A Knowledge Structure Map (KSM) is an organised visual map containing nodes and connecting lines (arcs) and showing a knowledge area. A complete map also contains expert opinion about the knowledge area being shown.
The nodes or boxes on the map represent specific, discrete pieces of knowledge. They may represent very complex pieces of knowledge or they may represent very simple pieces of knowledge depending on their place in the structure. Each knowledge node has between four and eight numeric parameter values associated with it and a descriptive name, a detailed definition, a brief summary of the knowledge itself, a link to where the actual knowledge can be found and optional additional notes.
A connecting line or arc between nodes is a directional link and indicates learning dependency. Learning dependency relates to the way that a human expert would acquire knowledge. A link shows that the knowledge represented by one knowledge node must be understood by the expert before it is possible that he or she can fully understand the knowledge represented by another node. A link leaving from the bottom of node ‘A’ and connecting to the top of node ‘B’ shows that an expert would need to know knowledge ‘B’ before he or she could fully understand knowledge ‘A’.
This simple map shows that in order to fully understand the knowledge represented by ‘Node A’, an expert would already know the knowledge represented by ‘Node B’, ‘Node C’, and ‘Node D’. An expert that knew the knowledge represented by ‘Node B’ would already know the knowledge represented by ‘Node E’ and because of the chain of learning dependency, also know the knowledge represented by ‘Node G’. The knowledge represented by ‘Node F’ must be known before it is possible to fully understand the knowledge represented by ‘Node C’ and ‘Node D’ but not by ‘Node B’. The knowledge represented by ‘Node D’ implies a prior understanding of the knowledge represented by ‘Node G’. However, the knowledge represented by ‘Node B’ does not imply a prior understanding of the knowledge represented by ‘Node H’. The arc or link from ‘Node D’ to ‘Node H’ is redundant because the arcs from ‘Node D’ to ‘Node F’ and then from ‘Node F’ to ‘Node H’ already show that the knowledge represented by ‘Node H’ must be known before it is possible to fully understand the knowledge represented by ‘Node D.
The name of a node should accurately reflect the actual knowledge being identified. The knowledge being represented may be typical and used widely or it may be unique to a particular study. There is a tendency to over generalise names, like naming a knowledge node ‘Physics’ when it is clear that it is not all of and everything about physics that needs to be known. A better name may be ‘Appropriate physics for water pump design’. This may be longer but it provides a much clearer indication of what knowledge is required. The word ‘appropriate’ could be left off to make the name more manageable. If names are too long, they will not fit in the node box on the map. Creating a name for a knowledge item is a compromise between close definition and brevity.
The definition should fully define the knowledge item being represented by the node. The definition will set the boundaries for the knowledge and include what knowledge is really needed and exclude knowledge that is not needed. The definition will clarify the meaning of the node name.
There should be a very brief summary of what the knowledge actually is or maybe what one actually has to know or know how to do. The requirement for brevity can make this summary difficult to create. However, the effort is worthwhile because it can help to clarify map structure as well as providing information about a particular knowledge node. The summary will probably reflect the amount that one could find out about the knowledge if an expert literally had only one minute to explain it. This means that there will hardly ever be any detail. The summary will only include the essential essence of the knowledge in question.
It is intended that the link will be a link to a web page that either contains the detail of the knowledge being represented by the node or contains an overview with additional appropriate links to greater detail. The link should allow access either directly or indirectly, to the detail of the knowledge being represented. It is possible that an organisation would wish to use the link to specify the internal location of a piece of knowledge or even the main person or team that has ownership of the knowledge (the experts). The default option should be to identify a web link but alternative links are acceptable as long as they do allow a system user to locate the knowledge.
This is an optional text parameter that must be enabled to allow access to it. The space is the same as that for summary, definition etc but the precise use of the parameter will not be specified within the methodology. The Notes parameter is therefore intended to be used for a specific study in a way that may benefit that particular study. It allows additional notes to be entered in context at the time that interviews are carried out and to be attached to the knowledge node that was the focus when the notes were generated. It would be possible for instance to record the name of the expert that generated the node or the names of the experts that hold the knowledge etc. However, both of these options are not recommended and indeed would be discouraged within the framework of the KSM methodology.
Each knowledge node on a Knowledge Structure Map (KSM) should be qualified by four primary numeric parameters. The default parameters used to qualify a node are, ‘Importance’, ‘Difficulty’, Study-Experience’ and ‘Known-By’. Each parameter should be assigned a number between 0 and 10 (inclusive). The meaning implied by these parameter names is:
Importance: How important is this piece of knowledge with respect to the knowledge area being studied. 0 = not important and 10 = highly important.
Difficulty: For the study of a business knowledge area this means – how difficult would it be for the organisation to replace this knowledge if it were lost to the organisation? 0 = very easily replaced and 10 = impossible to replace.
For the study of a general knowledge area this parameter means – what is the degree of difficulty (from a human perspective) associated with this knowledge? 0 = it is very easy to learn and understand and 10 = it is highly complex and very difficult to understand.
Study-Experience: How would a typical human expert acquire this knowledge? 0 = completely through formal study from books or from lectures etc. and 10 = only and fully from the experience of using the knowledge in a practical situation. Numbers in between reflect how much of each extreme represents the correct balance for knowledge acquisition. So a value of 5 shows that the knowledge needs to be studied just as much as it needs to be practiced in order to acquire it fully.
Known-By: What is the proportion of people from a given size of population that would know this piece of knowledge? 0 = no single person from the given population would know all of this knowledge and 10 = everyone (100%) in the given population would know this. The given population should be defined and agreed before a study is started. Experts being interviewed will need to estimate the proportion or percentage of the agreed population that are thought to have the knowledge in question.
Each of the four primary parameters can be represented by an alternative name. These are:
Importance: Critical
Difficulty: Complexity
Study-Experience: Tacit
Known By: Well Known
Up to four additional numeric parameters are selectable from a current list of six optional parameters. Additional parameters must be enabled before they become available for a study. The options currently available are as follows.
Long Term: Is the knowledge element expected to be in use within the knowledge area for the long term or is it only intended to be used short term. A knowledge element may be required by an organisation in the long term even though the knowledge area itself is only required short term. Knowledge may belong to a work area that is being phased out. 0 = immediate use only, 2 = for current short term work, 7 = Long Term use, 10 = indefinite
Specialised: Much of the knowledge required within any knowledge area is quite general and is also used in other knowledge areas or the knowledge area itself may be very general. Alternatively, knowledge may be needed that is highly specialised and only used in very restricted circumstances or applications. 0 = completely general, 2 = mostly general, 7 = specialised knowledge, 10 = highly specific
Stability: Technological growth is one thing that can make knowledge unstable. Knowing how to do something may regularly involve changes in tools and technique as new systems are introduced. This may be reflected, if restudied regularly, by a changing prerequisite structure for a knowledge element. The prerequisite structure itself being relatively short term knowledge yet the knowledge element in question may be a long term requirement. 0 = Changes regularly, 2 = changes frequently, 7 = generally stable, 10 = will not change.
Security: This needs to be defined more carefully in the pre study stages but in general relates to the existence of security issues that are associated with a knowledge element and/or its application. Security issues may relate to business protection, customer confidentiality or other aspects of security. The term or parameter can be used to indicate if there are any aspects of security associated with the knowledge at all or a particular security issue could be identified pre study. 0 = no security issue, 2 = minimal issue, 7 = significant issue, 10 = potential security problem
Hazard: This parameter may also need to be defined more carefully pre study but in general is intended to indicate if there are any hazards associated with the knowledge and/or its application. Hazards may relate to the protection of people, equipment or the environment. 0 = no hazard involved, 2 = unlikely to involve hazard, 7 = involves hazard, 10 = potential danger
Justifiable: Justifiable knowledge can be believed and therefore applied with more confidence. Some knowledge is applied but the reasons why it is applied are not known, it may be empirical knowledge and therefore may at some stage be inappropriate under new conditions. Other knowledge is backed up with sound theory or other justification. 0 = not justifiable knowledge, 2 = slightly justifiable, 7 = justifiable, 10 = proven and justifiable
The Knowledge Structure Map is a way of visualising and then analysing a knowledge area. The construction of a KSM involves the study of the knowledge area. Studying a knowledge area is not always seen as a valid thing to do, usually because there are few objective ways to do it. Processes concerning a knowledge area often centre on training or capture. The study of knowledge is however, potentially more beneficial than either of these things because it provides information about a knowledge resource that can support the more general activity of decision making and knowledge resource management. A knowledge area study can supplement information about other aspects of a business, an activity, a plan, a design or even a concept. The study of knowledge creates options and provides information to support decision making. It also helps to clarify a knowledge area and creates a common reference structure that facilitates objective debate.
The idea that an area of knowledge should be studied for some reason means that there is at least a rough idea concerning the boundaries for this knowledge. In order to initiate an objective study however, these boundaries must be clarified and stabilised. Once the area of knowledge is clear, that is, it is clear what is part of the knowledge to be studied and what will be excluded from the study, then a focus for the study itself must be identified. This means to provide both a starting place for the study and a concrete reference point that will act as a guide throughout the study. When a Knowledge Structure Map is being used as the study method, this focus involves creating a clear and unambiguous statement or question that will guide the study. Since the study will be based on human expert knowledge acquisition, the focus will often be a question about what needs to be known. For instance the focus for the study may be of the form ‘What does one need to know in order to fully understand X’, where ‘X’ is a knowledge area or more specifically, ‘What does one need to know in order to manage and run a small corner grocery store’.
The size of the knowledge area implied by the question need not be of great concern. For a very large knowledge area such as that needed in order to know everything necessary to be the prime minister of the UK , the question simply means that the study will remain at a fairly high level of knowledge throughout the study period. A smaller knowledge area such as the knowledge needed to change the wheel of a typical family car would much more rapidly start to identify quite fundamental and discrete knowledge items such as how wheel nuts attach to bolts.
The first node to be placed on the map represents the whole knowledge area to be studied. The rest of the map shows what this knowledge is composed of and how the knowledge is organised from a learning dependency perspective. The name of the first node will be the name of the knowledge area to be studied. It is useful if the definition box contains the question that has been agreed will focus the study and drive the exploration of the area. There is less point in providing a summary or link and there is also little point in providing parameter values for the first node. This is because the rest of the map is the thing that provides the information that helps to clarify the meaning and content of the first node.
Under normal circumstances, there will only be one ‘top’ node on a Knowledge Structure map. The rest of the map will represent knowledge that needs to be known before the ‘top’ (or fundamental) node can be fully understood.
For the majority of map exploration, new knowledge areas will be identified by considering the learning dependency of each knowledge node. This will be discussed later. Identifying the major knowledge sub areas of the top or fundamental knowledge node is often dealt with slightly differently. When carrying out a knowledge area study that is business based, it is often necessary to identify the main areas of activity within a knowledge area before beginning a more strictly focused learning dependency study. This is because business often organises knowledge in unique ways. The organisation of knowledge will often reflect business processes. For instance, a business may have a ‘Post Sales’ department and wish to study the knowledge necessary to fully understand ‘Post Sales’. In this case, it is important that the Knowledge Structure Map covers all of the knowledge needed within ‘Post Sales’ and does not show knowledge that is not needed. So the first questions to ask about the knowledge area of ‘Post Sales’ would concern the main activities that take place within this area. These activities will often reflect business structure and the way that departments and teams are organised. Once this is done then the KSM investigation can revert to its primary goal of uncovering the knowledge needed in order to fully understand a target knowledge item. For knowledge studies that do not have this business structure and business process restriction, the study can focus on learning dependency from the outset.
For business studies the aim is to identify between 3 and 6 major knowledge sub areas of the fundamental knowledge area. This needs to be done with great care and consultation. The important thing is that the sub areas identified MUST cover all of the knowledge from the fundamental area. This means that later in the investigation, there will not be any new knowledge item identified that does not fit correctly within the structures of the major knowledge sub areas identified. For ‘Post Sales’, Major knowledge sub areas may include ‘Installation’ and ‘Customer Care’ etc. and these may also be the names of departmental or team structures within the business. It may be necessary, on occasion, to invent a new term that covers several sub areas in a hierarchical way. This can become necessary if too many sub areas (> 6) are identified at the beginning of the study. In general, when too many sub areas are identified, it is often the case that there is really learning dependency and therefore structure within the sub areas. For instance, it may be revealed that ‘Site Surveying’ is a knowledge sub area of ‘Post Sales’. However, if one knows all about installation, one must also know about ‘Site Surveying’. Therefore, ‘Installation will be a sub area of ‘Post Sales’ and ‘Site Surveying’ will be prerequisite knowledge for ‘Installation’. If you do not agree with the hierarchy discussed here, this may be because your definitions for terms such as ‘installation’ differ from those used within the organisation.
The important thing here is to start the KSM off properly. In general this is simply part of discovering learning dependency but for business studies, it can involve using activities to identify and classify knowledge according to business needs. This way, the results from a study will be more useful to the business than if learning dependency was enforced strictly from the outset. The map will reflect the way that the business uses its knowledge.
The compromise of controlling the number of direct sub areas to the fundamental area is something that should still be attempted even for a non business knowledge study. This can again be achieved by inventing names that cover several other knowledge areas. In most cases it is not necessary to do this because structure will usually exist within a very large number of ‘so called’ direct sub areas. It is most likely that this structure has simply not been identified. Where it really is necessary, try to make sure that the name invented intuitively covers the sub area group to be considered as prerequisite knowledge to the invented term. In some cases, invented terms that group knowledge components can be enlightening for the business and can add value to the study.
With the practical business compromise defined, it is necessary to revert to a rigorous approach to the use of learning dependency as soon as possible. This will be discussed more carefully later.
Each node on a Knowledge Structure Map represents a piece of knowledge that is part of the knowledge area being studied. It is important that users of the map can clearly identify what the knowledge represented by the node actually covers. Ambiguity caused by the use of generalisations should be avoided. The meaning and boundaries of the knowledge can be made clear by the use of several text fields. A node should be assigned a knowledge name that is brief but descriptive. The name should not be too general as in the case of stating the requirement for a knowledge of ‘word-processing’ when the knowledge being represented is really ‘letter writing using a word-processor’.
A definition text field allows the name to be clarified and the meaning of the knowledge node to be made very clear and unambiguous. The definition should state what the knowledge represents and also define the boundaries of the knowledge. This means to state what is part of the knowledge item and what is excluded or where the boundary between inclusion and exclusion exists.
A summary field should hold a very brief explanation of the knowledge itself. This is not intended as a definition but an explanation. The explanation must be very brief and would be about as much as an expert would be able to tell you if you asked what the knowledge was and there was less than 1 minute to provide the brief answer. This explanation then is definitely a summary but should be enough to strengthen the users understanding of the knowledge being represented by the node. The provision of a good summary will also help to validate the map and confirm that the knowledge represented by the node is in its correct place in the learning hierarchy.
A link field is available to allow the location of the detail of the knowledge item to be identified. It is intended that this should be a web link to a page that either contains a detailed description of the knowledge or a page that links to other descriptive pages and resources that together provide the detail of the knowledge. In either case, it should be possible for a user of the map to learn the knowledge by studying the material available when the link is followed. Since the link field is simply a text field it would be possible to use the link to describe the location of other learning resources or even to state which experts to consult in order to learn the knowledge. The only firm requirement is that the link should allow a map user to gain access to the detail of the knowledge represented by the node.
An optional notes field is available to allow a particular knowledge study to be customised further to any more unusual requirements of the study or the organisation that has commissioned it.
In order to help with concentration during the construction of a KSM, it is useful to add parameter values for each node when 1 to 5 nodes have been placed and defined. This provides a break from knowledge structure study yet retains focus on the particular area of the map being explored. Parameter values should be entered in accordance with the values defined for each, above. It should be clear that the values being sought are the educated opinion of knowledge area experts and are not intended to be verifiable facts. Within the context of a knowledge study, it is useful to know where and why expert opinion may differ from other, say management, opinion. It is also clearly useful to know what experts think of the knowledge area that they are expert in, given that the opinion of experts is useful to consider. Because the parameter values being sough are expert opinion, allowable parameter values are integer numbers only. To specify a parameter value between 0 and 10, of 3.725 would be to overstate the usefulness of the expert opinion.
It is good practice to assign all parameters for each node before moving on to the next node (or assigned in small groups of 4 or 5 nodes). This allows a specific piece of knowledge to be considered in several ways in a single operation. Some people like to consider one parameter across several nodes so that relative values can be assigned. For instance, knowledge A is more important than knowledge B but less important than knowledge C. This is not really a valid thing to do because the comparison is being made across very few nodes from the map. A better way to utilise this sort of comparison of parameter values is during validation where values for a node can be contrasted with values from all of the other nodes on the map. Allowing the expert to assess a parameter value for a particular node within the context of the knowledge area being studied rather than the context of a few pieces of knowledge from that area is a reasonable position to adopt. When the expert is asked to assign a parameter value therefore, this should always be done within the context of the knowledge area being studied.
The value for importance is asked in relation to the knowledge area. The value for difficulty to replace is asked in connection with an organisation or that part of an organisation that utilises the knowledge area being studied. The value for study, experience should be asked in with respect to the piece of knowledge itself but within the context of the knowledge area. The value for known by should be asked relative to an agreed population of people or staff. Usually, this agreed population are those that work in the knowledge area in question but for a smaller company, the population may be all of the staff from that company.
If the knowledge study concerns an area of knowledge that is not associated with a company or organisation then the difficulty and known by parameters need to be considered differently. Difficulty to replace becomes ‘difficult’ as in how difficult or ‘complex’ a piece of knowledge do humans generally find this. Known by may become an estimation of the proportion of people in a much larger population group that this knowledge area is relevant to. For instance, if the knowledge area concerned a knowledge of politics in some context then the population may be those eligible to vote and the expert will be asked to provide an educated guess as to the proportion of this population that are likely to fully understand this piece of knowledge.
In general, a piece of knowledge represented by a node on the map implies a full and complete understanding. It is not appropriate to supply parameter values that relate to a partial or overview understanding of the knowledge in question. Therefore, the known by question for instance relates to the proportion of the given population that fully understand the piece of knowledge. The KSM itself should contain knowledge that must be fully understood. It is not appropriate to imply partial understanding since it would then be necessary to consider what parts of the knowledge are essential and what parts are not. The KSM should only include knowledge that is essential.
The Knowledge Structure Map is created by exploring the learning dependency of each knowledge item represented on the map. There are a few minor adjustments to this as discussed for the elicitation of the major knowledge sub areas of the map. However, learning dependency is the primary map structure. For any target knowledge node, the process of identifying dependent knowledge and developing the map involves finding out what an expert that knew the target knowledge would be expected to already know in order to fully understand the target knowledge. For instance if the knowledge item was ‘drive a car safely on the national road network’ then it would be expected that the person that knew this would already know ‘how to control a typical car’ and ‘national highway code’ etc. Equally, if a person knew ‘how to control a typical car’ then it would be expected that the person would already know ‘how to direct a car’, ‘how to stop a car’ etc.
In this way the KSM will be developed in accordance with the rules that provide a structure for the map based on learning dependency.
It is likely that in a restricted study time, a KSM may not be complete. That is, it may not show every conceivable item of knowledge that is prerequisite of another item. However, the goal to aim for in the construction of a KSM is that the knowledge that the nodes do represent and are shown on the map, MUST be placed correctly. This means that careful consideration should be given when any item of prerequisite knowledge is identified.
In some cases, people (experts) can confuse process with learning dependency. This means that they know that some activity involving some knowledge must be carried out before another activity involving other knowledge can be carried out. For instance, it is necessary to have a parts list before one can create a tender or estimate for a contract. This is a process dependency not a learning dependency. In this example it is still possible to know how to create a tender without knowing how to create a parts list. It is true that the parts list must be available before a tender can be created but the parts list could be created by someone else and so the person with the knowledge of how to create a tender does not need to know how to create a parts list.
When a KSM is being developed by exploring learning dependency, it is possible that knowledge that is prerequisite of a knowledge item being explored, is already on the map. During the exploration of another part of the map, many knowledge items will be placed. It is possible that some of these will also be prerequisite of new knowledge that is being explored currently. The figure showing ‘A Simple Map:’ contains knowledge ‘Node E’ that is prerequisite of both ‘Node B’ and ‘Node C’. When new knowledge is placed on the map it is necessary to look at the rest of the map to investigate whether existing prerequisite nodes should be linked to the new item to create new dependencies from existing knowledge. It may not always be possible to spot such dependencies when the focus is on a specific area of the map so it is useful to look over the whole map from time to time during its construction to try to identify new dependencies within the existing knowledge.
There are two main things that require periodic and final validation when constructing a KSM. These are the parameter values assigned and the learning dependencies indicated. It is also usual to give at least a final consideration to the content of the text fields and to at least ‘spell check’ these fields.
Periodic validation of all numeric parameters should be carried out several times during the construction of a KSM and then again at the end. Validation may involve considering each parameter in turn relative to all of the other values on the map of that same parameter. For example, listing all of the knowledge nodes in order of most important first is a good way to validate the importance parameter. Alternatively, each value can be considered relative to the values assigned to the parameters of knowledge nodes that surround a target node. For instance, it is not likely that a target item of knowledge can be known by nearly everyone if its prerequisite knowledge items are known by very few people.
Periodic and final validation of the map structure involves looking for new dependencies within the existing knowledge and also challenging existing dependencies that seem questionable. This can only be done by pondering the map in a reasonably methodical way. As a final stage, the map should be checked for redundant arcs as defined in the section that showed ‘A Simple Map:’.
There are several follow up activities that will not be explored further here. These activities are: