can assist in the identifica-
+tion of a contact where noticing changes in critical
+parameters is essential. Presentation of trend and
+history data, as well as threat assessment and
+comparison to norms, should also mitigate cogni-
+tive "tunnel vision" effects where the decision
+maker attends to a smaller number of cues when
+under stress.The notion of time is an important character-istic of situation awareness (Harwood, Barnett,
+and Wickens, 1988). The past is critical to under-
+standing the present, and both past and present
+information must be used to predict future events
+(Shrestha, et al, 1995). Endsley (1988) referred to
+the "projection of their (perceived elements) status
+in the near future" when discussing situational
+awareness. However, Endsley also noted the task
+of attending to incoming information and subse-
+quently predicting future events places a heavy
+load on working memory. Several decision sup-
+port modules were developed to assist the user in
+remaining aware of the contact's history and
+changes over time. Remembering which actions
+are to be taken at what time levies an additional
+burden by placing a heavy load on working mem-
+ory. A secondary time savings should be achieved
+by the DSS acting as an intelligent "advisor," that
+is, by assisting the decision maker in knowing
+what actions to take, when to take them, and
+which actions have already been taken. A tertiary
+time savings can be achieved by including a tem-
+plate depicting the weapons' release ranges so the
+decision maker does not need to rely on fallible
+human memory or query a team member regard-
+ing weapons ranges. By graphically depicting a
+synthesized view of a contact's kinematic history,
+with the focus on changes in the contact's behav-
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      15Monterey, Ca.ior over time, along with the contact's weaponsenvelope in relation to both own-ship's radar and
+weapon coverage, information processing time
+can be saved for the decision maker.8  ConclusionsThe research reported here focuses on developinga  DSS which reflects the natural decision-making
+strategies of humans. Presenting synthesized in-
+formation in the form of graphic presentations is
+expected to reduce the cognitive processing load
+for the decision maker when performing situation
+assessment. The intention is to aid the decision
+maker by providing information in a way that will
+minimize the need to maintain information in
+working memory, reduce information processing
+demands, help focus attentional resources on the
+highest priority contacts, remind the user of ac-
+tions which need to be taken, help make decisions
+under stress, and support higher levels of situation
+awareness.Decision support systems require that the hu-man's strengths be used in synergy with the ad-
+vantages offered by the DSS. Limitations associ-
+ated with the current generation of automated de-
+cision aids include the idea that (1) they cannot
+adequately capture the expertise developed by ex-
+perience over time and (2) since all contingencies
+cannot be anticipated, the expert's abilities to use
+intuition is indispensable (Mosier, in press). Mo-
+sier's review of the limitations of automated deci-
+sion systems delineates the characteristics of hu-
+man expertise that surpass the capabilities of
+automated systems. These include the human ca-
+pacity for creativity, adaptability, the ability to
+incorporate experience, the presence of a broad
+focus, analogical reasoning, and commonsense
+knowledge. The goal for the DSS is to capitalize
+on the strengths of the human along with the ad-
+vantages provided by the decision support system.9  Testing the DSSThe prototype DSS display modules are currentlybeing empirically evaluated in the simulated tacti-
+cal environment provided in the DEFTT Labora-
+tory. Experienced naval decision makers engage
+in experimental scenarios with and without access
+to the DSS display. The various decision supportmodules will be tested individually and in combi-nation in future experiments. Data on reduction of
+errors, improvements in users' situation awareness
+scores, changes in communication patterns, and
+subjective responses to the decision support sys-
+tem will be collected.10   Future ResearchWhile tools based on both the RPD and explana-tion-based reasoning models of decision making
+are included in the DSS there is no direct connec-
+tion between the two. Research is currently being
+conducted to extend schema theory to dynamic
+decision-making situations. This involves devel-
+oping and testing a hybrid model of cognitive be-
+havior in decision making to incorporate both
+types of knowledge, i.e., feature matching and
+story generation, as elements of the same schema
+model of naturalistic decision making (Smith &
+Marshall, in press). Schema theory as described
+by these authors, offers a context for integrating
+these two models which have typically been
+viewed as separate entities.AcknowledgmentsThe authors gratefully acknowledge the assistanceof Steve Francis, Brent Hardy, C.C. Johnson, Pat
+Kelly, Ron Moore, Connie O’Leary, Pat Marvel,
+Mike Quinn, and Will Rogers in data collection,
+interpretation, developing the DSS, and in con-
+ducting this research.ReferencesAdams, M. J., Tenney, Y. J., & Pew, R. W.(1995). Situation Awareness and the Cognitive
+Management of Complex Systems. HumanFactors, 37(1), 85-104.American Psychological Association, ScienceAgenda, Fall, 1988.Bower, G. H.  (1970).  Organizational Factors inMemory. Cognitive Psychology, 1, 18-46.Broadbent, D. E. (1957).  A mechanical model forhuman attention and immediate memory. Psy-chological Review, 64, 205-215.Buck, L.  (1989).  Human error at sea.  HumanFactors Bulletin, September, 32 (9), 12.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      16Monterey, Ca.Casey, S.  (1993).  Set Phasers on Stun and OtherTrue Tales of Design, Technology, and Hu-
+man Error.  Aegean Publishing Company,Santa Barbara, CA.Casner, S. (1991).  Task-Analytic Design ofGraphic Presentations. Technical Report  AIP-145.  Departments of Computer Sci-ence and Psychology, Carnegie Mellon Uni-
+versity and Learning Research and Devel-
+opment Center, University of Pittsburgh, PA.Cohen, M. S.  (1993).  The Bottom Line:  Natu-ralistic Decision Aiding.  In G. Klein, J. Ora-
+sanu, R. Calderwood, & C. E. Zsambok (Eds.)Decision Making in Action:  Models andMethods (pp. 138-147).  Ablex PublishingCorporation, New Jersey.Dawes, R. M.  (1971).  A case study of graduateadmissions:  Application of three principles of
+human decision making. American Psycholo-gist.  34, 571-582.Dotterway, K. A. (1992).  Systematic Analysis ofComplex Dynamic Systems:  The case of the
+USS Vincennes.  Unpublished master's thesis,Naval Postgraduate School, Monterey, CA.Endsley, M. R. (1988).  Design and evaluation forsituational awareness enhancement.  Pro-ceedings of the Human Factors 32nd Annual
+Meeting, 97-101.Federico, P. A.  (1995).  Expert and novice recog-nition of similar situations.  Human Factors,37(1), 105-122.Formal Investigation into the Circumstances Sur-rounding the Downing of a Commercial Air-
+line by the USS Vincennes, by Rear Admiral
+William M. Fogerty, USN, 28 July 1988, pp.
+4, 6.  Hereafter referred to as the Fogerty Re-
+port.Gruner, W. P. No Time For Decision Making.(1990, November). U.S. Naval Institute Pro-ceedings, 39-41.Harwood, K., Barnett, B., and Wickens, C.(1988). Situational awareness:  A conceptual
+and methodological framework. Proceedingsof the Symposium Psychology in the Depart-
+ment of Defense.Hockey, G. R. (1986).  Changes in Operator Effi-ciency as a Function of Environmental Stress,
+Fatigue, and Circadian Rhythms.  In:  K. R.Boff, L. Kaufman, J. P. & Thomas (Eds.):Handbook of Perception and Human Per-formance.  Wiley, New York.Hutchins, S. G. (in press). Decision-MakingEvaluation Facility for Tactical Teams. Naval
+Command, Control, and Ocean Surveillance
+Center, RDT&E Division Technical Report, in
+press, San Diego, CA.Hutchins, S. G., Kelly, R. T. and Morrison, J. G.(in press). Decision Support for Tactical Deci-
+sion Making Under Stress.  Proceedings of theSecond International Symposium on Com-
+mand and Control Research and Technology.June 1996, Monterey, CA.Hutchins, S. G. and Kowalski, J. T.  (1993).  Tac-tical Decision Making Under Stress:  Prelimi-
+nary Results and Lessons Learned.  Proceed-ings of the 10th Annual Conference on Com-
+mand and Control Decision Aids.  June 1993,Washington, D. C.Hutchins, S. G. and Rummel, B. K. (1995).  ADecision Support System for Tactical Deci-
+sion Making Under Stress.  Proceedings of the
+First International Conference on Command
+and Control Research and Technology.  June
+1995, Washington, D. C.Hutchins, S. G. and Westra, D. P. (1995). Patternsof Errors Shown by Experienced Navy Com-
+bat Information Center Teams.  Proceedingsof the 39th Annual Meeting of the Human
+Factors and Ergonomics Society, San Diego,CA. October 1995.Hutchins, S. G. and Westra, D. P. (in preparation).TADMUS Baseline Experimental Results.
+Naval Command, Control, and Ocean Sur-
+veillance Center, RDT&E Division Technical
+Report, in preparation, San Diego, CA.Kaempf, G. L. and Militelo, L. G., (1992). Deci-sion Making in Emergencies, First Offshore
+Installation Management Conference: Emer-
+gency Command Responsibilities Collected
+Papers, Aberdeen, Scotland.Kaempf, G. L. , Wolf, S. and Miller, T. E.  (1993).Decision Making in the Aegis Combat Infor-
+mation Center. In Proceedings of the Human
+Factors and Ergonomics Society 37th Annual
+Meeting (pp. 1107-1111).  Santa Monica, CA:
+Human Factors and Ergonomics Society.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      17Monterey, Ca.Kelly, R. T. , Hutchins, S. G., and Morrison, J. G.(1996). Decision Processes and Team  Com-
+munications with a Decision Support System.Proceedings of the Second International Sym-posium on Command and Control Research
+and Technology. June 1996, Monterey, CA.Kelly, R. T. , Morrison, J. G., and  Hutchins, S.G., (in press).  Impact of Naturalistic Decision
+Support on Tactical Situation Awareness.Proceedings of the 40th Annual Meeting of theHuman Factors and Ergonomics Society,Philadelphia, PA. September 1996.Kirshenbaum, S. S. (1992).  Influence of Experi-ence on Information-Gathering Strategies.Journal of Applied Psychology, 77, 343-352.Klein, G. A. (1989). Recognition-Primed Decisions. In W. R. Rouse (Ed.)  Advancesin Man-Machine Systems Research (pp. 47 - 92), Vol. 5. JAI Press, Inc.Klein, G. A. (1993).  A Recognition-Primed Deci-sion (RPD) Model of Rapid Decision Making.
+In G. A. Klein, J. Orasanu, R. Calderwood, &
+C. E. Zsambok (Eds.)  Decision Making inAction:  Models and Methods (pp. 138-147).Ablex Publishing Corporation, New Jersey.Larkin,  J. H. and Simon, H. A. (1987). Why adiagram is (sometimes) worth 10,000 words.Cognitive Science, 1, 65-99.Mosier, (in press). Myths associated with Auto-mated Decision Aids. In G. A. Klein, & C. E.
+Zsambok (Eds.) Advances in Naturalistic De-cision Making:  Research and Applications,Hillsdale, NJ:  Erlbaum.Mundy, C. E., Jr.  (1994).  Thunder and Light-ning:  Joint Littoral Warfare, Joint ForceQuarterly,  4, Spring, 45-50.Noble, D.  (1989).  Application of theory of cog-nition to situation assessment.  Vienna, VA:Engineering Research Associates.Noble, D.  (1993).  A Model to Support Devel-opment of Situation Assessment  Aids.  n G.
+A. Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.)  Decision Making in Action:Models and Methods (pp. 287-305). AblexPublishing Corporation, New Jersey.Norman, D. A.  (1988).  The Psychology of Eve-ryday Things.  Basic Books, Inc. New York.Norman, D. A. and Bobrow, D. G.  (1975).  Onthe data-limited and resources-limited proc-
+esses.  Cognitive Psychology, 7, 44-64.Office of Naval Technology.  (1992).  FY 1993Program Plan for Tactical Decision-Making
+Under Stress, Arlington, VA: July 1992.Paradies, M.  (1991).  Root Cause Analysis andHuman Factors. Human Factors Society Bul-letin, 34(8), 1-4.Paradies, M. & Unger, L. (1991).  TapRoot Inci-dent Investigation System Manual. Volumes1-7. System Improvements, Inc. Knoxville,
+TN.Pennington, N.  & Hastie, R (1992). Explainingthe Evidence: Tests of the Story Model of De-
+cision Making. Journal of Personality and So-cial Psychology, Vol. No. 2, 189-206.Pennington, N. & Hastie, R   (1993).  A theory ofExplanation-Based Decision Making.  In G. A.
+Klein, J. Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.) Decision Making in Action:Models and Methods (pp. 188-201).  AblexPublishing Corporation, New Jersey.Perrow, C.  (1984).  Normal Accidents:  Livingwith High Risk Technologies.  New York, Ba-sic Books, Inc.Rasmussen, J.  (1986). Information Processingand Human-Machine Interaction.  In A. P.
+Sage (Ed.) Series Volume 12, North-Holland,
+Amsterdam.Roberts, N. C. & Dotterway, K. A.  (1995). TheVincennes Incident:  Another Player on theStage?  Defense Analysis Vol 11, No. 1, pp.31-45.Salthouse, T. A. (1992).  Cognition and Context.Science, 257, 982-983.Sarter, N. B. and Woods,  D. D. (1991).  Situationawareness:  A critical but ill-defined phe-
+nomenon.  The International Journal of Avia-tion Psychology, 1(1), 45-57.Schneider, W., and Detweiler, M.  (1988). Therole of practice in dual-task performance:
+Toward workload modeling in a connectionist/
+control architecture. Human Factors, 30, 539-566.Senders, J.  W. & Moray, N. P.  (1991).  HumanError:  Cause, Prediction, and Reduction.Lawrence Erlbaum Associates, New Jersey.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      18Monterey, Ca.Shrestha, L. B., Prince, C., Baker, D. P. and Salas,E., (1995). Understanding Situation Aware-
+ness: Concepts, Methods, and Training. Hu-man/Technology Interaction in Complex Sys-
+tems.  Vol 7, W. B. Rouse (Ed.).  San Fran-cisco:  JAI Press.Simon, H. A.  (1978).   Information-processingtheory of human problem solving.  In W. K.
+Estes (Ed.), Handbook of Learning and Cog-nitive Processes, Vol 5, Human InformationProcessing, New York: Wiley.Smith, D. E. and Grossman, J. D. (1993).  Under-standing and Aiding Decision Making in
+Time-Constrained and Ambiguous Situations.Unpublished manuscript.Smith, D. E. and Marshall, S.  (in press).  Apply-ing Hybrid Models of Cognition in Decision
+Aids. In G. A. Klein, & C. E. Zsambok (Eds.)Advances in Naturalistic Decision Making:Research and Applications, Hillsdale, NJ:Erlbaum.Swets, J. A. (1984).  Mathematical models of at-tention. In R. Parasuraman and R. Davies
+(Eds.), Varieties of Attention (pp. 183-242).New York: Academic.Tolcott, M. A. (1991). Understanding and AidingMilitary Decisions. Office of Naval ResearchEuropean Office. 27th International Applied
+Psychology Symposium, Stockholm, Sweden,
+June 1991.Wilson, G. L. & Zarakas, P.  (1978).  Anatomy ofa blackout.  IEEE Spectrum, February, 339-346.
+Paper presented to the Second International Command and Control Research and Technology Symposium,                      19Monterey, Ca.The requirement to interleave a mul-tiplicity of tasks—although not necessary
+an ongoing characteristic of shipboard
+scenarios—represents the type of situation
+where providing decision support may
+make the critical difference in the outcome
+for a scenario. For example, during the
+experimental scenarios the decision mak-
+ers may have to perform the following:  monitor ship location (relative toother ships and objects in the vicinity)  monitor and apply rules of en-gagement to all applicable tracks in the
+local        operating area
+
+  receive and send radio messagesto the battle group commander and other        operating units in the area
+  monitor tracks on the Aegis dis-play system and maintain situation 
+                   awareness for all contacts of
+interest  monitor performance of actionstaken by team members to assess the       situation
+  monitor the tactical action offi-cer's/commanding officer's performance  maintain communications withCIC team members regarding their       assessment of tracks and vari-ous actions taken
+In broad perspective, although teammembers spend much of their time in rou-
+tine activities, a number of different atten-
+tionally demanding, knowledge-intensive,
+and procedurally complex tasks may de-
+mand attention at any moment. Each of
+these tasks is usually triggered by a
+stimulus event, such as a communication
+from a team member or an alert, and, in
+order to obtain proper interpretation, mayrequire additional information-seeking be-havior.  The cognitive challenge of select-
+ing and interpreting information to main-
+tain and revise one's SA is inherently
+complex. (Jager, Tenney, and Pew, 1995
+HF) Problems arise when, in the dynamic
+and multidimensional environments of
+some littoral antiair warfare scenarios, the
+situation-critical data become more time-
+compressed or ambiguous than humans
+can handle within the inherent time con-
+straints of the evolving scenario.Resources are such things as processing effort,the various forms of memory capacity, and
+communications channels (Bobrow & Norman,
+1975, in Rasmussen et al).Topics to be discussed include:  (1) a descriptionof the difficult tasks identified for analysis; (2) the
+general methodological approach; (3) develop-
+ment of the performance measures and issues
+related to their development; (4) discussion of the
+; and (5) discussion of the types of errors made by
+decision makers and interpretations for the cause
+of these errors based in the cognitive psychology
+literature.
+Paper presented to the Second International Command and Control Research and Technology Symposium,20Monterey, Ca.tactical operations require decision makingconditions of time pressure, stress, am-
+biguous, inaccurate and missing informa-
+tion, uncertain communications, and
+shifting conditions. These conditions make
+it difficult to perform careful analysis prior
+to making decisions.  Traditionally, most
+decision research and decision support
+system development have focused on well-
+defined tasks in carefully controlled envi-
+ronments.  Recent research has suggested
+that tactical decision makers use experi-
+ence to generate a likely course of action
+and then evaluate its feasibility using
+mental simulation.One way to prevent the same type of casualtyfrom being repeated is to thoroughly investigate
+and analyze root causes of actual mishaps, as well
+as data collected in a simulated tactical
+environment, and apply the findings in a concrete
+manner to improve tactical decision making.According to the recognition-primed decision-making (RPD) model, experienced decision
+makers can make rapid, high-quality decisions by
+associating a situation directly with the actions
+that normally work well in that kind of situation.
+Roberts, K. H., Stout, S. K., and Halpern, J. J.(1994)  Decision Dynamics in Two High
+Reliability Military Organizations.  Man-agement Science, Vol. 40, No. 5, May1994, 614-624.Tetlock, P. E., Accountability and the Persever-ance of First Impressions, Social Psycho-logical Quarterly, 26(1983), 285-292.Tetlock, P. E., Accountability:  The NeglectedSocial Context of Judgment and Choice, in
+L. L. Cummings and B. M. Straw (Eds.),
+Research in Organizational Behavior, Vol.7, JAI Press, Greenwich, CT, 1985, 297-332.The comparison to norms tool is based on acognitive model of human information processing
+which uses a feature matching strategy.  The
+model proposes a data-driven process.  As such,
+the comparison to norms tool is a knowledge-
+based tool with the knowledge represented as
+templates. Each template is a linked timeline
+associating individual features, through a series of
+feature matches, with expected actions. This
+allows the tool to make an assessment of the
+situation presented to the decision maker. A
+related module is the response manager. This
+assessment consists of a categorization of the
+situation (e.g., “hostile aircraft attacking”) and
+presentation of an associated template. The
+response manager module shows a template for an
+assessment of the current situation. This is a
+timeline display for template features, or events.
+The response manager module also provide inputs
+to the Prioritized track list, Alerts, and Responses
+and Tripwires.SABER is a model of another cognitivestrategy employed in making decisions. This
+strategy is known as explanation-based reasoning,
+or story generation.  In this approach, available
+data are assembled into explanatory structures,
+with one structure for each possible conclusion.
+Each of the explanations attempts to explain how
+every piece of data can be accounted for in
+support of each conclusion, even though some of
+the data items would naturally contradict reaching
+some conclusions. Contradictory data are
+explained through the of internal assumptions.  It
+is assumed that there are a fixed number of pre-
+defined possible conclusions and each data item
+points directly to one of those possible conclu-
+sions.Once the explanations are constructed,SABER evaluates them to determine which seems
+most plausible. Plausibility is based on three
+Paper presented to the Second International Command and Control Research and Technology Symposium,21Monterey, Ca.criteria:  simplicity, completeness, and impor-tance.  SABER provide data products which relate
+data to explanatory hypotheses for the current
+situation.  Hypotheses are presented in rank order,
+with evidence for and against each hypothesis and
+missing data is presented below the corresponding
+hypothesis. For a more detailed description of the
+DSS the reader is referred to Hair & Pickslay,
+1992; Hair, Pickslay, & Chow, 1992; Hutchins
+and Rummel, 1994.Hair, D. C. and Pickslay, K. (1992). Explanation-Based Reasoning in Decision Support
+Systems, Proceedings of the 9th AnnualConference on Command and Control De-cision Aids, June 1992, Monterey, CA.Hair, D. C.,  Pickslay, K. & Chow, S.  (1992)Explanation-Based Decision Support in
+Real Time Situations.  Proceedings of the1992 IEEE International Conference onTools with AI.  Nov. 1992, Arlington, VA.Various studies have indicated that as much as 90
+percent of industrial and system failures are
+produced by human error (Senders & Moray,
+1991).
+At the same time, a shift was occurring in U.S.
+Navy doctrine, away from a "blue-water" strategy
+to a doctrine of littoral operations aimed at
+potentially hostile regional powers.Tactical decision makers in today's oper-ating environment are required to perform
+complex tasks in a highly dynamic envi-
+ronment. Numerous interactive surface
+units and aircraft whose parameters are in
+flux and which must be continually
+sensed, processed, their future status pro-
+jected (development of hypotheses re-
+garding their future behavior) and actions
+taken to assure the successful outcome.Operating close to land presents additionalchallenges to the tactical decision maker.
+In littoral (i.e., near-land) settings, which
+most likely represent the majority of future
+anticipated naval conflicts, the decisions to
+be made are even more complex than they
+would be in full-scale warfare. When ci-
+vilian and neutral nation resources are in
+the conflict area incoming information
+carries an added element of uncertainty. In
+these situations, interpretation of the rules
+of engagement, contact identification, de-
+termining the capability and possible in-
+tent of the potential threat, and the
+shoot/no-shoot decision often pose ex-
+tremely difficult decision problems. (Since
+70 percent of the world's population lives
+within 200 miles of the sea, most future
+contingencies are likely to involve littoral
+warfare (Mundy, 1994).)as opposed to replacing "the user's approach tothe problem" (emphasis in original, Cohen, 1993)with tools based on decision analysis or
+mathematical optimization,
+ A considerably smaller number are attributable to
+other causes such as mechanical, electrical and
+materials failure (Meshkati, 1993). The purpose for this phase of the TADMUSprogram is to empirically evaluate the effective-
+ness of a DSS based on these recent approaches to
+decision support, to the extent possible,"This focus on why errors occuris...different from...the typical study of
+human errors which solely emphasize what
+occurs, a point of view which has received
+considerable criticism." (Rouse & Rouse,
+1983, p. 539)
+Paper presented to the Second International Command and Control Research and Technology Symposium,22Monterey, Ca. DSS which will mitigate typical types oferrors. The objective during this phase of
+the research was to develop an under-
+standing of the decision-making problems
+presented by current and future Navy sce-
+narios in order to identify the types and
+forms of information that are likely to fa-
+cilitate performance of these activities.
+Rouse, W. B. & Valusek, J. (1993). Evo-lutionary Design of Systems to Support
+Decision Making.  In G. A. Klein, J.
+Orasanu, R. Calderwood, & C. E.
+Zsambok (Eds.)  Decision Making inAction:  Models and Methods  (pp.270-286).  Ablex Publishing Corpora-
+tion, New Jersey.The Naturalistic Decision Making (NDM)model (Klein, 1989, 1993) seems more
+applicable than traditional decision-
+making models to the types of decisions
+involved in tactical decision-making.
+Early work conducted under the TADMUS
+program to determine the cognitive strate-
+gies employed by Navy AAW decision-
+makers found that when performing situa-
+tion assessment 87% of the time a recog-
+nitional strategy was used and 13% of the
+time story generation was used (Kaempf,
+Wolf, and Miller, 1993). We have applied
+these new models of human decision
+making—which parallel the cognitive
+strategies used by domain experts—to the
+design of a DSS for enhancing antiair war-
+fare tactical decision making. These two
+models for situation assessment are feature
+matching and story generation.
+(The scenarios were intentionally devel-
+oped to have many tracks with a high de-
+gree of uncertainty associated where it is
+not always clear whether a particular track
+should be engaged. Our interest in theTADMUS program was in gaining insightinto and aiding the decision process. The
+reader is referred to Hutchins, 1995, for a
+detailed coverage of performance meas-
+urement issues and scenario development
+issues.)
+These conditions include dynamic, fluid
+situations, time pressure, high risk, multi-
+ple decision makers, shifting and compet-
+ing goals, action-feedback loops, and
+situations with uncertain and incomplete
+data (Orasanu & Connolly, 1993).

adp5_0-1-30

@@ -0,0 +1,30 @@
+DISTRIBUTION  RESTRICTION:  Approved  for  public  release;;  distribution  is  unlimited.HEADQUARTERS,  DEPARTMENT  OF  THE  ARMYMAY  2012THE  OPERATIONS  PROCESSADP  5-­0
+                       This publication is available at Army Knowledge Online (https://armypubs.us.army.mil/doctrine/index.html). 
+                       *ADP 5-0 (FM 5-0) Army Doctrine Publication Headquarters Department of the Army No. 5-0 (FM 5-0) Washington, DC, 17 May 2012 The Operations Process Contents Page   PREFACE ..................................................................................................... ii  Definition and Purpose ........................................................................... 1  Principles of the Operations Process ..................................................... 2  Activities of the Operations Process....................................................... 6  Conclusion ........................................................................................... 16  GLOSSARY .................................................................................. Glossary-1  REFERENCES .......................................................................... References-1  Figures Figure 1. The operations process underlying logic ...............................................iv  Figure 2. The operations process ......................................................................... 1  Table Table 1. Preparation activities ............................................................................ 11  DISTRIBUTION RESTRICTION. Approved for public release; distribution is unlimited. *This publication supersedes FM 5-0, dated 26 March 2010. i 
+                                                   Preface Army Doctrine Publication (ADP) 5-0, The Operations Process, constitutes the Army’s view on planning, preparing, executing, and assessing operations. (See figure 1 on page iv.) It accounts for the complex, ever-changing, and uncertain nature of operations and recognizes that a military operation is foremost a human undertaking. As such, this publication emphasizes the philosophy of mission command to include the central role of commanders (supported by their staffs) in driving the operations process.  To comprehend the doctrine contained in ADP 5-0, readers must first understand the foundations of unified land operations described in ADP 3-0, Unified Land Operations. Readers must also fully understand the principles of mission command described in ADP 6-0, Mission Command.  For a detailed explanation of the operations process, readers should refer to Army Doctrine Reference Publication (ADRP) 5-0, The Operations Process. The principal audience for ADP 5-0 includes Army commanders, leaders, and unit staffs (officers, noncommissioned officers, and Soldiers). Commanders and staffs of Army headquarters serving as a joint task force or multinational headquarters should also refer to applicable joint or multinational doctrine concerning the range of military operations as well as joint or multinational forces. Trainers and educators throughout the Army will also use this manual.  Commanders, staffs, and subordinates ensure their decisions and actions comply with applicable U.S., international, and, in some cases, host nation laws and regulations. Commanders at all levels ensure their Soldiers operate in accordance with the law of war and the rules of engagement. (See Field Manual [FM] 27-10.) ADP 5-0 uses joint terms where applicable. Selected joint and Army terms and definitions appear in both the glossary and the text. Terms for which ADP 5-0 is the proponent publication (the authority) are marked with an asterisk (*) in the glossary. Definitions for which ADP 5-0 is the proponent publication are boldfaced in the text. These terms and their definitions will be in the next revision of FM 1-02. For other definitions shown in the text, the term is italicized and the number of the proponent publication follows the definition. ADP 5-0 applies to the Active Army, Army National Guard/Army National Guard of the United States, and United States Army Reserve unless otherwise stated. The proponent of ADP 5-0 is the United States Army Combined Arms Center. The preparing agency is the Combined Arms Doctrine Directorate, United States Army Combined Arms Center. Send comments and recommendations on a DA Form 2028 (Recommended Changes to Publications and Blank Forms) to Commander, U.S. Army Combined Arms Center and Fort Leavenworth, ATTN: ATZL-MCK-D (ADP 5-0), 300 McPherson Avenue, Fort Leavenworth, KS 66027-2337; by e-mail to usarmy.leavenworth.mccoe.mbx.cadd-org-mailbox@mail.mil; or submit an electronic DA Form 2028. ADP 5-0 17 May 2012 ii 
+��������������������������������ACKNOWLEDGEMENT Cover photo courtesy of the U.S. Army at http://www.flickr.com/photos/soldiersmediacenter/6846045865/.   
+�������  iv ADP 5-0 17 May 2012  Figure 1. The operations process underlying logic  
+                          This publication defines and describes the operations process. It provides principles commanders and staffs consider to effectively plan, prepare, execute, and continuously assess operations. DEFINITION AND PURPOSE   1. The Army’s framework for exercising mission command is the operations process— the major mission command activities performed during operations: planning, preparing, executing, and continuously assessing the operation. Commanders, supported by their staffs, use the operations process to drive the conceptual and detailed planning necessary to understand, visualize, and describe their operational environment; make and articulate decisions; and direct, lead, and assess military operations.  Figure 2. The operations process 2. The activities of the operations process are not discrete; they overlap and recur as circumstances demand. Planning starts an iteration of the operations process. Upon completion of the initial order, planning continues as leaders revise the plan based on changing circumstances. Preparing begins during planning and continues through execution. Execution puts a plan into action by applying combat power to seize, retain, and exploit the initiative to gain a position of relative advantage. Assessing is continuous and influences the other three activities.   3. Both the commander and staff have important roles within the operations process. The commander’s role is to drive the operations process as depicted in figure 2. The 17 May 2012 ADP 5-0 1 
+                                                          ADP 5-0 staff’s role is to assist commanders with understanding situations, making and implementing decisions, controlling operations, and assessing progress. In addition, the staff assists subordinate units (commanders and staffs), and keeps units and organizations outside the headquarters informed throughout the operations process. (ATTP 5-0.1 discusses the duties and responsibilities of the staff in detail.) PRINCIPLES OF THE OPERATIONS PROCESS  4. The philosophy of mission command guides commanders, staffs, and subordinates as they plan, prepare, execute, and assess operations. Mission command requires an environment of mutual trust and shared understanding among commanders, staffs, and subordinates. It requires a command climate in which commanders encourage subordinates to accept prudent risk and exercise disciplined initiative to seize opportunities and counter threats within the commander’s intent. Through mission orders, commanders focus their instructions on the purpose of the operation rather than on the details of how to perform assigned tasks. Doing this minimizes detailed control and allows subordinates the greatest possible freedom of action. Finally, when delegating authority to subordinates, commanders set the necessary conditions for success by allocating appropriate resources to subordinates based on assigned tasks. 5. Commanders and staffs use the operations process to integrate numerous tasks that are executed throughout the headquarters and with subordinate units. Commanders must organize and train their staffs and subordinates as an integrated team to simultaneously plan, prepare, execute, and assess operations. In addition to the principles of mission command discussed in ADP 6-0, commanders and staffs consider the following principles for the effective use of the operations process: �Commanders drive the operations process. �Build and maintain situational understanding. �Apply critical and creative thinking. �Encourage collaboration and dialogue. COMMANDERS DRIVE THE OPERATIONS PROCESS 6. Commanders are the most important participants in the operations process. While staffs perform essential functions that amplify the effectiveness of operations, commanders drive the operations process through understanding, visualizing, describing, directing, leading, and assessing operations. Understand 7. To understand something is to grasp its nature and significance. Understanding includes establishing context—the set of circumstances that surround a particular event or situation. Throughout the operations process, commanders develop and improve their understanding of their operational environment and the problem. An operational environment is a composite of the conditions, circumstances, and influences that affect the employment of capabilities and bear on the decisions of the commander (JP 3-0). Both conceptual and detailed planning assist commanders in developing their initial ADP 5-0 17 May 2012 2 
+                                                 The Operations Process  understanding of the operational environment and the problem. Based on personal observations and inputs from others (to include running estimates), commanders improve their understanding and modify their visualization throughout the conduct of operations.  Visualize 8. As commanders begin to understand their operational environment and the problem, they start visualizing a desired end state and potential solutions to solve the problem. Collectively, this is known as commander's visualization—the mental process of developing situational understanding, determining a desired end state, and envisioning an operational approach by which the force will achieve that end state. Commander’s visualization begins in planning and continues throughout the operations process until the force accomplishes the mission. During planning, commander’s visualization provides the basis for developing plans and orders. During execution, it helps commanders determine if, when, and what to decide, as they adapt to changing conditions.  Describe 9. After commanders visualize an operation, they describe it to their staffs and subordinates to facilitate shared understanding and purpose. During planning, commanders ensure subordinates understand their visualization well enough to begin course of action development. During execution, commanders describe modifications to their visualization resulting in fragmentary orders that adjust the original order. Commanders describe their visualization in doctrinal terms, refining and clarifying it as circumstances require. Commanders express their visualization in terms of— �Commander’s intent. �Planning guidance, including an operational approach. �Commander’s critical information requirements. �Essential elements of friendly information. 10. The commander’s intent is a clear and concise expression of the purpose of the operation and the desired military end state that supports mission command, provides focus to the staff, and helps subordinate and supporting commanders act to achieve the commander’s desired results without further orders, even when the operation does not unfold as planned (JP 3-0). During planning, the initial commander's intent drives course of action development. In execution, the commander’s intent guides disciplined initiative as subordinates make decisions when facing unforeseen opportunities or countering threats. 11. In addition to issuing their commander’s intent, commanders provide planning guidance that conveys the essence of their visualization. Effective planning guidance broadly describes when, where, and how the commander intends to employ combat power to accomplish the mission within the higher commander’s intent. Planning guidance includes an operational approach—a description of the broad actions the force must take to transform current conditions into those desired at end state (JP 5-0). The 17 May 2012 ADP 5-0 3 
+                                                       ADP 5-0 operational approach forms the basis of the unit’s concept of operations and serves as the link between conceptual and detailed planning.  12. Commanders also describe gaps in their visualization by stating their commander’s critical information requirements (CCIRs). Commanders use CCIRs to focus information collection on the relevant information they need to make critical decisions throughout the conduct of operations. The two components of CCIRs are friendly force information requirements and priority intelligence requirements. 13. In addition to information commanders need, commanders also describe the information they want protected as essential elements of friendly information (EEFIs). EEFIs establish an element of information to protect rather than one to collect. EEFIs identify those elements of friendly force information that, if compromised, would jeopardize mission success. Direct 14. Commanders direct all aspects of operations by establishing their commander’s intent, setting achievable objectives, and issuing clear tasks to subordinate units. Throughout the operations process, commanders direct forces by— �Preparing and approving plans and orders. �Establishing command and support relationships. �Assigning and adjusting tasks, control measures, and task organization. �Positioning units to maximize combat power.  �Positioning key leaders at critical places and times to ensure supervision. �Allocating resources to exploit opportunities and counter threats. �Committing the reserve as required. Lead 15. Through leadership, commanders provide purpose, direction, and motivation to subordinate commanders, their staff, and Soldiers. In many instances, a commander’s physical presence is necessary to lead effectively. Where the commander locates within the area of operations is an important leadership consideration. Commanders balance their time between leading the staff through the operations process and providing purpose, direction, and motivation to subordinate commanders and Soldiers away from the command post. Assess  16. Commanders continuously assess the situation to better understand current conditions and determine how the operation is progressing. Continuous assessment helps commanders anticipate and adapt the force to changing circumstances. Commanders incorporate the assessments of the staff, subordinate commanders, and unified action partners into their personal assessment of the situation. Based on their assessment, commanders modify plans and orders to adapt the force to changing circumstances.  ADP 5-0 17 May 2012 4 
+                                                            The Operations Process  BUILD AND MAINTAIN SITUATIONAL UNDERSTANDING 17. Success in operations demands timely and effective decisions based on applying judgment to available information and knowledge. As such, commanders and staffs seek to build and maintain situational understanding throughout the operations process. Situational understanding is the product of applying analysis and judgment to relevant information to determine the relationships among the operational and mission variables to facilitate decisionmaking. Building and maintaining situational understanding is essential for establishing the situation’s context, developing effective plans, assessing operations, and making quality decisions throughout the operations process. Commanders continually strive to maintain their situational understanding and work through periods of reduced understanding as the situation evolves. 18. Commanders and staffs use the operational and mission variables to help build their situational understanding. They analyze and describe an operational environment in terms of eight interrelated operational variables: political, military, economic, social, information, infrastructure, physical environment, and time (PMESII-PT). Upon receipt of a mission, commanders filter information categorized by the operational variables into relevant information with respect to the mission. They use the mission variables, in combination with the operational variables, to refine their understanding of the situation and to visualize, describe, and direct operations. The mission variables are mission, enemy, terrain and weather, troops and support available, time available, and civil considerations (METT-TC).  APPLY CRITICAL AND CREATIVE THINKING 19. Commanders and staffs apply critical and creative thinking throughout the operations process to assist them with understanding situations, making decisions, and directing action. Critical thinking is purposeful and reflective judgment about what to believe or what to do in response to observations, experience, verbal or written expressions, or arguments. Creative thinking involves creating something new or original. Creative thinking leads to new insights, novel approaches, fresh perspectives, and new ways of understanding and conceiving things. 20. Critical and creative thinking are indispensible to the operations process. For both commanders and staff, these two skills begin with a rigorous analysis of friendly and enemy forces, as they relate to one another in time and space. This analysis includes weapons system ranges, mobility options afforded by terrain and weather, operational reach, communications system range, sustainment, and other considerations of the operational and mission variables. Disciplined and focused analysis of the operational and mission variables, coupled with critical and creative thinking about the challenges and opportunities resulting from that analysis, is essential to developing a full appreciation of the range of alternatives available to accomplish assigned missions. ENCOURAGE COLLABORATION AND DIALOGUE 21. Throughout the operations process, commanders encourage continuous collaboration and dialogue among commanders, staffs, and unified action partners to create shared 17 May 2012 ADP 5-0 5 
+                                                        ADP 5-0 understanding and facilitate unity of effort. Collaboration is two or more people or organizations working together toward common goals by sharing knowledge and building consensus. Dialogue is a way to collaborate by involving the candid exchange of ideas or opinions among participants that encourages frank discussions in areas of disagreement.  22. Commanders, staffs, and unified action partners collaborate and dialogue actively, sharing and questioning information, perceptions, and ideas to better understand situations and make decisions. Collaboration and dialogue assist in developing shared understanding and purpose, building teams, and making rapid adjustments during execution. ACTIVITIES OF THE OPERATIONS PROCESS 23. The operations process consists of the major mission command activities: planning, preparing, executing, and assessing. PLANNING 24. Planning is the art and science of understanding a situation, envisioning a desired future, and laying out effective ways of bringing that future about. Army leaders plan to create a common vision among subordinate commanders, staffs, and unified action partners for the successful execution of operations. Planning results in a plan or order that communicates this vision and directs actions to synchronize forces in time, space, and purpose for achieving objectives and accomplishing missions.  Integrated Planning 25. Planning consists of two separate, but closely related, components: a conceptual component and a detailed component. Conceptual planning involves understanding the operational environment and the problem, determining the operation’s end state, and visualizing an operational approach. Conceptual planning generally corresponds to operational art and is the focus of the commander with staff support. Detailed planning translates the broad operational approach into a complete and practical plan. Generally, detailed planning is associated with the science of operations including the synchronization of the forces in time, space, and purpose. Detailed planning works out the scheduling, coordination, or technical problems involved with moving, sustaining, and synchronizing the actions of force as a whole toward a common goal. Effective planning requires the integration of both the conceptual and detailed components of planning. Planning and Operational Art 26. Operational art is the cognitive approach by commanders and staffs—supported by their skill, knowledge, experience, creativity, and judgment—to develop strategies, campaigns, and operations to organize and employ military forces by integrating ends, ways, and means (JP 3-0). Operational art guides the conceptual and detailed aspects of planning to produce executable plans and orders. Operational art applies to all aspects of ADP 5-0 17 May 2012 6 
+                                                       The Operations Process  operations and integrates ends, ways, and means, while accounting for risk and opportunities, across the levels of war. 27. The elements of operational art (see ADRP 3-0) assist commanders and staffs in the application of operational art. These conceptual tools help commanders think through the challenges of understanding their operational environment, defining the problem, developing an operational approach, and articulating their planning guidance that drives more detailed planning.   Army Planning Methodologies Elements of operational art •  End state and conditions •  Center of gravity •  Decisive points •  Lines of operations and lines of effort •  Operational reach •  Basing •  Tempo •  Phasing and transitions •  Culmination •  Risk 28. Army leaders employ three methodologies for planning. Commanders and staffs determine the appropriate mix of these methodologies based on the scope of the problem, their familiarity with it, the time available, and the availability of a staff. Methodologies that assist commanders and staffs with planning include— �Army design methodology. �Military decisionmaking process (MDMP). �Troop leading procedures (TLP). Army Design Methodology 29. The Army design methodology is a methodology for applying critical and creative thinking to understand, visualize, and describe unfamiliar problems and approaches to solving them. Army design methodology is an iterative process of understanding and problem framing that uses elements of operational art to conceive and construct an operational approach to solve identified problems. Commanders and their staffs use Army design methodology to assist them with the conceptual aspects of planning.   30. Army design methodology entails framing the operational environment, framing the problem, and developing an operational approach to solve the problem. Army design methodology results in an improved understanding of the operational environment, a problem statement, an initial commander’s intent, and an operational approach that serves as the link between conceptual and detailed planning. Based on their understanding and learning gained during Army design methodology, commanders issue planning guidance, to include an operational approach, to guide more detailed planning using the MDMP.  31. The understanding developed through Army design methodology continues through preparation and execution in the form of continuous assessment. Assessment, to include 17 May 2012  ADP 5-0 7 
+                                                              ADP 5-0 updated running estimates, helps commanders measure the overall effectiveness of employing forces and capabilities to ensure that the operational approach remains feasible and acceptable within the context of the higher commander’s intent and concept of operations. If the current operational approach fails to meet these criteria, or if aspects of the operational environment or problem change significantly, the commander may decide to reframe. Reframing involves revisiting earlier hypotheses, conclusions, and decisions that underpin the current operational approach. Reframing can lead to a new problem statement and operational approach, resulting in an entirely new plan. Military Decisionmaking Process 32. The military decisionmaking process is an iterative planning methodology to understand the situation and mission, develop a course of action, and produce an operation plan or order. The MDMP combines the conceptual and detailed aspects of planning and integrates the activities of the commander, staff, subordinate headquarters, and other partners throughout the planning process. The MDMP helps leaders apply thoroughness, clarity, sound judgment, logic, and professional knowledge to understand situations, develop options to solve problems, and reach decisions. The MDMP results in an improved understanding of the situation and a plan or order that guides the force through preparation and execution. 33. The MDMP facilitates collaborative and parallel planning as the higher headquarters solicits input and continually shares information concerning future operations with subordinate and adjacent  units, supporting and supported units, and unified action partners through planning meetings, warning orders, and other means. Commanders encourage active collaboration among all organizations affected by the pending operations to build shared understanding, participate in course of action development and decisionmaking, and resolve conflicts before publication of the plan or order. 34. The MDMP consists of a series of steps that have various inputs and outputs. The outputs lead to an increased understanding of the situation facilitating the next step of the MDMP. Commanders and staffs generally perform these steps sequentially; however, they may revisit several steps in an iterative fashion, as they learn more about the situation before producing the plan or order. The steps of the MDMP are— �Step 1 – Receipt of mission. �Step 2 – Mission analysis. �Step 3 – Course of action development. �Step 4 – Course of action analysis. �Step 5 – Course of action comparison. �Step 6 – Course of action approval. �Step 7 – Orders production, dissemination, and transition. Troop Leading Procedures 35. Troop leading procedures are a dynamic process used by small-unit leaders to analyze a mission, develop a plan, and prepare for an operation. TLP are used by commanders and leaders without a staff. These procedures enable leaders to maximize ADP 5-0 17 May 2012 8 
+                                                              The Operations Process  available planning time while developing effective plans and preparing their units for an operation. Like the MDMP, troop leading procedures consist of a series of steps: �Step 1 – Receive the mission. �Step 2 – Issue a warning order. �Step 3 – Make a tentative plan. �Step 4 – Initiate movement. �Step 5 – Conduct reconnaissance. �Step 6 – Complete the plan. �Step 7 – Issue the order. �Step 8 – Supervise and refine the plan. 36. The sequence of the steps of troop leading procedures is not rigid. Leaders modify them as required. Higher headquarters issue frequent warning orders to optimize available time for subordinates to conduct their TLP. Guides to Effective Planning 37. Planning helps commanders understand and develop solutions to problems, anticipate events, adapt to changing circumstances, task-organize the force, and prioritize efforts. Effective planning requires dedication, study, and practice. Planners must be technically and tactically competent within their areas of expertise and disciplined in the use of doctrinally correct terms and symbols. The following guides aid in effective planning: �Commanders focus planning. �Develop simple, flexible plans through mission orders. �Optimize available planning time. �Continually refine the plan. Commanders Focus Planning 38. Commanders are the most important participants in effective planning. They focus the planning effort by providing their commander’s intent, issuing planning guidance, and making decisions throughout the planning process. Commanders apply discipline to the planning process to meet the requirements of time, planning horizons, simplicity, level of detail, and desired outcomes. Commanders ensure that all operation plans and orders comply with applicable domestic and international laws. They also confirm that the plan or order is relevant and suitable for subordinates. Generally, the more involved commanders are in planning, the faster staffs can plan. Through personal involvement, commanders ensure the plan reflects their commander’s intent. Develop Simple, Flexible Plans Through Mission Orders 39. Effective plans and orders are simple and direct. Staffs prepare clear, concise orders that communicate an understanding of the operation through the use of doctrinally correct operational terms and symbols. Doing this minimizes chances of 17 May 2012 ADP 5-0 9 
+                                         ADP 5-0 misunderstanding. Clarity and brevity are important. Shorter, rather than longer, plans aid in simplicity. Shorter plans are easier to disseminate, read, and remember. 40. Flexible plans help units adapt quickly to changing circumstances. Commanders and planners build opportunities for initiative into plans by anticipating events that allow them to operate inside of the enemy’s decision cycle or to react promptly to deteriorating situations. Identifying decision points and designing branches ahead of time—combined with a clear commander’s intent—help create flexible plans. 41. Commanders stress the importance of using mission orders as a way of building simple, flexible plans. Mission orders are directives that emphasize to subordinates the results to be attained, not how they are to achieve them (ADP 6-0). Mission orders clearly convey the unit’s mission and the commander’s intent. Mission orders focus subordinates on what to do and the purpose of doing it, without prescribing exactly how to do it. Commanders establish control measures to aid cooperation among forces without imposing needless restriction on freedom of action. Optimize Available Planning Time 42. Time is a critical variable in operations. Therefore, time management is important in planning. Whether done deliberately or rapidly, all planning requires the skillful use of available time to optimize planning and preparation throughout the unit. Taking more time to plan often results in greater synchronization; however, any delay in execution risks yielding the initiative—with more time to prepare and act—to the enemy. When allocating planning time to the staff, commanders must ensure subordinates have enough time to plan and prepare their own actions prior to execution. Commanders follow the “one-third—two-thirds rule” as a guide to allocate time available. They use one-third of the time available before execution for their planning and allocate the remaining two-thirds of the time available before execution to their subordinates for planning and preparation. Continually Refine the Plan 43. Planning does not cease with production of a plan or order. It continues throughout an operation as the order is refined based on confirmation briefings, rehearsals and changes in the situation. In addition, staffs are always refining plans for branches and sequels throughout an operation. During preparation and execution, the plan is continuously refined as situational understanding improves. Through assessment, subordinates and others provide feedback on the progress of operations. In some circumstances, commanders may determine that the current order (to include associated branches and sequels) is no longer relevant to the situation. In these instances, instead of modifying the current plan, commanders reframe the problem and develop an entirely new plan. PREPARING 44. Preparation consists of those activities performed by units and Soldiers to improve their ability to execute an operation. Preparation creates conditions that improve friendly forces’ opportunities for success. It requires commander, staff, unit, and 10 ADP 5-0 17 May 2012 
+                                                                The Operations Process  Soldier actions to ensure the force is trained, equipped, and ready to execute operations. Effective preparation helps commanders, staffs, and subordinate units better understand the situation and their roles in upcoming operations. The major activities of preparation are listed in table 1. Table 1. Preparation activities Continue to coordinate and conduct liaison Conduct rehearsals Initiate information collection Conduct plans-to-operations transitions Initiate security operations Revise and refine the plan Initiate troop movement Integrate new Soldiers and units Initiate sustainment preparations Complete task organization Initiate network preparations Train Manage terrain Perform pre-operations checks and inspections Prepare terrain Continue to build partnerships and teams Conduct confirmation briefs 45. Mission success depends as much on preparation as on planning. Higher headquarters may develop the best of plans; however, plans serve little purpose if subordinates do not receive them in time. Subordinates need enough time to fully comprehend the plan, rehearse key portions of the plan, and ensure Soldiers and equipment are positioned and ready to execute the operation. The following guidelines aid in effective preparation: �Secure and protect the force. �Improve situational understanding. �Understand, rehearse, and refine the plan. �Integrate, organize, and configure the force. �Ensure forces and resources are ready and positioned. Secure and Protect the Force 46. The force as a whole is often most vulnerable to surprise and enemy attack during preparation. As such, security operations—screen, guard, cover, area security, and local security—are essential during preparation. In addition, commanders ensure the various tasks of the protection warfighting function are fully integrated to safeguard bases, secure routes, and protect the force, while it prepares for operations. Improve Situational Understanding 47. During preparation, commanders may realize that their initial understanding developed during planning may be neither accurate nor complete. As such, commanders strive to validate assumptions and improve their situational understanding as they prepare for operations. Information collection (to include reconnaissance, surveillance, and intelligence operations) helps improve understanding of the enemy, terrain, and civil 17 May 2012 ADP 5-0 11 
+                                                          ADP 5-0 considerations. Inspections, rehearsals, liaison, and coordination help leaders improve their understanding of the friendly force.   Understand and Rehearse the Plan 48. A successful transition from planning to execution requires those charged with executing the order to understand the plan fully. The transition between planning and execution takes place both internally in the headquarters and externally between the commander and subordinate commanders. Rehearsals, to include confirmation briefings and plans-to-operations transition briefings, help improve understanding of the concept of operations, control measures, decision points, and command and support relationships. Integrate, Organize, and Configure the Force  49. During preparation, commanders allocate time to put the new task organization into effect. This includes detaching units, moving forces, and receiving and integrating new units and Soldiers into the force. When units change task organization, they need preparation time to learn the gaining unit’s standard operating procedures and the plan the gaining unit will execute. The gaining unit needs preparation time to assess the new unit’s capabilities and limitations and to integrate new capabilities.   Ensure Forces and Resources are Ready and Positioned 50. Effective preparation ensures that the right forces are in the right place, at the right time, with the right equipment and other resources ready to execute the operation. Concurrent with task organization, commanders use troop movement to position or reposition forces to the correct locations prior to execution. This includes positioning sustainment units and supplies. EXECUTING 51. Execution is putting a plan into action by applying combat power to accomplish the mission. During execution, commanders, staffs, and subordinate commanders focus their efforts on translating decisions into actions. They apply combat power to seize, retain, and exploit the initiative to gain and maintain a position of relative advantage. This is the essence of unified land operations (see ADP 3-0). Decisionmaking During Execution 52. Decisionmaking is tied to disciplined initiative and inherent in executing operations. Commanders observe the progress of operations and intervene when necessary to ensure success. Because operations never unfold exactly as envisioned and because understanding of the situation changes, a commander’s decisions made during execution are critical to an operation’s success. During execution, commanders direct their units forcefully and promptly to overcome the difficulties of enemy action, friendly errors, and other changes in their operational environment. 53. Commanders make execution and adjustment decisions throughout execution. Execution decisions implement a planned action under circumstances anticipated in the 12 ADP 5-0 17 May 2012 
+                                                           The Operations Process  order. An execution decision is normally tied to a decision point—a point in space or time the commander or staff anticipates making a key decision concerning a specific course of action (JP 5-0). An adjustment decision is the selection of a course of action that modifies the order to respond to unanticipated opportunities or threats. An adjustment decision may include a decision to reframe the problem and develop an entirely new plan. 54. Executing, adjusting, or abandoning the original operation is part of decisionmaking in execution. By fighting the enemy and not the plan, successful commanders balance the tendency to abandon a well-conceived plan too soon against persisting in a failing effort too long. Effective decisionmaking during execution— �Relates all actions to the commander’s intent and concept of operations to ensure they support the decisive operation. �Is comprehensive, maintaining integration of combined arms rather than dealing with separate functions.  �Relies heavily on intuitive decisionmaking by commanders and staffs to make rapid adjustments. �Is continuous and responds effectively to any opportunity or threat. Guides to Effective Execution 55. During execution, the situation may change rapidly. Operations the commander envisioned in the plan may bear little resemblance to actual events in execution. Subordinate commanders need maximum latitude to take advantage of situations and meet the higher commander’s intent when the original order no longer applies. Effective execution requires leaders trained and educated in independent decisionmaking, aggressiveness, and risk taking in an environment of mission command. During execution, leaders must be able and willing to solve problems within the commander’s intent without constantly referring to higher headquarters. Subordinates need not wait for top-down synchronization to act. The following guides aid in effective execution: �Seize the initiative through action. �Accept prudent risk to exploit opportunities. Seize the Initiative Through Action 56. Commanders create conditions for seizing the initiative by acting. Without action, seizing the initiative is impossible. Faced with an uncertain situation, people naturally tend to hesitate and gather more information to reduce the uncertainty. Although waiting and gathering information might reduce uncertainty, it will not eliminate it. Waiting may even increase uncertainty by providing the enemy with time to seize the initiative. It is far better to manage uncertainty by acting and developing the situation. Accept Prudent Risk to Exploit Opportunities  57. Uncertainty and risk are inherent in all military operations. Successful commanders are comfortable operating under conditions of uncertainty, as they balance various risks while taking advantage of opportunities. Prudent risk is a deliberate exposure to potential 17 May 2012  ADP 5-0 13 
+                                            ADP 5-0 injury or loss when the commander judges the outcome in terms of mission accomplishment as worth the cost (ADP 6-0). Reasonably estimating and intentionally accepting risk is not gambling. Gambling, in contrast to taking prudent risk, is staking the success of an entire action on a single event without considering the hazard to the force should the event not unfold as envisioned. Therefore, commanders avoid taking a gamble. Commanders carefully determine risks, analyze and minimize as many hazards as possible, and then take prudent risks to exploit opportunities.  ASSESSING 58. Assessment is the determination of the progress toward accomplishing a task, creating an effect, or achieving an objective (JP 3-0). Assessment is a continuous activity of the operations process. The focus of assessment, however, changes for each operations process activity. During planning, assessment focuses on understanding current conditions of an operational environment and developing an assessment plan, including what and how to assess progress. During preparation, assessment focuses on determining the friendly force’s readiness to execute the operation and on verifying the assumptions on which the plan is based. During execution, assessment focuses on evaluating progress of the operation. Based on their assessment, commanders direct adjustments to the order, ensuring the operation stays focused on accomplishing the mission. Assessment Process 59. Assessment involves the continuous monitoring and evaluation of the current situation to determine progress of an operation. Broadly, assessment consist of the following activities: �Monitoring the current situation to collect relevant information. �Evaluating progress toward attaining end state conditions, achieving objectives, and completing tasks. �Recommending or directing action for improvement. 60. Primary tools for assessing include running estimates, after action reviews, and the assessment plan. Running estimates provide information, conclusions, and recommendations from the perspective of each staff section. Running estimates help to refine the common operational picture and supplement it with information not readily displayed. Both formal and informal after action reviews help identify what was supposed to happen, what went right, and what went wrong for a particular action or operation, and how the commander and staff should do things differently in the future. The assessment plan includes measures of effectiveness, measures of performance, and indicators that help the commander and staff evaluate progress toward accomplishing tasks and achieve objectives. (See ATTP 5-01.1 for doctrine on building assessment plans.) Running Estimates 61. Effective plans and successful preparation, execution and assessment hinge on accurate running estimates. A running estimate is the continuous assessment of the 14  ADP 5-0 17 May 2012 
+                                                    The Operations Process  current situation used to determine if the current operation is proceeding according to the commander’s intent and if planned future operations are supportable. Detailed running estimates begin in planning and are continuously updated during preparation and execution. In their running estimates, the command and staff continuously consider the effects of new information and update facts, assumptions, and conclusions. Running estimates from the staff always include recommendations to the commander. Guides to Effective Assessment 62. Throughout the conduct of operations, commanders integrate their own assessments with those of the staff, subordinate commanders, and other partners in the area of operations. The following guides aid in effective assessment: �Commanders prioritize the assessment effort. �Incorporate the logic of the plan. �Use caution when establishing cause and effect.  �Combine quantitative and qualitative indicators. Commanders Prioritize the Assessment Effort 63. Commanders establish priorities for assessment in their planning guidance, CCIRs, and decision points. By prioritizing the effort, commanders avoid excessive analyses when assessing operations. Committing valuable time and energy to developing excessive and time-consuming assessment schemes squanders resources better devoted to other operations process activities. Commanders reject the tendency to measure something just because it is measurable. Effective commanders avoid burdening subordinates and staffs with overly detailed assessments and collection tasks. Generally, the echelon at which a specific operation, task, or action is conducted should be the echelon at which it is assessed. Incorporate the Logic of the Plan 64. Effective assessment relies on an accurate understanding of the reasons and logic used to build the plan. Each plan is built on assumptions and an operational approach. The reasons and logic why the commander believes the plan will produce the desired results are important considerations when staffs determine how to assess operations. Recording and understanding this logic helps the staffs recommend the appropriate measures of effectiveness, measures of performance, and indicators for assessing the operation.  Use Caution When Establishing Cause and Effect 65. Although establishing cause and effect is sometimes difficult, it is crucial to effective assessment. Sometimes, establishing causality between actions and their effects can be relatively straightforward, such as in observing a bomb destroy a bridge. In other instances, especially regarding changes in human behavior, attitudes, and perception, 17 May 2012 ADP 5-0 15 
+                               ADP 5-0 establishing links between cause and effect proves difficult. Commanders and staffs must guard against drawing erroneous conclusions in these instances. Combine Quantitative and Qualitative Indicators 66. Effective assessment incorporates both quantitative (observation-based) and qualitative (opinion-based) indicators. Human judgment is integral to assessment. A key aspect of any assessment is the degree to which it relies upon human judgment and the degree to which it relies upon direct observation and mathematical rigor. Rigor offsets the inevitable bias, while human judgment focuses rigor and processes on intangibles that are often key to success. The appropriate balance depends on the situation— particularly the nature of the operation and available resources for assessment—but rarely lies at the ends of the scale. CONCLUSION 67. The doctrine in this publication provides a starting point for the execution of the operations process. It establishes a common frame of reference and the intellectual tools Army leaders use to plan, prepare for, execute, and assess operations. By establishing a common approach and language, the doctrine in this publication promotes mutual understanding and enhances the effectiveness during the conduct of operations. The doctrine in this publication is a guide for action rather than a set of fixed rules. While it provides an authoritative guide for leaders and Soldiers, it requires original application adapted to circumstances. In operations, effective leaders recognize when and where doctrine, training, or even their past experience no long fits the situation and adapt accordingly. 16 ADP 5-0 17 May 2012 
+                                                Glossary The glossary lists acronyms and terms with Army or joint definitions. Where Army and joint definitions differ, (Army) precedes the definition. Terms for which ADP 5-0 is the proponent are marked with an asterisk (*). The proponent publication for other terms is listed in parentheses after the definition. SECTION I – ACRONYMS AND ABBREVIATIONS ADP Army doctrine publication ADRP Army doctrine reference publication CCIR commander’s critical information requirement DA Department of the Army EEFI essential element of friendly information FM field manual JP joint publication MDMP military decisionmaking process METT-TC mission, enemy, terrain and weather, troops and support available, time available, and civil considerations PMESII-PT political, military, economic, social, information, infrastructure, physical environment, and time TLP troop leading procedures U.S. United States SECTION II – TERMS *Army design methodlogy A methodology for applying critical and creative thinking to understand, visualize, and describe unfamiliar problems and approaches to solving them.  assessment The determination of the progress toward accomplishing a task, creating an effect, or achieving an objective. (JP 3-0) commander’s intent A clear and concise expression of the purpose of the operation and the desired military end state that supports mission command, provides focus to the staff, and helps subordinate and supporting commanders act to achieve the commander’s desire result withou furthor order, even when the operation does not unfold as planned. (JP 3-0) 17 May 2012 ADP 5-0 Glossary-1 
+                                            Glossary *commander’s visualization The mental process of developing situational understanding, determining a desired end state, and envisioning an operational approach by which the force will achieve that end state. decision point A point in space or time the commander or staff anticipate making a key decision concerning a specific course of action. (JP 5-0) *execution Putting a plan into action by applying combat power to accomplish the mission. *military decisionmaking process An iterative planning methodology to understand the situation and mission, develop a course of action, and produce an operation plan or order. mission orders Directives that emphasize to subordinates the results to be attained, not how they are to achieve them. (ADP 6-0) operational approach A description of the broad action the force must take to transform current conditions into those desired at end state. (JP 5-0) operational art The cognitive approach by commanders and staffs—supported bytheir skill, knowledge, experience, creativity, and judment—to develop strategies, campaigns, and operations to organize and employ military forces by integrating ends, ways, and means. (JP 3-0) operational environment A composite of the conditions, circumstances, and influences that affect the employment of capabilities and bear on the decisions of the commander. (JP 3-0) *operations process The major mission command activities performed during operations: planning, preparing, executing, and continuously assessing the operation. *planning The art and science of understanding a situation, envisioning a desired future, and laying out effective ways of bringing that future about. *preparation Those activities performed by units and Soldiers to improve their ability to execute an operation.  prudent risk A deliberate exposure to potential injury or loss when the commander judges the outcome in terms of mission accomplishment as worth the cost. (ADP 6-0) Glossary-2 ADP 5-0 17 May 2012 
+                        Glossary *running estimate The continuous assessment of the current situation used to determine if the current operation is proceeding according to the commander’s intent and if planned future operations are supportable.  *situational understanding The product of applying analysis and judgment to relevant information to determine the relationships among the operational and mission variables to facilitate decisionmaking. *troop leading procedures A dynamic process used by small-unit leaders to analyze a mission, develop a plan, and prepare for an operation. 17 May 2012 ADP 5-0 Glossary-3 
+ This page intentionally left blank.  
+                     References Field manuals and selected joint publications are listed by new number followed by old number. REQUIRED PUBLICATIONS These documents must be available to intended users of this publication. FM 1-02 (101-5-1). Operational Terms and Graphics. 21 September 2004. JP 1-02. Department of Defense Dictionary of Military and Associated Terms. 8 November 2010. RELATED PUBLICATIONS These documents contain relevant supplemental information. JOINT PUBLICATIONS Most joint publications are available online:  <http://www.dtic.mil/doctrine/new_pubs/jointpub.htm.>  JP 3-0. Joint Operations, 11 August 2011.  JP 3-33. Joint Task Force Headquarters. 16 February 2007.  JP 5-0. Joint Operation Planning, 11 August 2011.  ARMY PUBLICATIONS Most Army doctrinal publications are available online: <http://www.apd.army.mil>. ADP 3-0 (FM 3-0). Unified Land Operations, 10 October 2011.  ADP 6-0 (FM 6-0). Mission Command. 17 May 2012.  ADRP 3-0. Unified Land Operations. 16 May 2012. ADRP 5-0. The Operations Process. 17 May 2012. ADRP 6-0. Mission Command. 17 May 2012.  ATTP 5-0.1. Commander and Staff Officer Guide, 14 September 2011.   FM 27-10. The Law of Land Warfare, 18 July 1956. WEB SITE Cover photo courtesy of the U.S. Army at http://www.flickr.com/photos/soldiersmediacenter/6846045865/. REFERENCED FORMS DA Form 2028. Recommended Changes to Publications and Blank Forms. 17 May 2012 ADP 5-0 References-1 
+ This page intentionally left blank.  
+                           ADP 5-0 17 May 2012 By order of the Secretary of the Army: RAYMOND T. ODIERNO General, United States Army Chief of Staff Official: JOYCE E. MORROW  Administrative Assistant to the     Secretary of the Army 1211423 DISTRIBUTION: Active Army, Army National Guard, and U.S. Army Reserve: To be distributed in accordance with the initial distribution number (IDN) 110412, requirements for ADP 5-0. 
+PIN: 102805-000