LEARNING AND THE COORDINATION OF SEQUENTIAL INFORMATION

In 4 experiments, participants performed a list-processing task that r equired alternation between 2 sets of information. Previous research ( R.A. Carlson, J.L. Wenger, & M.A. Sullivan, 1993) attributed the major demands of this task to the need to coordinate 2 sets of information. In the present study, learning was manipulated at the level of indivi dual sets; participants acquired consistent sets of information in var ious contexts. This consistency facilitated performance, and the prima ry benefit was that they learned to select items faster. The selection benefit was limited to acquisition contexts that required alternation and was most pronounced when the consistent set was learned as a memo ry set. The results are compatible with Carlson et al.'s model describ ing the organization of control processes and suggest that coordinatin g information from different sources imposes a fundamental constraint on the working memory system.