Running Header: Cognitive Mapping                            



Cognitive Mapping of Ideas     

John Weber

University of North Dakota




List of Figures



Illustrative information processing analysis of human visual perception

Page 4


The research cycle

Page 6


Conceptual dictionary structure

Page 8


Conceptual dictionary

Page 9



Pg. 11


Cognitive map: “What causes depression?”

Pg. 12


Conceptual dictionary of the author’s proposed research

Pg. 15


Scaling exercise from the author’s proposed research

Pg. 16



            This paper is about the use of cognitive maps in social work research. There are three parts. The first part attempts to define what cognitive maps are. There will be an analysis of their theoretical underpinnings. To save time, the mechanics of cognitive mapping will be discussed in part two. Part two is devoted to the application of cognitive maps in social work research. The process of constructing a cognitive map will be described step-by-step. Part three will be a presentation of the rough draft of the cognitive map for my own proposed research project. Successful use of the cognitive map should lead seamlessly and economically into the next step of the research project, the literature review.


Part One: What are cognitive maps?

            Cognitive maps are a recent addition to human thinking, but their origin is as ancient as the mind itself. “Thought is imaginative, and it is through metaphor, metonymy, and mental imagery that we transcend the limitations of direct experience to enable abstract thought and creativity… Language and images are active agents in constructing our experience of ourselves and others.” (Gould, 1996). Cognitive maps are based on the methods of cognitive psychology, which supplanted Behaviorism in the 1960’s as the dominant approach in psychology. Behaviorism was limited by its emphasis on observable behavior and animal learning; such an approach was inadequate to explain human behavior “without resort to such concepts as ‘memory’, ‘imagery’, ‘reasoning’, ‘intelligence’, and ‘knowledge’.” (Slatter, 1987). Cognitive psychology uses “an information processing view of human cognition. An information processing system consists of a set of memories, receptors, and affectors, and processes for acting on them.” (pg. 18). One of the most intrinsic operations of the human mind is vision. Figure 1 shows the process of how the mind sorts light from the eyes into “knowable” images. Essentially the same process of assembling, identifying, and assimilating is used in cognitive mapping.

Figure 1.

            Cognitive mapping is a way of sharing innately human forms of knowledge. “Concept mapping is a structured process [italics added], focused on a topic or construct of interest, involving input from one or more participants, that produces an interpretable pictorial view (concept map) of their ideas and concepts and how these are interrelated.” (Trochim, 2002). “A cognitive map is a product—a person’s organized representation of some part of the spatial environment… Cognitive mapping refers to a process of doing: it is an activity that we engage in rather than an object that we have.” (Downs & Stea, 1977). Intrinsically, cognitive maps do not presume scientific uninvolvement by those who construct them but specifically requires their subjective interest: “Purposes are always an integral part of the mapping process… Cognitive mapping can only be understood as a purposive and hence goal-directed activity.” (pp. 68-69). “Individuals evolve [cognitive maps] for themselves that allow them to impose meaning on their personal experiences. Through the cognitive structure, individuals construct their own truth.” (Rodwell, 1998).

            Cognitive mapping is used by many disciplines. Researchers in the field of artificial intelligence emulate human cognition by using cognitive maps as they design expert systems. Business people use cognitive maps to devise organizational models. Politicians use cognitive maps to aid in decision-making. Physics researchers use them for problem representation. Cognitive maps are useful for psychologists, particularly counselors, who work with clients who want to relearn behaviors. The human mind, with its intricate parallel and overlapping systems, can best be projected as an active, evolving system, rather than by static, linear representations. Social work research, which seeks to understand the person-in-environment, within the perspective of the profession’s code of ethics and helping mandate, seems well-suited to the cognitive mapping mode of discovery, with its recognition of directionality, creativity, and change.

Part Two: Using cognitive maps in social work research

            We are at the first phase—the problem formulation phase—of our research

projects. Problem formulation does not begin at a zero point. As Figure 2 illustrates, social work research occurs as part of an ongoing greater cycle. Problems are suggested from prior research, but they have relevance only because they are important to an active social worker or social workers as part of daily work with clients in an evolving society. Not mere scientific curiosity but professional obligation and personal commitment to social justice stirs the social worker’s interest. “At this stage of research, creativity is at a premium.” (Wechsler,, 1976). Inevitably, the researcher’s professional and private experiences contribute to his or her chosen field of interest. “The problem may be

Figure 2.

relevant only to the researcher, but that is all right as it is the researcher who wants to find the answer.” (Friedman, pp. 8-9). The researcher’s experiences will be conscious and the result of much analysis and deliberation, but also, and perhaps more importantly, they will arise from feelings, biases, and unconscious or subconscious awarenesses. Self-reflection is difficult. Why are some of us committed to children’s welfare issues, others to political activism, others to therapeutic work?  “Few of us are aware of the techniques we have used to organize information…” (McGraw & Harbison-Biggs, 1989). Cognitive maps are tools which transform this amalgam of personal and professional ambiguities into clearly identifiable relationships that can become the object of research.

            The second step in creating the cognitive map is to create concepts that seem to relate to the chosen problem. At this point, the researcher simply lists the concepts in the indiscriminate order in which they occur to him or her. McGraw and Harbison-Biggs define concepts as “abstract representations that define objects, elements, or events according to attributes and values. They are exceedingly difficult to speak about, much less to represent.” (pg. 136). Clearly, the researcher’s own viewpoint is vital here: “Concepts are words that symbolize ideas created by the researcher.” (Friedman, pg. 6). The researcher should trust his or her own intuitions when constructing concepts, which can have behavioral components, cognitive (or thought process) components, and affective (emotions, feelings) components. (Lecture notes from Research I class, Oct. 11, 2002). The key is comprehensiveness; later stages of the process impose order and logic. “Very rarely does the problem initially present itself too narrowly. The initial conception of the problem generally takes a global or all-encompassing form… The problem must be carefully examined and recast by narrowing and focusing, so that it can be studied empirically.” (Wechsler,, pg. 6).

            The third step in creating a cognitive map is to make a concept dictionary. This is

            “a mechanism to visualize an abstraction of the primary concepts and the

            terminology used to label them. Such a ‘dictionary’ is based on the notion that

            words may be grouped by common elements of reference into recognizable

            concepts. Some concepts are very closely aligned with one another, while others

            will be more unrelated. Once they have been exposed, conceptual associations or

            groups may be classified together under a broader umbrella that reflects all of the

            concepts that should be depicted.” (McGraw & Harbison-Biggs, pp. 136-137).

Figure 3 is an example of a simple concept dictionary from the McGraw and Harbison-Biggs book. Note that spatial location serves the function of suggesting relationships between the concepts. The concepts that social workers create, much more complex and elusive than those in Figure 3, will form a concept dictionary more suggestive than final.

Figure 3.

“At this point, the concepts are very abstract; it is hard to observe them because the definition of what is being observed is not precise.” (Friedman, pg. 6). Figure 4 is a non-pictorial dictionary of concepts relating to socioeconomic status and health that was taken from our reading for HBSE class this week. Note the range of quantitative and qualitative concepts, and the reasoning behind the various groupings.

Figure 4.


            The next step in making a cognitive map is to operationalize the concepts. Visualization of the concepts as a heuristic, ecological system enables the researcher to better concretize the component concepts:

            “To operationalize, the researcher defines the concepts precisely, and then states

            the assumed relationship between the concepts. This is accomplished by assigning

            measurable terms to the concept… When the researcher has been able to define

            the concepts clearly, he or she has created a variable. A variable is a concept that

            can be measured.” (Friedman, pg. 7).

How does the researcher find which variables measurably impact the problem and which do not? One technique is called ‘scaling’. In scaling, the researcher generates a broad series of statements about the problem, and then rates on a scale—usually between 1 to 5—their perceived importance. (Thus nominal concepts cannot be scaled.)  Quantity but also discreteness of statements are important, so that patterns or clusters might emerge. Concepts that repeatedly rate high numbers are likely to be concepts that are significant and could be investigated further as variables. Concepts that consistently rate low numbers are probably not relevant to the problem. Scaling is the beginning of hypothesis. “Scaling techniques may be useful in providing a method of feedback… Abstracting and representing domain concepts and conceptual relationships can open the [researcher] to refine, extend, and clarify… initial understanding.” (McGraw & Harbison-Biggs, pg. 145). Figure 5 shows an abbreviated example of scaling. The listed concepts are some postulated as impacting depression. The researcher is ranking their importance regarding the statement “Depression recurs in families across generations.”                       

History of depression  1            Age  1              Frequency of occurrence  1

                                     2                  2                                                       2

                                     3                  3                                                       3

                                     4                  4                                                       4

                                     5                  5                                                      5

Figure 5.



The process repeated over many, many questions would begin to reveal clusters that might be termed variables. Some researchers use focus groups to generate statements and programs to sort data.

            The final step is to determine functional relationships between variables; that is, to hypothesize which variables might be independent variables and which might be dependent variables. Friedman (pg. 7) makes an analogy of understanding relationships between variables to making a cake:

            “The eggs, flour, sugar, and water are independent variables. By themselves they

            stand alone. When mixed together, they interact in a way that creates a cake. The

            cake then is dependent… it would not exist without them… Answering a research

            question is similar to baking a cake. A relationship exists between variables. It is

            up to the researcher to determine which are the independent variables and which

            are the dependent variables—that is, which ones are affecting the others.” (pg. 7).

As the researcher reflects on the variables and the possible relationships, it is essential to keep in mind the original question or problem: “The type of research (exploratory, descriptive, or explanatory) you do will determine how you look at the relationship between independent and dependent variables.” (pg. 8). “How a problem initially gets represented is crucial in determining how, and even if, it eventually gets solved.” (Slatter, 1987). Let’s return to our depression example to clarify this point. Suppose our original problem or question was “What causes depression?” Treatment might be an independent variable here, and frequency of episodes a dependent variable. But suppose the researcher reformulated the problem as “Who gets treatment for depression?” Now frequency of episodes might become an independent variable, and treatment might become a dependent variable. This, of course, does not mean that questions cannot be revised as the cognitive map suggests new possibilities, but only that the map and the question must be consistent.

            A simplified cognitive map for “What causes depression?” might look like this (Figure 6). Note that some of the variables are more suited for quantitive analysis, and others are better suited for qualitative analysis. Also, the map could be reconfigured with feedback loops in one chose to investigate advanced types of regression analyses, such as discriminant or stepwise.

Figure 6.

            To recap the process of constructing a cognitive map:

  1. Formulate the problem.
  2. Create concepts.
  3. Make a concept dictionary.
  4. Operationalize the concepts (scaling).
  5. Determine functional relationships.

Now the researcher has the tools to ask qualitative and quantitative questions in the problem area. He or she can go to the literature knowing much more precisely what he or she is looking for.


Part Three: Construction of a cognitive map

            This section presents the construction of the author’s own cognitive map for this proposal writing class. The process will follow the five steps discussed in the previous section.

Formulate the problem. The author’s area of interest is how elderly, lifelong, “traditional” North Dakotans conceive themselves as they age in a world that has changed significantly from the world in which they grew up. Why is this topic of interest to me? Remember that personal and professional experiences can raise questions that lead to interest. One particular personal experience among many stands out. I used to visit my father every few months of the last years of his life. When he was in his eighties, he spent a lot of time in his recliner or wheelchair watching TV. He would spend hours each day watching the Minneapolis news channels broadcast their stories of crime and disaster. There were times I would walk into the room where he was watching this stuff and he would say “Goddammit, John, I think sometimes I’m going crazy.” (This from a man who never cursed in his house.) Then we might have a discussion where he would reveal his profound frustration at losing touch with a world beyond his comprehension, at a time of his life when he expected to have it better figured out.

An example from my professional career is my work with a man at the nursing home where I intern. This man has been on kidney dialysis for over ten years. There are days when he is in so much pain that all he can do is lie in bed and moan. He spends most of his “better” time sitting in his room writing letters, reading, and looking at favorite objects from his early life. He is a religious man. He pays attention to the outside world, but is not judgemental. Like my father, he grew up on a poor North Dakota farm. I often ask myself where this man gets his strength from.

Create concepts. This is the “brainstorming” phase of the process. I listed all the components—behavioral, cognitive, and affective—that seemed meaningful in the problem area, in the random order in which they occurred to me.

“traditional” North Dakotan




sickness, frailty

loss of skills

male-dominant paradigm



poverty as beneficial





Ethnocentric, bigoted

identification with country of origin

            respect for elderly

            didactic toward youth


            “hi-context” style of speech

            don’t go to doctor. Stay out of hospitals.

            traditional family-oriented


            flu epidemic


            WW II


            media (John Wayne movies)

            Make a concept dictionary. Here I have grouped concepts into categories (Figure 7). An overview of the problem emerges as the categories align themselves into schema. Detail sharpens as contrasts emerge. New concepts are added in response to further reflection.



“traditional” North Dakotan: agrarian, religious, identification with country of emigration, hi-context style of speech

Male-dominant paradigm: self-sufficient, individualistic, winning, traditional family-oriented

Values: poverty as beneficial, anti-welfare, respect for the elderly, didactic toward youth, traditional family-oriented

Don’t go to doctor. Stay out of hospitals.

History: flu epidemic, Depression era,

WW II, patriotic

Media: John Wayne movies.

Elderly: sickness, loss of skills, awareness of death, spirituality

Death. Spirituality

Current culture

Current values

Current events.

Current media. Fargo.


Figure 7.


            Operationalize the concepts.  Here I assess the importance of prospective variables through scaling. Room does not permit this process to be fully re-enacted here. Figure 8 shows the author’s scaling exercise for the variables pertaining to a “traditional” North Dakotan as they might relate to feelings of frailty at the end of the life span.

Agrarian  1                   religious  1                    hi-context style of speech  1

                2                                  2                                                             2

                3                                  3                                                             3

                4                                  4                                                             4

                5                                  5                                                             5

identification with                      Male-dominant paradigm          traditional family-

country of emigration   1                                                     1    oriented                  1

                                      2                                                  2                                  2

                                      3                                                  3                                  3

                                      4                                                  4                                  4

                                      5                                                  5                                  5

Figure 8.

            Determine functional relationships. This is the step where quantitative variables begin to suggest themselves and qualitative variables suggest exploration. If the variables developed thus far have discreteness, several plausible functional relationships should occur to the researcher. Listed below are five variables which I have preliminarily determined are worth pursuing in my area of interest and five concepts which I have discarded as not functionally significant.

Functionally significant variables:


            Media portrayals

            Current events

            Feelings of significance (self-worth)




            Hi-context style of speech

            Identification with country of emigration

            Don’t go to doctor. Stay out of hospitals.





            Cognitive mapping is not merely a tool but, as I have tried to emphasize, a mental process that is innately human. “Cognitive mapping is an attempt to represent the cognitive structures in memory.” (Carley & Palmquist, 1992). As such, the process uses symbols and meanings shared by others in the human fabric. “Cognitive mapping is perhaps the most useful means of exploring the nature of shared knowledge in several groups. It can be used on relatively large numbers of individuals, and the results can be compared to each other, to the cognitive maps of experts, or to idealized representations.” (pg. 605). Properly used, they can be inventive, but their inventiveness leads only to truth.


Carley-K. & Palmquist-M. (1992). Extracting, representing, and analyzing mental models.

            Social Forces, 70(3): 601-636.

Downs-R. & Stea-D. (1977). Maps in the Minds: Reflections on Cognitive Mapping.

            New York: Harper & Row.

Friedman-B. (1998). The Research Tool Kit: Putting It All Together. Pacific Grove, CA:


Gould-N. (1996). Using Imagery in Reflective Learning. In “Reflective Learning for

            Social Work”, Gould-N. & Taylor-I., eds. Brookfield, VT: Ashgate.

McGraw-K. & Harbison-Biggs-K. (1989). Knowledge Acquistion: Principles and

            Guidelines. Englewood Cliffs, NJ: Prentice-Hall.

Rodwell-M. (1998). Social Work Constructivist Research. New York: Garland.

Slatter-P. (1987). Building Expert Systems: Cognitive Emulation. New York: Halsted


Trochim-M.K. (2002). Concept Mapping. Retrieved Jan. 20, 2003 from Cornell

            University website:

Wechsler-H., Reinherz-S.M., and Dobbin-D. (1976). Social Work Research in the

            Services. New York: Human Sciences Press.