THE THEORY OF COGNITIVE FIT - Human-Computer Interaction and Management Information Systems: Foundations - page 150

Figure 8.3 Cognitive Fit Model Extended to Include Internal Representation of the Task

External
Problem
Representation

Mental
Representation
for Task Solution

Problem-Solving
Performance

Internal
Task
Representation

External
Task
Representation

accuracy, or time, or a combination of the two. One of the separation tasks was facilitated by the
separate control interface and the combination task was best supported by the joystick interface, as
expected. Because ceiling effects were evident in the other two separation tasks, the experimental
manipulations were ineffective. Hence, the two meaningful findings from this experiment support
the theory of cognitive fit.
The third study of this nature is that of Khatri et al. (2006). This study provides an example of
a further match that can be examined within the theory of cognitive fit, that between the internal
representation of the task and the external problem representation used to address the task (Khatri
et al., 2006). See Figure 8.3. Specifically, the researchers assessed whether geospatio-temporal
extensions to traditional conceptual models—in this case, entity-relationship (ER) models—are
best presented by annotating the conceptual schema or by providing the geospatio-temporal data as
an additional external representation separate from the schema. The authors examined perform-
ance on both syntactic and semantic schema comprehension tasks that were presented as questions
in the form of English-language text.
The nature of the problem solver's internal representation of the task can be addressed by
drawing on the theory of text comprehension (Kintsch, 1974; Rumelhart et al., 1972) and the
related way in which text is stored in long-term memory. Text is viewed as being composed of
propositions. In similar fashion, the knowledge acquired from the text is viewed as being stored
in long-term memory as abstract conceptual propositions, consistent with Anderson and Bower's
model of memory, that is, human associative memory (HAM) (1980).
In HAM, each proposition tree is divided into two sub-trees, one of which represents a fact,
while the other represents a context; the tree therefore signifies that the fact is true in the given
context. A fact can be subdivided into a subject and a predicate, and a predicate can be further
subdivided into a relation and an object. Note that, in terms of the ER model, a relation and an
object correspond to a relationship and an entity, respectively. The fact sub-tree therefore matches
the information presented in the traditional ER model. The context sub-tree can be subdivided
into location and time. Hence the context sub-tree represents geospatio-temporal data.