Search

We describe four major approaches to the study of science–historical accounts of scientific discoveries, psychological experiments with nonscientists working on tasks related to scientific discoveries, direct observation of ongoing scientific laboratories, and computational modeling of scientific discovery processes–by viewing them through the lens of the theory of human problem solving. We compare and contrast the different approaches, indicate their complementarities, and provide examples from each approach that converge on a set of principles of scientific discovery.

A common motivation for developing computational frameworks for diagrammatic reasoning is the hope that they might serve as re-configurable tools for studying human problem solving performance. Despite the ongoing debate as to the precise mechanisms by which diagrams, or any other external representation, are used in human problem solving, there is little doubt that diagrammatic representations considerably help humans solve certain classes of problems. In fact, there are a host of applications of diagrams and diagrammatic representations in computing, from data presentation to visual programming languages. In contrast to both the use of diagrams in human problem solving and the ubiquitous use of diagrams in the computing industry, the topic of this review is the use of diagrammatic representations in automated problem solving. We therefore investigate the common, and often implicit, assumption that if diagrams are so useful for human problem solving and are so apparent in human endeavour, then there must be analogous computational devices of similar utility.

This paper discusses how Herbert Simon's initial interest in decision making became transformed into a focus on understanding human problem solving in response to the concrete conditions of the Cold War and the practical goals of the military.In particular, it suggests a connection between the seachange in Simon's interest and his shift in patronage.As a result, Simon is portrayed as a component of the scientific-military World War II cyborg that further evolved during the Cold War.Moving from decision making to problem solving, Simon's cyborg science not only required large sums of money, but also managed to acquire these.

Over the years, research in mathematical problem-solving has examined expert/novice problem-solving performance on various types of problems and subjects. In particular, DeFranco examined two groups of Ph.D. mathematicians as they solved four mathematics problems and found that although all were content experts , only one group were problem-solving experts . Based on this study, this article posits the notion that one distinguishing feature between experts and novices is that experts tend to look for special features of a problem and use algorithms only as a “fail-safe” system while novices act like “machines” relying on algorithms to solve the problems. Why? The article explores the idea that novice problem solvers learned to solve problems the way they learned proof, that is, in a formal, abstract and mechanizable way. Beliefs about proof and the culture in which it is practiced help frame a mathematician's view of the discipline and ultimately impacts classroom practice. The authors believe that current classroom instruction tends to create a culture that fosters algorithmic proficiency and a “machine-like” approach to the learning of mathematics and problem-solving. Further, they argue that mathematicians need to be aware of the distinction between knowing a proof is true and explaining why it is true. When these distinctions are appreciated and practiced during classroom instruction, then and only then will students begin to acquire the mathematical knowledge to become better problem solvers.

This paper presents a bibliography of 263 references related to human problem solving, arranged by subject matter. The references were taken from PsycInfo and Academic Premier database. Journal papers, book chapters, and dissertations are included. The topics include human development, education, neuroscience, and research in applied settings. It is argued that researchers are more and more engaged with problem solving research because of its centrality in human actions and because society needs advice from science in understanding and solving complex problems.

Pictorial representations are widely used in human problem solving. For blind and visually impaired people, haptic interfaces can provide perceptual access to graphical representations. We propose line-based graphics as a type of graphics, which are suitable to be explored by blind and visually impaired people, and which can be successfully augmented with auditory assistance by speech or non-verbal sounds. The central prerequisite for realizing powerful assistive interaction is monitoring the users' haptic exploration and in particular the recognition of exploratory events. The representational layers of line-based graphics as well as of exploration-event descriptions are specified by qualitative spatial propositions. Based on these representations, event recognition is performed by rule-based processes.

The research question addressed in this paper is: Given a problem, can we automatically predict how difficult the problem will be to solve by humans? We focus our investigation on problems in which the difficulty arises from the combinatorial complexity of problems. We propose a measure of difficulty that is based on modeling the problem solving effort as search among alternatives and the relations among alternative solutions. In experiments in the chess domain, using data obtained from very strong human players, this measure was shown at a high level of statistical significance to be adequate as a genuine measure of difficulty for humans.

A chance has two contrary aspects: suddenness as an accidental finding and gradualness as the result of a prepared mind. Such duality of chance discovery resembles the insight process treated by problem solving researches. In this paper, we focus on the insight process in human problem solving, present a broad overview of its suddenness and the gradualness, and introduce our experimental results from the viewpoint of the duality of insight. We believe that our research findings will contribute to studies of chance discovery.

Design educators often comment on the difficulties that result from a premature commitment by students to a solution to a design problem. Similarly practitioners can find it difficult to move away from an idea they have developed or precedents in a field. In the psychology of problem solving this effect is called functional fixedness or fixation. It is not surprising that these effects should occur in design problem solving. However, while these types of issues have been discussed in the context of design, there has been little systematic evidence available about whether or not and under what conditions design fixation does occur. The paper reviews the results of a series of recent experiments which begin to address these issues. The results of the experiments are examined in terms of what insights they provide into the design process, what implications they have for design education and how they relate to the larger and more general area of human problem solving.

This paper argues that all discoveries, if they can be viewed as autonomous learning from the environment, share a common process. This is the process of model abstraction involving four steps: act, predict, surprise, and refine, all built on top of the discoverer’s innate actions, percepts, and mental constructors. The evidence for this process is based on observations on various discoveries, ranging from children playing to animal discoveries of tools, from human problem solving to scientific discovery. Details of this process can be studied with computer simulations of discovery in simulated environments.

SUMMARY An electronic performance support system (EPSS) is an interactive computer‐based environment that is intended to facilitate and/or improve human problem solving capability within some target application domain. EPSS tools can be used by individuals, groups and organizations in order to improve various facets of their operation. The nature of an EPSS facility will therefore depend upon who it is intended to serve and the job tasks for which it is to be used. Bearing these requirements in mind, this paper outlines the basic nature of performance support and then describes a generic model that can be used to facilitate EPSS development. Use of this model is supported by the use of a collection of performance support guidelines that are summarized in this paper in the form of basic principles of performance support. Use of the model and principles are illustrated using a case study approach.

Complex problem solving in naturalistic environments is fraught with uncertainty, which has significant impacts on problem-solving behavior. Thus, theories of human problem solving should include accounts of the cognitive strategies people bring to bear to deal with uncertainty during problem solving. In this article, we present evidence that analogy is one such strategy. Using statistical analyses of the temporal dynamics between analogy and expressed uncertainty in the naturalistic problem-solving conversations among scientists on the Mars Rover Mission, we show that spikes in expressed uncertainty reliably predict analogy use (Study 1) and that expressed uncertainty reduces to baseline levels following analogy use (Study 2). In addition, in Study 3, we show with qualitative analyses that this relationship between uncertainty and analogy is not due to miscommunication-related uncertainty but, rather, is primarily concentrated on substantive problem-solving issues. Finally, we discuss a hypothesis about how analogy might serve as an uncertainty reduction strategy in naturalistic complex problem solving.

Rather than the future of organization design (academics make justifiably reluctant futurists), I want to discuss a possible future for the science of organization design – one that I hope will come to pass.I understand organization design to refer to a particular form of human problem solving in which the problem is one of getting multiple individuals with diverse knowledge and interests to collectively achieve something that they could not by acting individually. Because bounded rationality affects not only the members but also the designers of organizations, solutions may be imperfect and unsuccessful, and many may have arisen almost unintentionally. But clearly there are better and worse solutions, and given the predominantly organizational nature of our economy, many good solutions exist in the form of the organizations that surround us.

This paper presents a bibliography of 100 references related to human problem solving, arranged by subject matter. The references were taken from PsycInfo database. Journal papers, book chapters, and dissertations are included. The topics include human development, education, neuroscience, and research in applied settings, as well as animal studies.

AFS (Axiomatic Fuzzy Sets) -based formal concept is a generalization and development of classical concept lattice and monotone concept, which can be applied to represent the logic operations of queries in information retrieval. Granular computing is an emerging field of study that attempts to formalize and explore methods and heuristics of human problem solving with multiple levels of granularity and abstraction. The main objective of this paper is to investigate and develop AFS-based formal concept by using granule logics. Some generalized formulas of granular computing are introduced, in which AFS-based formal concept and AFS-based formal concept on multi-valued context are interpreted from the point of granular computing, respectively.

Artificial intelligence and data mining techniques offer a chance to make education tailored to every student. One of possible contributions of automated techniques is a selection of suitable problems for individual students based on previously collected data. To achieve this goal, we propose a model of problem solving times, which predicts how much time will a particular student need to solve a given problem. Our model is an analogy of the models used in the item response theory, but instead of probability of a correct answer, we model problem solving time. We also introduce a web-based problem solving tutor, which uses the model to make adaptive predictions and recommends problems of suitable difficulty. The system already collected extensive data on human problem solving. Using this dataset we evaluate the model and discuss an insight gained by an analysis of model parameters.

Metaphor graphics are data displays designed to look like corresponding variables in the real world, but in a non-literal sense of “look like”. Evaluation of the impact of these graphics on human problem solving has twice been carried out, but with conflicting results. The present experiment attempted to clarify the discrepancies between these findings by using a complex task in which expert subjects interpreted respiratory data. The metaphor graphic display led to interpretations twice as fast as a tabular (flowsheet) format, suggesting that conflict between earlier studies is due either to differences in training or to differences in goodness of metaphor, Findings to date indicate that metaphor graphics work with complex as well as simple data sets, pattern detection as well as single number reporting tasks, and with expert as well as novice subjects.

Two problem-solving experiments investigated the relationship between planning and the cost of accessing goal-state information using the theoretical framework of the soft constraints hypothesis (Gray & Fu, 2004; Gray, Simms, Fu, & Schoelles, 2006). In Experiment 1, 36 participants were allocated to low, medium, and high access cost conditions and completed a problem-solving version of the Blocks World Task. Both the nature of planning (memory based or display based) and its timing (before or during action) changed with high goal-state access cost (a mouse movement and a 2.5-s delay). In this condition more planning before action was observed, with less planning during action, evidenced by longer first-move latencies, more moves per goal-state inspection, and more short (≤0.8 s) and long (>8 s) “preplanned” intermove latencies. Experiment 2 used an eight-puzzle-like transformation task and replicated the effect of goal-state access cost when more complex planning was required, also confirmed by sampled protocol data. Planning before an episode of move making increased with higher goal-state access cost, and planning whilst making moves increased with lower access cost. These novel results are discussed in the context of the soft constraints hypothesis.

Pragmatists, most notably John Dewey and Richard Rorty, propose overcoming the modern split between science and values with a new image of ourselves as language users. In this new self-understanding, both our scientific and evaluative vocabularies are integral parts of self-reliant human problem solving and coping with the larger natural environment. Our language is not the medium of any higher power from which it derives its legitimacy. On this view, the principal matter at issue between pragmatists and realists so far as interaction between religion and science is concerned is the moral one of human self-reliance.

Formally, the orthodox rational agent's “Olympian” choices, as Simon has called orthodox rational choice, are made in a static framework. However, a formalization of consistent choice, underpinned by computability, suggests by, satisficing in a boundedly rational framework is not only more general than the model of “Olympian” rationality, it is also consistently dynamic. This kind of naturally process-oriented approach to the formalization of consistent choice can be interpreted and encapsulated within the framework of decision problems—in the formal sense of metamathematics and mathematical logic—which, in turn, is the natural way of formalizing the notion of Human Problem Solving in the Newell-Simon sense.

This paper presents a bibliography of more than 200 references related to human problem solving, arranged by subject matter. The references were taken from PsycInfo database. Journal papers, book chapters, books and dissertations are included. The topics include human development, education, neuroscience, research in applied settings, as well as animal studies.

"Practice wisdom" continues to be a meaningful concept for social work. Herbert Simon's work on bounded rationality, Donald T. Campbell's evolutionary epistemology, and Donald Schon's descriptive work on reflective practice are combined to explain how a social worker's "practice wisdom" evolves from a problem-solving practice model built upon practice and life experiences.

The study focuses on the identification of the underlying representational properties of human problem solving and their application to expert systems. In this study the interaction between problem representation (procedural, conceptual, unstructured) and problem type (transformation, arrangement, inducing structure) was observed. The results of this study indicate partly that quantitative and qualitative differences in problem-solving performance can be attributed to the form of knowledge representation employed by the problem solver. It is suggested that modularized expert systems could be designed with different problem-solving modules organized by problem characteristics or type, exploiting the representational differences identified in this study.

One-hundred and nine adult subjects were tested on two occasions using a set of computer administered letter series problems. Using methods introduced by Furneaux [In Eysenck, H. J. (Ed.), Handbook of abnormal psychology . New York: Basic Books, 1961], response times were analysed in order to provide estimates of item difficulties based on time to correct solution, rather than on probability of correct solution. Subject parameters of speed, continuance and accuracy were also calculated using both Furneaux (directly and indirectly) and classical methods. These ‘time based’ parameters were compared with subscale scores from the WAIS-R IQ test and parameters from two reaction time tasks: 1 and 2-bit choice reaction time and the ‘Odd-Man-Out’ task. It was concluded that the original Furneaux model performed sub-optimally with regard to parameter estimation. However, with some modifications made to these estimation procedures, the Furneaux parameters were shown to be both practical and useful estimates of subject performance on letter series problems.

意思決定において人間-コンピュータ協調型の問題解決を行うことが多くなってきた. 本研究ではこのような問題解決において, 人間の処理を軽減し, 効率のよい問題解決が実行できるような画面表示形式を提案する. 問題解決を理解プロセスと探索プロセスに分け, これらのプロセスに影響を与える要因として, 表示自由度と位置表現性の2つを取り上げ, これら2つの要因の異なる画面表示形式に対して実験を実施した. その結果, 理解プロセスは位置表現性に関係し, 探索プロセスは位置表現性と表示自由度の両方に関係することがわかった.

Although human problem solving has been investigated in a behavior based approach, it has been recognized that ignoring what goes on in human brain and focusing instead on behavior has been a large impediment to understand how human being does complex adaptive, distributed problem solving and reasoning. In the paper, we propose a methodology for investigating human problem solving process by combining ERP mental arithmetic tasks, as a case study, with multi-aspect data analysis. Preliminary results show the usefulness of our methodology.

This paper presents a bibliography of a little more than 100 references related to human problem solving, arranged by subject matter. The references were taken from PsycInfo and Compendex databases. Only journal papers, books and dissertations are included. The topics include human development, education, neuroscience, research in applied settings, as well as animal studies. A few references from artificial intelligence are also given.

The objective of this paper is to present a series of proposed cognitive models for specific components of design ideation. Each model attempts to explain specific cognitive processes occurring during ideation. Every model presented here is constructed with elements (i.e. cognitive processes) and theories available from cognitive psychology, human problem solving, mental imagery, and visual thinking. Every model in turn is an element of a higher-level cognitive model of design ideation. These models provide a better understanding of the components involved during ideation and their relationships.Copyright © 2007 by ASME

Several psychophysical studies on human problem solving were performed. These studies involved the following problems: the Traveling Salesman Problem, the 15-puzzle and variants of this puzzle with different sizes, and finally, the TSP with obstacles. All these problems are difficult combinatorial problems and are considered intractable. However, human subjects were found to produce near-optimal solutions very quickly. For all these problems, a pyramid algorithm was used as a model of the mental representation of the problem, and of the global-to-local process of producing the solution. This approach represents a paradigm shift in the study of human problem solving because the prior research (1) neglected mental representation of problems and (2) it concentrated on easy problems that are not solved well by humans. As a part of this project, a workshop on human problem solving was held at Purdue. After the workshop, a new journal was launched. This is the first journal dedicated to human problem solving.

Games are a field in which players are implicitly trained and compelled to solve hard problems optimally. By making an appropriate training experiment using a computer game, players can reach expert performance. A computer game tournament that contains Euclidian traveling salesman problem (ETSP) instances is developed and solutions made by expert players are analyzed with human-behavior hypotheses and data mining techniques. A model for the expert players' behavior that combines apparently different fields is proposed and tested with TSPLIB problems.

Future robots should have common sense about the world in order to handle the problems they will encounter. A large part of this commonsense knowledge must be naive physics knowledge, since carrying out even the simplest everyday chores requires familiarity with physics laws. But how should one start codifying this knowledge? What kind of skills should be elicited from the experts (each and every one of us)? This paper will attempt to provide some hints by studying the mental models of force and motion.

Diagrams often play an important role in human problem solving. A diagram can make a task more difficult, however, when it obscures important problem features, or requires repeated effort to interpret what it represents. Moreover, the nature and origin of diagrams can be as important as their exploitation during problem solving. This paper chronicles the complex interleaving of visual cognition with high-level reasoning in three subjects. Their diagrams and subsequent verbal protocols offer insight into human cognition. The diversity and richness of their response, and their ability to address the task via diagrams, provide an incisive look at the role diagrams play in the development of expertise. This paper recounts how their diagrams led, and misled, them; how their diagrams both explained and drove explanation; and how a spatial approach reported here can lead to deeper understanding of other simple games.

Prior theories have assumed that human problem solving involves estimating distances among states and performing search through the problem space. The role of mental representation in those theories was minimal. Results of our recent experiments suggest that humans are able to solve some difficult problems quickly and accurately. Specifically, in solving these problems humans do not seem to rely on distances or on search. It is quite clear that producing good solutions without performing search requires a very effective mental representation. In this paper we concentrate on studying the nature of this representation. Our theory takes the form of a graph pyramid. To verify the psychological plausibility of this theory we tested subjects in a Euclidean Traveling Salesman Problem in the presence of obstacles. The role of the number and size of obstacles was tested for problems with 6-50 cities. We analyzed the effect of experimental conditions on solution time per city and on solution error. The main result is that time per city is systematically affected only by the size of obstacles, but not by their number, or by the number of cities.

As mentioned before, DPlan is intended as a tool to support the generation of finite programs. That is, plan generation is for small domains - with three or four objects - only. Generation of a universal plan covering all possible states of a domain, as we do with DPlan, is necessarily a complex problem because for most interesting domains the number of states grows exponentially with the number of objects. In the following, we first discuss what extensions and modifications would be needed to make DPlan competitive with state of the art planning systems. Afterwards, we discuss extensions which would improve DPlan as a tool for program synthesis. Finally, we discuss some relations to human problem solving.

This paper presents a new formalism for the design of automated manufacturing systems (AMS). This formalism is based on results from artificial intelligence and cognitive psychology works on human problem solving strategies. It is described in a validation context which is a small tooling model that allows to put forward its assets. A computer support of this formalism is also presented. >

Presents a small survey of commonly used design methods applied to the field of automated manufacturing systems (AMS). Putting forward the lack of methods really suited to the field, the authors then introduce a new specific formalism. This formalism is based on results from artificial intelligence and cognitive psychology works on human problem solving strategies. It is described in a validation context. Some extensions to the formalism are then presented. These ones allow a more formal description of the AMS in order to ease the transition toward the realization of the system. >

Which are the fundamental principles of life? This is the main question to be addressed if one tries to create artifcial life on computers. Though it has been answered only partially, evolutionary algorithms are substantially contributing already to many kinds of human problem solving by means of virtual organisms. Besides looking back on that success story and extrapolating it a bit into the future - both endeavors obviously being subjective -, a new result will be presented in the following showing the importance of multicellularity, which helps to self-adapt the error rates of the replication step to what is needed for eficacious and efficient optimum seeking without individual learning., Reihe Computational Intelligence ; 18

The term ‘everyday problem solving’ signals the importance of environmental context in determining human problem solving processes. The founders of the cognitive approach to problem solving, Newell and Simon (1972) recognized the importance of context, and incorporated an analysis of context as a central construct of their framework (under the name ‘task environment’). Nevertheless, the puzzles studied by Newell and Simon, and in the majority of the experimental literature on human problem solving have tended to strip away some important aspects of context. However, there has been a major effort to study richer contexts, including, in particular, the use of external representations and artifacts to support problem solving. This article will briefly review some of the theoretical frameworks which have been put forward to underpin this enterprise. The emerging importance of a rational perspective in problem solving will be outlined.

Die „Logisierung des Denkens“, die mit der Computertheorie des Geistes verbunden war, speiste sich nicht aus empirischen Befunden, sondern aus apriorischen Uberlegungen, die uber den Umweg der automatischen Beweiser in die Kunstliche Intelligenz und die Kognitionswissenschaft eingedrungen waren. Man kann diese Entwicklung beispielsweise anhand der Biographien von Allen Newell und Herbert Simon verfolgen.1 Newell und Simon stellten 1956 den „Logical Theorist“ vor, einen automatischen Beweiser, der mit heuristischen Mitteln „similar to those that have been observed in human problem solving“ Beweise generiertes und die Grundlage des spateren „General Problem Solver“3 wurde. Bei Zenon Pylyshyn wird diese „Logisierung“ sogar explizit als Grundlage der Computertheorie des Geistes genannt: „... many believe that the discrete, languagelike nature of the postulated carriers of thought arises by false analogy with the discrete symbol structures of computers, and thus object to them on this ground alone. They point out that to adopt metaphors from current technology is to ensure the rapid obsolescence of ones ideas. The wax-tablet metaphor of the ancients gave way to the lever-and-gear metaphor of the eighteenth century, which gave way to the telephone-switchboard metaphor of the early twentieth century. Now we have the computer-software metaphor with its discrete languagelike, symbolic expressions. The sentential-representation format has little to do with actual computing devices, however; rather, as has been mentioned, it is more closely related to formalist ideas in mathematics and logic. These ideas, which form the basis of symbolic logic, are the only scheme we have for capturing normative patterns of reasoning.“4

In this paper we examine a previously published algorithm which addresses the problem of network growth by implementing a clustering algorithm to operate on time dependant data. The computational constraints of the problem forced the development of an architecture, which in retrospect can be analysed in terms of a computational and symbolic module operating symbiotically. Here we attempt to identify the computational constraints that necessitate the use of this architecture, and any further merits it has. Further, we analyse the nature of the interaction between the two modules and highlight the manner in which the behaviour the symbiotic modules correlates with what is known of human problem solving behaviour.

Human problem solving involves search. Therefore, to simulate human problem solving in a computer we need to develop algorithms for search. In this chapter we will lay a foundation for later chapters on advanced problem solving techniques by discussing basic search methods. Before a formal definition of problem solving is provided, we will discuss some concepts with a small, yet illustrative, example.

Study on building powerful and robust problem solver for ITSs is described. The proposed problem solver is based on parallel computing technology to simulate some of the intelligent characteristics of human problem solving including the concurrency of thought, the case based reasoning, and the awareness of the key to problem solving. We propose the multiagent problem solver which divides a given problem into several small sub-problems and solves them concurrently. This system is especially useful for ITS which teaches procedural problem solving to novice students, because it is capable of showing the reason of achieving an idea for solution by thinking aloud by the “trial-and-error” method.

This paper aims at reviewing several results concerning the development of practical Decision Support Systems (DSS) for real time production control in manufacturing. Several characteristic features of DSS that are highly demanded and/or accepted by human factors and mimic the human problem solving are identified. The perspective includes technology, business and human factors. A practical DSS illustrates the approach.

Rather than replacing a human problem solver, computers are increasingly seen as intelligent aids to problem solving. In this role the computer’s capacity for storing and rapidly retrieving large amounts of information can potentially significantly augment human problem solving. In the specific context of design problem solving, this approach has particular attractions when associated with recent advances which allow the inclusion of graphic as well as textural material. Rather than simply providing information about principles and descriptions of examples of their use to solve problems, computers could generate sets of plans, perspective drawings and even sets of pictorial representations of actual design solutions to particular types of problems. However, while this may make such systems very attractive to designers, this very richness of information may produce a significant problem for design. Design fixation involves the reproduction of both appropriate and inappropriate aspects of an example design when the example solution is shown as part of the statement of the design problem. The results of the experiment to be reported indicate that the fixation effect does not simply depend on the pictorial representation of a possible solution to a problem. Rather fixation depends on the picture embodying principles which form a part of the knowledge base of the design discipline. It is likely that the cases representing previous solutions to a problem contained in an intelligent design aid would predominantly be of this form and could therefore establish the conditions for fixation to occur. While this could in fact be beneficial where routine design is involved, it places severe constraints on innovative or creative problem solving.

Simplification by effective structuring of information in order to reduce mental effort is one basic component of human problem solving. This simplification depends on the ability to find representative structures in the learning material. The more effective the structures, the more effectively can problems be solved. Similarly uneffective structuring should demand higher cognitive effort to solve problems. Cognitive effort may change the basic rhythm of the electroencephalogram (EEG) (Creutzfeld et al. 1969). These changes can be found at different locations depending on the brain structures involved during different kinds of activation. Topographical differences can be described by the EEG mapping method. Thus, in the first part of the study, we examined whether ineffective structuring by untrained persons is accompanied by more EEG desynchronization than effective structuring by trained persons.

Many critical real world human problem solving situations take place in dynamic event-drivers environments, where the evidence comes over time and situations can change rapidly. In these situations people must amass and integrate, uncertain, incomplete and changing evidence. A major source of human error in dynamic domains seems to be a failure to revise situation assessment as new evidence comes in. This paper will be concerned with the identification of the main descriptive patterns of fixation errors and with how to build new sytems to reduce this type of error. It will also begin the process of building a theory of fixation errors.