Marchionini Electronic Browsing Model
1. MODELS OF INFORMATION BEHAVIOR/INFORMATION SEEKING BEHAVIOR 1. 1 Introduction of Information seeking According to Marchionini, information seeking is a “process driven by life itself” (Marchionini, 1995). “A process in which humans purposefully engage in order to change their state of knowledge” (Marchionini, 1995). Information seeking is an essential and ongoing activity where humans seek to address their needs through the information seeking process.
In the context of the electronic environment, Marchionini identifies the “problem” as the kernel of information seeking; to address the problem, the seeker engages in search tasks in systems, whose outcomes are organized and contextualized by domains and settings (Marchionini, 1995). Notably, information seeking occurs in a wide range of environments, both online and offline (in earlier formulations, electronic or non-electronic). 1. 2 Model of Information Seeking
The influence of new technology on information seeking is also providing a new set of alternative models that more accurately describe the information seeking process as a dynamic activity. Models of information seeking attempt to describe the process a user follows to satisfy an information need. The information seeking models in this section focus on the behavior of information seeking activities. There are a large number of models of information behavior (Fisher et al. , 2005).
These models demonstrate the difficulty of representing information needs; they also explain how the information seeking process reduces uncertainty regarding the information need. These models are Kuhlthau’s Information Search Process, Ellis’ Behavioral Model of Search, Wilson’s Model of Information Behavior, Sonnenwald’s Information Horizons, Allen’s Information Tasks (1996) and others. 1. 3 Issues on information seeking model According to David Johnson (1997), theoretical models of information seeking must address three key issues.
First, models should provide a sound theoretical basis for predicting changes in information seeking behavior. Second, models should provide guidance for designing effective strategies for enhancing information seeking. Third, models should explicitly conceptualize information seeking behavior, developing rich descriptions of it. Finally, models should answer the “why” question; they should explicitly address the underlying forces that impel particular types of information seeking. 2. BACKGROUND OF MARCHIONINI 2. 1 Biography of Marhionini
Gary Marchionini is Dean and Cary C. Boshamer Professor in the School of Information and Library Science at the University of North Carolina where he serves as Dean of the school. He formerly taught courses in human-information interaction, interface design and testing, and digital libraries. He heads the Interaction Design Laboratory at SILS. Marchionini earned a doctorate in curriculum development, focusing on mathematics education, and a master’s degree in secondary mathematics education from Wayne State University in 1974 and 1981, respectively.
He graduated with a bachelor’s degree in mathematics and English from Western Michigan University in 1971. Before arriving at UNC, he was a faculty member at the University of Maryland for 15 years. He served on the faculty and as a researcher at Wayne State from 1978 to 1983 and taught mathematics at the East Detroit Public Schools for seven years. His research interests are on information seeking in electronic environment, human-computer interaction, digital libraries, information design and information policy.
Marchionini received a Google Faculty Research Award in 2010 to support work on banner blindness in web search (with Felix Portnoy) He received an IBM Faculty Research Award for 2006-07 to work on digital video surrogate creation and metadata evaluation. He also received a Google Research Award to develop the Information in Life Video Series (2007-08) for the UNC-CH YouTube Educational Video Channel. He is author of a books titled Information Seeking in Electronic Environments published by Cambridge University Press, and Information Concepts: From books to cyberspace identities published by Morgan-Claypool.
INFORMATION SEEKING MODEL BY MARCHIONINI 3. MARCHIONINI’S SEEKING PROCESS MODEL (1995) 3. 1 Introduction The information seeking model put forth by Marchionini closely describes a formal online search from a cognitive perspective, which we found useful to articulate the general components of such a search at a high level. The model is meant to provide an overview of the online information seeking process, rather than a detailed representation of the search itself. Marchionini’s model outlines some of the types of knowledge that are necessary for the information seeking process.
For example, the knowledge that needs to be applied to choose an appropriate search system is clearly distinct from the knowledge that one needs to formulate an effective query. Marchionini’s model is eight step information seeking framework that begins with the recognition and acceptance of the problem and continues until the problem is resolved or abandoned. It composed of the following sub processes:- 3. 2 Steps of information seeking framework First: Recognize and accept an information problem. Recognizing and accepting an information problem can be internally motivated or externally motivated.
The problem may be characterized as a gap (Dervin, 1977), a visceral need (Taylor, 1962), an anomaly (Belkin, 1980), as a defect in a mental model, or as an unstable collection of noumenal clouds, but it is manifested as a resource demand on the perceptual or memory systems–the person becomes “aware” of the problem. At this point, the problem may be suppressed or accepted. Suppression is influenced by setting and the information seeker’s judgment about the immediate costs (physical and mental) to initiate search.
In the case where the information seeker judges the situation appropriate, they accept the problem and begin to define it for the purposes of search. Acceptance is influenced by knowledge about the task domain, by the setting, by knowledge of search systems, and by the information seeker’s confidence in his/her personal information infrastructure. Recognition and acceptance are typically ignored by system designers since they are viewed as user-specific and uncontrollable.
However, systems that invite interaction and support satisfying engagement will lead users to accept information problems more readily. Attention to this subprocess also reinforces user control and volition in any intellectual activity. Problem acceptance initiates problem definition. Second: Define and understand the problem Problem definition is a critical step in the information-seeking process. This subprocess remains active as long as information seeking progresses. Understanding is dependent on knowledge of the task domain and may also be influenced by the setting.
The cognitive processes that identify key concepts and relationships lead to a definition of the problem that is articulated as an information-seeking task. For intermediated information seeking, the intermediary conducts a reference interview to accomplish this subprocess (Auster & Lawton, 1984; White, 1985). In end-user searching this step is often assumed or abbreviated–a major cause of end user frustration and failure. To understand and define the problem, it must be limited, labeled, and a form or frame for the answer determined.
The problem may be limited by identifying related knowledge or similar problems or by listing what specific knowledge is not related. Concepts, words, phrases, events, or people related to the problem can be listed and grouped into categories that serve as the basis for assigning labels and problem statements. This process represents what Taylor called the conscious need. The information seeker may hypothesize what the answer will be, but at least creates an expectation of what the answer will “look like”.
There may also emerge an expectation of the physical form of the answer which, in turn, strongly influences the selection of a search system. These expectations about outcomes ultimately guide action. The limiting, labeling, and framing of solution properties lead to the articulation of an information-seeking task, what Taylor referred to as the formalized need. During problem definition the information seeker represents the problem internally as a task with properties that allow progress to be judged and determines a general strategy to use for subsequent steps. Third: Choose a search system
Choosing a search system is dependent on the information seeker’s previous experience with the task domain, the scope of his/her personal information infrastructure, and the expectations about the answer that may have been formed during problem definition and task development. Domain knowledge is a powerful variable in selecting a search system and focusing search. Experts in a domain have experience with the primary search systems specific to the domain. Economists in our studies were able to make spontaneous judgments about whether information required for assigned information problems was likely to be found in one journal or another.
Likewise, attorneys were readily able to determine whether information in their assigned searches would be found in case law, statutes, or treatises. In both these cases, some professional intermediaries who regularly conducted searches in these domains were also able to predict where relevant information would be found (Marchionini, Dwiggins, Katz, ; Lin, 1993). Given the constraints of domain knowledge, general cognitive conditions, and previous search experience, information seekers endeavor to map the search task onto one or more search systems.
The mapping process takes into consideration the type of task, and characteristics of available or familiar search systems. In actual practice, information seekers consult several search systems as they move toward solutions to their problems. For example, in libraries, information seekers may ask a reference librarian where to begin searching, they may consult an index or a card catalog, and eventually one or more journal or book primary sources. As electronic search systems and network access proliferate, there are a plethora of potential sources available to information seekers.
It is becoming increasingly important to use secondary systems to limit the time and effort spent locating and using primary systems. With a few exceptions today’s electronic systems are specific to one or two particular levels of search rather than providing a common interface to many levels of systems. For example, there are expert systems that emulate a reference service, thousands of online bibliographic databases, and hundreds of online or CD-ROM full-text databases. Filtering, ordering, and selecting the collection of sources will become increasingly important to mapping tasks onto search systems.
Fourth: Formulate a query Query formulation involves matching understanding of the task with the system selected. In many cases, the first query formulation identifies an entry point to the search system and is followed by browsing and/or query reformulations. Query formulation involves two kinds of mappings: a semantic mapping of the information seeker’s vocabulary used to articulate the task onto the system’s vocabulary used to gain access to the content; and an action mapping of the strategies and tactics the information seeker deems best to forward the task onto the rules and features the system interface allows.
Semantic mapping is similar to moving from Taylor’s formalized need to the compromised need (1962) and is highly influenced by earlier mappings from the sensation that causes attention to a problem and the mappings during problem definition from fuzzy noumena and general concepts to specific terms and concept classes. In general, this mapping takes as domain the entire set of identifiers available to an individual information seeker, and the complete set of identifiers available to a system as range.
The mapping function most commonly takes words associated with the task onto the set of words that serve as entry points to the system content. For static search systems such as books, the information seeker has total control for the mapping and aims to match words/phrases from the task statement itself, with words/phrases in the title, index, table of contents, headings, list of keywords, and text. For dynamic search systems such as people, the intelligence of both parties can be applied to enrich the mapping function since the controlled vocabulary of a human is both large more associationally connected.
Thus, experts in a domain not only know more terms that directly relate to the information seeker’s query formulation, but they can also add additional terms and interact with the information seeker to clarify and verify the query. In the case of professional intermediaries, the process of developing a query formulation is part of the reference interview and has been shown to be an important determinant of intermediary performance. In the case of electronic search systems, the query formulation process is partially dynamic and there are a wide range of techniques system designers have used to assist the information seeker.
Such techniques include: expert system intermediaries (Croft & Thompson 1987; Marcus, 1983); online suggestions (Meadow, 1988); query-by-example (Zloof, 1977); dynamic queries (Shneiderman, 1992); and hypertext (Croft & Turtle, 1989; Frisse, 1988; Marchionini & Shneiderman, 1988). An electronic system may have a strictly controlled vocabulary or a full-text vocabulary, each clearly affecting the cardinality of the resultant set of items retrieved as a result of applying a mapping.
The problem of representing concepts in document sets is a fundamental problem in information science and is considered from several perspectives in subsequent chapters. Action mappings take possible sets of actions to the inputs a search system can recognize. If semantic mappings are thought of as “what” or declarative in nature, then, action mappings are “how” or procedural in nature. Just as a search system constrains the vocabulary an information seeker may use in query formulation, search systems also constrain how queries may be expressed.
For example, humans recognize spoken or written expressions but books do not, and electronic systems have not yet accomplished any but rudimentary or highly constrained recognitions. Electronic systems may support Boolean expressions and provide special syntax for how these expressions may be made. Electronic systems may allow users to enter any terms they wish, or provide a menu that completely specifies all possible terms, or provide traversable links among various partitions of the database.
At even more detailed levels, the system demands that users specify terms or previous sets using explicit characters, type cases, or punctuation. Fifth: Execute search Execution of the physical actions to query an information source is driven by the information seeker’s mental model of the search system. Execution is based on the semantic and action mappings developed during query formulation. Conducting lookup requires actions like articulating a question verbally, picking up a volume, or pressing a key.
For a card catalog, execution may entail selecting proper drawers and using alphabetical ordering rules; for an online database, execution may entail typing the query and sending it with a special keypress for a hypertext, execution may entail browsing the database by following available links provided by the author. Communication and computing technology has greatly affected how searches are executed by altering the physical actions necessary.
Phone calls, telefacimilies and electronic mail make execution of a search with a human search system much more feasible, and electronic networks allow direct queries of remote collections from a home or office. Although interfaces for these devices are often complicated and frustrating, the effects of executing information-seeking tasks in physically proximate space cannot be overestimated. Search execution is one of the most obvious changes wrought by electronic environments since information seekers perform much more constrained physical actions at workstations than they do in libraries or offices. Sixth: Examine results
Executing a query results in a response from the search system. This response is an intermediate outcome and must be examined by the information seeker to assess progress toward meeting the goal of the information-seeking task. This examination is dependent on the quantity, type, and format of the response and involves judgments about the relevance of information contained in the response. Responses are provided by information systems in units specific to the type of database, for example, numeric values, bibliographic records, fixed-length fields, entire documents, specific images, or verbal expositions on a topic.
A response to a query may contain zero, one, a few, or many of these units, often referred to as “hits”. The information problem and the user’s personal information infrastructure cause the information seeker to have expectations about the number of units required to complete a task, although these expectations often change as information seeking progresses. For example, information seekers typically expect zero or one hit when searching a card catalog for a specific title, and zero, one, or many for a query about a topic.
Users of print encyclopedias typically expect to find zero, one, or a few articles on a topic, and may be quite surprised to find hundreds of hits when using a full-text electronic encyclopedia for the same topic. A significant difference in printed and full-text electronic encyclopedias is that electronic systems often retrieve many articles, thus requiring another major decision point in the examination of results (Marchionini, 1989). When multiple hits are returned, they are usually presented as a set made up of document surrogates such as titles, bibliographic records, or descriptive identifiers.
The way these sets are organized and presented affects how information seekers examine individual units, make relevance judgments, and decide what steps to take next. In a library, a set of catalog cards on a broad topic are ordered alphabetically according to the main entry for that document. In a set of bibliographic records retrieved from an online search system, the items are often ordered in chronological order beginning with the most recent. In more advanced electronic retrieval systems, items may be ranked according to query term frequencies.
In hypertext systems, explicit links to other units and implicit links such as next page, previous unit, or index lookups are provided by the database designer. The ordering of resultant sets becomes more important as the size of the set increases, and the ability to manipulate orderings of sets of items is recommended for all electronic search systems. The propensity of electronic systems to report large sets of documents significantly affects the examination of results subprocess, complicating the decision-making associated with selecting relevant items of information.
The information seeker must judge the relevance of individual retrieved units with respect to the information-seeking task at hand. Relevance is a central theme of information science and has been considered from both theoretical and practical perspectives. Cooper (1971) provided a definition of logical relevance as a formal basis for evaluation of retrieval systems, and Wilson (1973) described situational relevance as dependent on the particular information problem at hand. Situational relevance is more specific to the relevance judgments that information seekers make as they examine intermediate results of search.
From a practical perspective, relevance serves as the main criterion measure for computing performance measures such as recall and precision. From an information seeker’s perspective, relevance may be considered as a decision on what action to take next in the information-seeking process. Alternatives include: terminate search due to goal achievement; pursue the document more fully, pursue the document later, and location and continue examining other results; pursue implications of the document to the continuation of search and either continue examining other results in this iteration, formulate a new query, or redefine the roblem; reject the document completely and continue examining results; or reject the document and stop information seeking without accomplishing the task. The examination of specific items for relevance is obviously affected by the type and the quantity of information in the retrieved set. For small sets of results, items can be scanned quickly, browsed systematically, or inspected comprehensively. For large sets of results, the set may be reduced by reformulating the query or semantically-related surrogates be scanned to identify those that suggest more comprehensive relevance assessment.
Marchionini and his colleagues have argued that information seekers are willing to scan substantial sets of textual or graphic documents if they are given appropriate display and control mechanisms (Marchionini, 1989; Liebscher ; Marchionini, 1988). As with query formulation, electronic systems have made substantial progress in supporting examination of results. Ranked output and alternative orderings of output offer substantial advantages to experienced information seekers because they assist in managing large result sets.
Display techniques such as highlighting query words in retrieved documents, presenting different levels of organizational details (e. g. , table of contents and full text–Egan et al, 1989), fisheye views that cluster potentially relevant items in a spatially ordered manner (Furnas, 1986), and high-resolution graphic views of information in hierarchical displays (e. g. , Card, Robertson, & Mackinlay, 1991). Seventh: Extract Information There is an inextricable relationship between judging information relevant and extracting the relevant information for all or part of the problem solution.
Assessments about relevance cause information extraction actions to be taken, although, information can be relevant to the problem, but not fully meet the conditions of the task goal. If a retrieved document is judged relevant, the information seeker may choose to continue assessing its relevance by extracting and saving information or to defer extraction and continue examining results. In the latter case, the document will eventually be re-examined and a revised relevance assessment made based on what other documents were added to the relevant list and what experiential events the nformation seeker experienced since the previous relevance judgment. To extract information, an information seeker applies skills such as reading, scanning, listening, classifying, copying, and storing information. In the case of secondary databases, extraction may entail copying or printing bibliographic citations to facilitate retrieval of actual documents. In the case of verbal questions to human experts, listening skills, clarification requests or restatements of the information in one’s own words aid in extracting the information relevant to the task.
In full-text systems, basic reading skills, scanning skills, use of structural features such as headings and outlines, and jumping from section to section aid in extracting relevant information. As information is extracted, it is manipulated and integrated into the information seeker’s knowledge of the domain. As more information is extracted and stored, new items may not be as relevant as they would have been previous to other items being manipulated and integrated.
Information extraction often includes some physical action such as copying to paper or other medium and saving those copies in larger structures according to well-defined organizational rules. For electronic systems, some of the techniques for ordering and display mentioned in the previous section on examination will assist users by allowing them to cut and paste items easily, including the contextual components that may appear in other windows on the screen.
Thus, saving a section outline from a table of contents, the paragraphs around the most relevant information, and a path or query statement that retrieved the document can all be extracted and aggregated easily. Electronic tools for cutting and pasting already offer substantial advantages for information extraction of text, static and moving images, and sound. Eight: Reflect/iterate/stop An information search is seldom completed with only a single query and retrieved set.
More often, the initial retrieved set serves as feedback for further query formulations and executions. Deciding when and how to iterate requires an assessment of the information-seeking process itself, how it relates to the acceptance of the problem and the expected effort, and how well the extracted information maps onto the task. Monitoring the progress of information seeking is particularly crucial to browsing strategies that are highly interactive and opportunistic.
Determination of a stopping function may depend on external functions like setting or search system, or on internal functions like motivation, task domain knowledge, and information-seeking ability. Stopping decisions in full-text electronic systems are more complex because retrieval is both physically easier and yields more robust outcomes. FIGURE 1 Information-seeking process model (based on Marchionini, 1995) INFORMATION SEEKING MODEL BY MARCHIONINI 4. MARCHIONINI’S ELECTRONIC BROWSING MODEL (1995) 4. 1 Introduction of browsing
The influence of new technologies, notably the World Wide Web with its search engines and navigational links is changing what scholars mean when referring to information seeking. Marchionini (1995) begins to distinguish information seeking as a superset of activities including information retrieval and browsing. He contends that “the term information seeking is preferred to information retrieval because it is more human oriented and open ended”. He points out that information retrieval assumes the information sought has been known at some point, while seeking information can cover this activity and beyond.
In most cases, learning about a new topic is involved, meaning that the information was not known, it may not be discovered in one measurable activity and may eventually end up not satisfying the initial answer to the question. Browsing is the other component of Marchionini’s idea information seeking, an accompaniment to information retrieval. With the new graphical browsing applications, most notably web browsers, information seeking can be more fluid and less structured than traditional information retrieval processes. Chun, Detlor, Turnbull, 2000) 4. 2 Why Do People Browse Marchionini (1987) discusses three primary reasons why people browse: * First, they browse because they cannot or have not defined their search objective; they have what Belkin, Oddy and Brooks have called anomalous states of knowledge (1982). * Second, people browse because it takes less cognitive load to browse than it does to plan and conduct an analytical, optimized search. * Third, people browse because the information system supports and encourages browsing.
Particular information sources like encyclopedias invite browsing by supplying indexes, outlines, section headings, tables and graphs, which help users quickly filter information. (Hildreth, 1995) 4. 3 Modes of Browsing and Searching Marchionini (1995) reviewed the research on browsing and observed that “there seems to be agreement on three general types of browsing that may be differentiated by the object of search and by the systematicity of tactics used. ” Directed browsing occurs when browsing is systematic, focused, and directed by a specific object or target.
Examples include scanning a list for a known item, and verifying information such as dates or other attributes. Semidirected browsing occurs when browsing is predictive or generally purposeful: the target is less definite and browsing is less systematic. An example is entering a single, general term into a database and casually examining the retrieved records. Finally, undirected browsing occurs when there is no real goal and very little focus. Examples include flipping through a magazine and “channel-surfing. ” (Chun, Detlor, Turnbull, 2000) 4. 4 Stages of Electronic Browsing Model
Figure 2 Electronic Browsing Model First: Scanning Scanning is the most basic browsing strategy. It is fundamentally a perceptual recognition activity that compares sets of well-defined objects to an object that is clearly represented in the information seeker’s mind. Scanning benefits from highly organized environments that provide clear and concise representations. It can proceed in sequential fashion, according to some structural feature of the content, or through some sampling method. Two such tactical approaches are linear and selective scanning.
Linear scanning applies perceptual attention in continual and sequential fashion. For example, information seekers may scan title lists to identify potentially relevant documents to examine, or may fast-forward through television channels to locate an interesting program. Linear scanning is most applicable to sequential arrangements of similar objects with precise attributes that are recognizable with a single glance, and when the collection is reasonably small or the information seeker is confident about being in the general neighborhood that contains the object.
Linear scanning is effective since the eye can recognize simple patterns in as little as 50 milliseconds, although eye movement plus recognition times vary from 125 to 500 ms. (Potter & Levy, 1969). An alternative to linear scanning is selective scanning. This tactic applies perceptual attention according to either an inherent or imposed stratification of the search space. Most search systems offer some inherent partitions that support selective sampling. For example, information seekers may scan the section headings within a book or the reference lists of different journal articles.
This tactic is particularly useful to gain an overview of content or to identify a neighborhood for linear scanning. A variation of this tactic is to scan selectively by sampling the search space according to a partitioning rule. Rather than depending on some features or organizations in the search space, random or purposive samples of the search space are identified and scanned for relevance. For example, information seekers may fast-forward a videotape 100 feet at a time to locate a particular sequence, or may examine every first sentence of each screen of text to get a sense of content.
The partitioning rule may take advantage of spatial memory and may be particularly useful in locating objects previously seen in the system. This tactic could be applied to gain an unbiased overview of database coverage or to make determinations of accuracy or integrity. Sample scanning may also be appropriate in discovering promising information neighborhoods in extremely large, unstructured databases. Scanning tactics are most applicable to organized search spaces of reasonable size.
What constitutes “reasonable” is often underestimated, especially in electronic environments, nonetheless one of the most obvious ways that electronic search systems can amplify human information-seeking abilities is by emphasizing and facilitating scanning tactics. Wiberley and Daugherty (1988) summarize literature related to library patron across-document scanning behaviors and suggest that users are willing to scan longer lists of citations from online searches than from OPAC searches and longer lists from electronic than manual systems.
They report ranges of 7 to 50 items scanned with few above 20 for manual systems, but occasionally up to 200 for electronic systems. They also note that poor interfaces in electronic systems minimize scanning persistence. It seems plausible that if systems are specifically designed to support scanning rather than simply permit it, substantial lists of references will be scanned by users. There is little evidence on how users scan within electronic documents, although simple affordances such as highlighted query terms and indicators of progress through the document clearly assist within document scanning.
Scanning tactics are most useful for tasks where recognizable and discrete attributes are available. Scanning tactics are most often applied during systematic browsing strategies or for intermediate examinations during opportunistic or casual browsing. One measure of the costs of scanning tactics is attention time and research on scanning in electronic environments will be well-served by analyses of these costs as influenced by different information-seeking factors. Second: Observation
Observational strategies are the most general of all browsing strategies in that they have minimal thresholds for all the browsing dimensions except for cognitive effort. Browsers who apply these strategies assume they are in a promising neighborhood and react to stimuli from that neighborhood. Observational strategies depend on a great deal of parallel input. In busy street scenes, a browser may attend to a variety of sounds, sights, and motions rather than systematically scanning the scene.
Advertisements in newspapers and magazines and provocative titles of books on shelves attract attention as we passively apply observational tactics. Like scanning, observational strategies are rooted in our physiological and psychological survival instincts and thus are naturally and easily applied as defaults. Observation does require interpretation and reflection to make sense of what is observed and to relate it to information-seeking objectives. Observations may lead to interesting discoveries but yield most initiation control to the environment.
For this reason, it is most important for the information seeker to be in a relevant neighborhood. Observations can be executed in systematic fashion but for ill-defined objects and fuzzy purposes, for example, regularly reading the morning newspaper or watching a news broadcast. Observation is the primary strategy used in opportunistic and casual browsing since it admits the widest range of objects and unorganized environments. Serendipitous observations are lauded as a benefit of browsing, but they occur as a result of reflection and association rather than simple perception and pattern matching.
Interfaces that aim to support observational strategies should provide alternative views of information. Because observational strategies yield significant control to the environment, information should be clearly represented and demarcated. Alternative representations should be available for user to transform so that pattern recognition may lead to association and reflection. Third: Navigation The navigation strategy balances the influence of the user and the environment. The environment constrains browsing by providing possible routes and the user exercises some control by selecting which routes to follow.
Navigation is a term that is used broadly and differently in the literature. In much of the human-computer interaction literature, navigation is considered as synonymous with browsing rather than one of several specific browsing strategies. One model that does distinguish browsing and navigation was proposed by Waterworth & Chignell (1991). In their model of information exploration, they distinguish querying from browsing, object specification from object recognition and navigation from mediation.
In their model, navigation is taken to mean a high degree of user control and mediation depends more on system control. As considered here, navigation is defined by relatively high thresholds for all the browsing dimensions. Objects must be specifable; information seekers must know what they are seeking, take an active role in interacting with the environment, and regularly reflect and make decisions about progress; and although navigation may occur in unstructured environments, high levels of organization greatly aid efficiency and effectiveness.
Physical navigation is often used as a metaphor for traversal of a hypertext or a database. The activity users engage in as they follow links in a hypertext is the navigation that we consider to be a browsing strategy. In this strategy, the process and the information found along the way are what is important. A better metaphor for this strategy may be “grazing” or “berry-picking” (Bates, 1989), but the activity is commonly called navigation in the human-computer interaction and hypertext literature and so is adopted here.
This type of navigation is in contrast to the goal-oriented physical navigation that uses cues from the environment only to check progress toward the goal. In physical navigation, the object sought is most often a predetermined destination and the process involves adjusting course according to attributes provided by the environment. Navigation connotes goal-oriented behavior where plans and subplans respond to the environment. This dynamic interplay between navigator and environment is what makes navigation a popular metaphor for human-computer interaction.
Navigation is a useful name for a browsing strategy in so far as it denotes observations of the environment and adjustments in behavior based on these observations. It is a weak metaphor for browsing in an information space in two respects. Most importantly, the destination for information seeking is seldom predetermined–intellectual space is highly amorphous and much of what is called navigation involves problem definition and clarification. Second, information-seeking activity draws information to the information seeker rather than transporting the information seeker to the information.
It is especially true in electronic environments that we remain physically stationary and gather information to our screen rather than travel to some location and idea. Using navigation as a global metaphor for browsing admits the side effects of becoming lost. In electronic systems this effect has been most often referred to as lost in hyperspace (Nielsen, 1990a). It is useful to distinguish being lost and confused as a result of the environment and being lost or confused as a result of the information problem.
Although it is likely that improved interfaces will minimize disorientation due to the system, being lost in a collection of thoughts will remain a human problem regardless of system advances. This point of view that understanding content is more critical than system orientation was reinforced by two investigations we conducted to determine the effects of blatant metaphors on learning. Two versions of a hypertext explanation of the information seeking framework were prepared using the GUIDE system. One version, called the “jump” implementation, reinforced the navigation metaphor.
The screen consisted of a single text window and link anchors were immediately followed by the word “jump” in bold upper case letters. Clicking on the link anchor replaced the text with the text in the linked node. The other implementation, called the “bring” version, reinforced the metaphor of bringing information to the screen rather than jumping to it. Link anchors were immediately followed by the word “bring” in bold upper case letters. Clicking on the link anchor opened a new window that overlapped the active window immediately below the anchor.
Navigation as a browsing strategy refers to the ongoing observation of environmental attributes, adjustments to the mental problem representation based on these observations, and resulting behavioral actions. Navigation is thus information seeking that proceeds incrementally based on feedback from the information system. The ways the system provides feedback is a critical factor in navigation. This strategy is most clearly supported by the structural layouts of museums. Navigation has become an important metaphor for electronic systems and is the primary inspiration for static hypertexts.
In one navigation variation, users are invited to follow paths through the database by selecting one of possibly many links from a current node to other nodes. Thus, the tactics used are simple selections from the choices active on the path. Navigation strategies can be applied in casual or systematic fashion, although they depend on taking advantage of existing links or the user’s ability to create new links. Navigation is an attractive compromise between user and system responsibility since the system invites or suggests links to follow but the user is free to choose from among many links or continue in linear fashion.
Different systems provide distinct levels of navigational freedom, ranging from highly systematic, mandatory paths to absolutely no guidance whatsoever. Successful systems fall somewhere in between, providing some choice to users but suggesting directions and providing informative cues about progress. Systems that provide explicit hypertextual links support more systematic browsing strategies and those that provide implicit links support more exploratory or casual browsing. Fourth: Monitor A monitor strategy is often applied in conjunction with systematic browsing or other primary activities such as reading.
Monitoring is most similar to scanning except it tolerates poorly structured environments. For example, while reading text related to a specific topic, a monitor browsing strategy “listens” for concepts related to another topic of interest. Professionals in any field often apply the monitor strategy automatically by spontaneously relating what they read in an unrelated area to their own field. It is this strategy that is perhaps most important for discovery and creative connections among disparate ideas, and is the reason scholars revere serendipity.
Although this strategy may be partly subconscious, it depends on the user making associations among concepts in the mind and representations in the information space. Monitoring is enabled by perspectives or views of the database, which in turn are made up of various cues such as words or phrases, movements, or visual characteristics, depending on the information environment. Monitor strategies are focused on attributes of interest to the information seeker and are less dependent on stimuli in the environment than observational and navigation strategies.
INFORMATION SEEKING MODEL BY MARCHIONINI 5. MARCHIONINI’S CONCEPT OF THE PERSONAL INFORMATION INFRASTRUCTURE Mental model(s) Events, etc. Search system Knowledge domain general Specific, e. g. info. seeking Cognitive skills Material resources Time, etc Money Equipment Documents Cognitive executive & attitudes FIGURE 3 Marchionini’s Personal Information Infrastructure 5. 1 Introduction Marchionini’s (1995) conception of a personal information infrastructure (PII) elaborates the relationship of the information seeker, the information system, and a range of contextual factors.
According to Marchionini, the personal information infrastructure is “a collection of interacting mental models for specific information systems; mental models for events, experience and domains of knowledge, general cognitive skill and specific cognitive skills related to the organizing and accessing information; material resources such as information systems, money and time; metacognitive resources and attitudes towards information seeking and knowledge acquisition. ” (Marchionini, 1995, p. 11).
An individual’s PII is the relationship he or she has with information technology, his or her mastery of the technology, and how the information sourced from the technology affects and shapes the individual. At the time of writing, ubiquitous information systems were nascent, but it is not hard to imagine our everyday life without ongoing informing interactions with technology. Marchionini has advanced the concept of the PII into the human information interaction perspective (Marchionini 2006; Marchionini 2008). (Stutzman)
An individual’s personal information infrastructure reflects an interaction of the individual’s cognitive abilities and skills, mental models of technologies, and resources available at hand. These factors are mutually shaped over time; as an individual gains greater master of an information technology, the mental model of the technology will shift, as well as cognitive reliance on the technology. Marchionini finds that an individual’s general cognitive ability,domain expertise, and systems expertise structure the information-seeking process.
To understand this interaction, we can study the information seeking behavior at the following levels of granularity: * Patterns “most often reflect internalized behaviors that can be discerned over time and across different information problems and searches” * Strategies are “sets of ordered tactics that are consciously selected, applied, and monitored to solve an information problem. ” * Tactics are “discrete intellectual choices or prompts manifested at behaviors actions during an information-seeking session. * Moves are “finely grained actions manifested as discrete behavioral actions such as walking to a shelf, picking up a book, pressing a key, clicking a mouse, or touching an item from a menu. ” (Marchionini, 1995, pp. 71-75) 5. 2 Mental model Mental models are dynamic mental representations of the real world (Johnson-Laird, 1983; Norman, 1983). People construct, then draw upon mental models to predict the effects of contemplated actions. Information seekers develop and use mental models for a variety of mental and physical objects, including information objects and different domains of knowledge.
A mental model for a particular information object such as a book allows one to base expectations about how to begin and proceed in reading and to estimate how much effort will be required. A mental model for the domain (topic area) related to the book’s content allows the reader to integrate information (understand) as reading progresses. Mental models account for expectations and therefore learning and change in behavior. A personal information infrastructure includes the various mental models an individual has developed for different information systems and domains of knowledge.
Experience with a variety of information problems and systems leads people to develop general knowledge about how information is organized, and skills for facilitating access. We learn to recognize the advantages and limitations of general organizational structures such as lists, arrays, hierarchies, and networks and how to leverage the advantages and mitigate the limitations. At the most basic levels, skills include memory processes such as rehearsal, association, and chunking (Lindsay ; Norman, 1977) and strategies such as use of mnemonics.
At more formal levels, they include the strategies and relationships we learn and develop throughout our lives. Greeno (1989) distinguishes symbolic/abstract knowledge from mental models to account for these types of knowledge. For example, we learn filing rules such as alphabetical, chronological, or positional orderings that facilitate subsequent retrieval. These rules are generic and serve as defaults for orderings when we encounter a new domain. Particular domains for example biological classes, library classifications and chemical structures.
It use specific orderings that must be integrated in mental models for the domain. As we will see in Well-developed personal information infrastructures allow people to look for organizational rules in new domains before applying default rules. The organizational rules become defaults for experts and thus organizations in our personal lives reflect the domain organizations in our professional lives. Our general knowledge about what is typical about these generic type of ystems overlaps with our particular mental models for those systems. Furthermore, the formal strategies for using particular information sources (for example use of a back of the book index) that are part of our mental models for those systems are generalized and serve as the basis for heuristic or analogical strategies when we encounter new systems. We develop knowledge about areas of personal interest and relationships with others with expertise in those areas so we can exchange information when needed.
As we gain experience with information problems, we strengthen our general information-seeking knowledge and skills just as we develop our knowledge and skills in other general cognitive processes such as listening, reading, writing, speaking, reasoning, and decision-making. 5. 3 Cognitive theories Many cognitive theories include an executive process that controls and monitors the various perception, memory, computation, and motor processes. A popular view of this executive process is termed metacognition (Flavell, 1985).
Metacognitive activity refers to our ability to reflect on our own thoughts and actions in the past, monitor them as they proceed, and plan which ones to take to meet our needs. Our personal information infrastructure is guided by metacognitive activity directed at meeting situated information needs. Metacognition determines that we need information, enables our general information-seeking knowledge and our mental models for systems and domains, and monitors progress. Metacognition is influenced by affective states such as motivation and attitude and by physical states such as fatigue and comfort levels. 5. 5 Material resources
Material resources that make up the personal information infrastructure include: people, books, computers, telecommunications lines, and all the other tangible things we use to gather, generate, manage, and communicate information. Material resources also include money and time we have available to use and maintain these resources. These physical components of our personal information infrastructures are most readily affected by sociological and technological developments. To augment our memories, we accumulate huge collections of paper covered with relatively permanent, visually accessible symbols and marks.
To organize these collections of paper, we acquire drawers, cabinets, shelves, libraries, and archives. To replicate and distribute items from these collections we use copiers, mail and courier services, and telefacimile. To acquire new information, we maintain personal reference collections, hire clerical support staff, nurture networks of colleagues, contract with research companies, and visit libraries. All these objects, people, communication channels, and strategies are part of a personal information infrastructure that individuals develop to accomplish their goals.
Individuals support many layers of personal information infrastructure to serve long and near-term goals and many intermediate goals within the infrastructure itself. 6. COMPARISON OF MARCHIONINI’S MODEL BETWEEN OTHERS MODEL 6. 1 Norman’s model To further understand how users may carry out different information seeking stages in an IR system to complete tasks, the eight information seeking stages in Marchionini’s model are viewed using Norman’s generic model of user interaction in interactive systems.
The subprocess of “extract information” bears the same name as Ellis’ “extracting” activity but the two processes are different. Marchionini (1995) describes extracting thus: “There is an inextricable relationship between judging information to be relevant and extracting it for all or part of the problem’s solution. To extract information, an information seeker applies skills such as reading, scanning, listening, classifying, copying, and storing information. As information is extracted, it is manipulated and integrated into the information seeker’s knowledge of the domain. In Ellis’ model, “browsing” and “differentiating” are activities separate from “extracting,” which is “systematically working through a particular source or sources to identify material of interest. ” (Ellis 1989, p. 242) On the Web, we expect extracting (in Ellis’ sense) to mean systematically working through a selected website or set of web pages (typically using search engines) in order to search and retrieve material of interest. (Chun, Detlor, Turnbull, 1999) 6. 4 Kuhlthau vs Ellis vs Marchhionini’s model
One limitation of the models of both Kuhlthau and Ellis is that they seem to suggest that the information seeking process is a series of linearly ordered, sequential activities whilst Marchionini more accurately identifies the information seeking cycle as being iterative and recursive. Another potential shortcoming of Kuhlthau’s framework vis-a-vis the others is that it doesn’t explicitly present in its stages of information processing criteria on evaluating the quality of the information offered by the resources. Muthu Kumar, Uma Natarajan & Sunita Shankar,2005) 7. STRENGTHS OF MARCHIONINI’S MODEL BETWEEN OTHERS MODEL 7. 1 According to Hung, 2007 The information seeking model put forth by Marchionini closely describes a formal online search from a cognitive perspective, which we found useful to articulate the general components of such a search at a high level. The model is meant to provide an overview of the online information seeking process, rather than a detailed representation of the search itself.
Marchionini’s model outlines some of the types of knowledge that are necessary for the information seeking process. For example, the knowledge that needs to be applied to choose an appropriate search system is clearly distinct from the knowledge that one needs to formulate an effective query. 7. 2 According to Donald Marchionini model are meant to apply to a particular task, typically searching electronic information in databases or online library catalogs. 7. 3 According to Whittaker
Marchionini’s (1995) model is focused on more recent technologies, discussing how information seeking moves from high level framing of information needs to expressing those as some form of query, evaluation of the results of executing that query and reiteration depending on the outcome of that evaluation. 7. 4 According to Komlodi & Carlin, 2004 Marchionini’s (1995) task model has been selected because of its linear main structure (with many feedback loops added) and clarity. While this model mostly represents searching, the analysis presented in the paper creates the foundation for studying browsing as well.
This model provides a clear distinction of steps that can be easily related to the cultural dimensions. 7. 5 According to Wilson While most of the key ISP models awkwardly cater for, ignore, or even abstract-out the fact that users switch frequently between stages, Marchionini’s model is the most explicit in representing the reasons and conditions in which changes occur. Marchionini “crudely”, as he describes it, models these switches by identifying both more and less likely paths that users may follow backward through the stages. Further, the absence of arrows between certain states implicitly highlights switches that do not occur. . WEAKNESSES OF MARCHIONINI’S MODEL BETWEEN OTHERS MODEL 8. 1 According to White The information seeking process model captures many important elements of information seeking, including aspects of collection exploration in examine results, and aspects of knowledge acquisition in extract information. However, it does not have the same emphasis on learning and understanding as exploratory search nor does it fully represent search context or information use REFERENCES Barajas, M. & Higueras, E. (2003). Initial taxonomy of information seeking behaviours. Retrieved February 16, 2011 from http://www. b. es/euelearning/seeks/FINAL%20VERSION%20UB_DEL3. pdf Case, D. O. (2007). Looking for Information: A Survey of Research on Information Seeking, Needs and Behavior. Chun W. C. , Detlor, B. & Turnbull, D. (2000). Web work: information seeking and knowledge work on the World Wide Web. Springer. Chun W. C. Detlor, B. & Turnbull, D. (1999). Information Seeking on the Web: An Integrated Model of Browsing and Searching. Paper presented at ASIS Annual Meeting Contributed Paper in Faculty of Information Studies, University of Toronto, Toronto, Ontario, Canada Hildreth C. R. (1995).
Browsing and exploring: a new paradigm for ir/opac system design. Online Catalog Design Models: Are We Moving in the Right Direction?. Retrieved February 16, 2011 from http://152. 106. 6. 200:8080/dspace/bitstream/10210/1866/4/MyersCh3. pdf Hung, P. W. et. al. (2007). A multi level model of information seeking in the clinical domain. Journal of Biomedical Informatics, 41, 357-370. Retrieved February 16, 2011 from http://www. sciencedirect. com. Irgens, B. The use and usability of the browsing features of bibliographic search systems. J. Long. (1995). Information Seeking in Electronic Environments.
Cambridge Series on Human-Computer Interaction: Cambridge University Press. Komlodi, A. & Carlin, M. (2004). Identifying Cultural Variables in Information-Seeking. AMCIS 2004 Proceedings. Retrieved February 16, 2011from http://aisel. aisnet. org/amcis2004/68 Kules, B. Speaking the same language about exploratory information seeking. Kules, B. et. al. From Keyword Search to Exploration: How Result Visualization Aids Discovery on the Web. Kules, B. , Tunkelang, D. & White, R. W. (2009). HCIR 2009: The Third International Workshop on Human-Computer Interaction and Information Retrieval. Retrieved February 26, 2011 from http://cuaslis. rg/hcir2009/HCIR2009Report. pdf Lee, S. S. et al. (2005). Subjective relevance: Implications on interface design for information retrieval systems. Proceedings of the 8th International Conference on Asian Digital Libraries ICADL 2005, Bangkok, December 12-15 (424-434). Marchionini, G. (1999). Browsing Strategies. Information Seeking in Electronic Environments (pp. 100-138). New York: Cambridge University Press. Marchionini, G. (2008). Human–information interaction research and development. Library & Information Science Research, 30, 165–174. Retrieved February 16, 2011 from http://comminfo. rutgers. du/~tefko/Courses/e530/Readings/Marchionini%20Inf%20 interact%20LISR%202008. pdf Marchionini, G. (2006). Toward Human-Computer Information Retrieval. Bulletion of American Society for Information Science and Technology. Retrieved February 16, 2011 from http://www. asis. org/Bulletin/Jun-06/marchionini. html Muthu Kumar, Uma Natarajan, Sunita Shankar. Information literacy: A key competency to students’ learning. Malaysian Online Journal of Instructional Technology, 2, (2), 50-60. Stutzman, F. Social information seeking during a life transition. Tunkelang, D. (2009). To Search, Ask: Library science will improve online search.
Retrieved February 25, 2011 from http://www. techreview. com/web/22848/? mod=more&a=f (2008, June). Information Seeking Support System. Paper presented at An invitational workshop sponsored by the National Science Foundation, Chapel Hill, NC USA. White, R. & Roth, R. A. Exploratory search: beyond the query-response paradigm. Wikiversity. (2011). Digital Libraries/Online info seeking behavior and search strategy. Retrieved January 13, 2011 from http://en. wikiversity. org/wiki/Digital_Libraries/Online_info_seeking_behavior_and_search_strategy Wilson, M. L. The Tetris model of the information seeking process.