Cognitive Theory and Word Processing Training: When Prediction Fails
|Neil Charness||Elizabeth Bosman|
|Florida State University||Boycrest Centre for Geriatric Care|
|Catherine Kelley||Melvin Mottram|
|University of Waterloo||Alberta Safety Council|
One way to judge the progress of a field is to see whether we can predict accurately from principles discovered in the laboratory to real-world situations. If we cannot, then it may be time to go back to the laboratory and re-examine our experimental paradigms and our theories. A parallel can be seen with biological research, comparing in vitro with in vivo testing. The effect should not disappear as you scale up from cell cultures to laboratory animals to human populations, and there should be no unintended side effects, such as was the case for the drug thalidomide.
One of the major problems in scaling up laboratory results in social sciences such as psychology is that we usually deal with people who can change the method of doing nearly any task, and they typically perform a given task in a vast range of environmental settings. To circumvent this problem, we typically attempt to hold constant as many variables as possible that might be involved in performance, manipulating just the few we feel are critical to the process we are investigating. We can often constrain task strategy with such controls. Unfortunately, we usually do not have this level of control in the field where strategies can and do vary freely to fit the environmental constraints.
Also, in aging research in particular, given the difficulty in locating participants for our experiments, and the heterogeneity of older adults, we rely heavily on repeated measurements with the same people: within-subject experimental designs. Such experiments usually involve many (10-1000) massed trials of similar or identical experi