THE THEORY OF COGNITIVE FIT - Human-Computer Interaction and Management Information Systems: Foundations - page 136

Table 8.2
Classification of Cognitive Fit Studies Examined
Testing the theory of cognitive fit
Simple information acquisition and information evaluation tasks
Vessey and Galletta (1991)
Mahoney et al. (2001)
Chan (2001)
More complex tasks
Wilson and Addo (1994)
Speier and Morris (2003)
Speier, Vessey, and Valacich (2003)
Wheeler and Jones (2003)
Extending the theory of cognitive fit to new domains
Human judgment tasks
Umanath and Vessey (1994)
Tuttle and Kershaw (1998)
Multi-criteria decision making tasks
Smelcer and Carmel (1997)
Dennis and Carte (1998)
Extending the theory of cognitive fit to new dimensions
Cognitive fit as traditional match-mismatch using new dimensions of fit
Hubona, Everett, March, and Wauchope (1998)
Beckman (2002)
Khatri, Vessey, Ram, and Ramesh (2006)
Cognitive fit based on the extent of fit between task and problem representation
Borthick, Bowen, Jones, and Tse (2001)
Dunn and Grabski (2001)
The second study in this category is that by Mahoney, Roush, and Bandy (2003). These researchers
also conducted a study that investigated the effect of task type (spatial and symbolic) and presentation
format (graphs and tables) on performance. Their results support cognitive fit for both accuracy and
time on both spatial and symbolic tasks.
The third study in this category, that of Chan (2001), assessed the effect of graphs and informa-
tion overload on decision quality. For the aspect of interest here, Chan hypothesized an interaction
effect between presentation format and information load. He found neither a main effect for presen-
tation format nor an interaction effect with information load (although he hypothesized both!), and
summarizes his results by stating that “presentation format has little effect on decision quality for
very simple or very complex tasks.” Graphs and tables performed equally well in both conditions.
We can gain insight into why this finding occurred by examining the tasks Chan used. Chan's
nominal task (low information load) consisted of predicting operating profit margin using a sin-
gle variable—the operating profit margin. The associated displays presented this variable over the
prior five years as a line graph and as values in a single row of a table. Hence the participants had
just five data points with which to work. No mention is made of the data used for high informa-
tion load. However, because no variable other than operating profit margin is mentioned in the
article, one might infer that it also uses a single variable with more years of data. Hence it is per-
haps not surprising that Chan's extremely simple information load conditions produced tasks that
could be addressed equally well using either graphs or tables.