Instructional compensation for age-related cognitive declines: Effects of goal specificity in maze learning.

Abstract

The differential effects of goal specificity on maze learning among 40 young adults and 40 old adults were investigated. Participants had to navigate through a computerized training-maze task. The finish point of the maze could be presented either as a specific location or in more general terms. After solving the maze problem, participants were required to solve the same problem again, either by moving from start to finish or backward from finish to start. The hypotheses that the presence or absence of a specific goal would disproportionately compromise or enhance, respectively, elderly people's performance were confirmed. Although young adults outperformed old participants in all conditions, these differences were much smaller in the nonspecific goal conditions. These results suggest that instruction based on cognitive load theory (J. Sweller, J.J.G. Van Merrienboer, & F. Paas, 1998) can compensate for age-related cognitive declines. Sweller and his colleagues (e.g., Bobis, Sweller, & Cooper, 1994; Owen & Sweller, 1985; Sweller & Levine, 1982) have provided evidence that the extent to which a goal is clearly specified to a problem solver as a problem state affects the problem-solving strategy used. In their experiments, Sweller and Levine (1982) used maze-tracing and numerical problems in which the finish point could be presented either as a specific location or in more general terms. These transformation problems are characterized by an initial problem state, a goal state, and a set of operators to transform the initial state into the goal state. The major mechanism used by problem solvers faced with transformation problems is means-ends analysis. The use of means-ends analysis and learning, that is, the construction of a cognitive schema of the underlying spatial structure of the maze, were independent. Under goal-specific conditions that facilitated the use of means-ends analysis, knowledge of the goal location was the primary factor controlling problem solvers' moves. This rendered the problem insoluble, and problem solvers were prevented from abstracting from the solutions the general rules used in problem solving. The nonspecific goal prevented the use of conventional means-ends analysis and resulted in fewer errors and more rapid learning of the structure of the problem. Sweller and Levine argued that in the absence of a goal, other aspects of the problem structure control moves. Under nonspecific goal conditions, the location of choice