Switching tasks and attention policies.

The determinants and costs of control were studied in 6 experiments examining the performance costs of changing stimulus dimension (digit value/number of elements) or attention strategies (speed/accuracy) on the first trial after task transition. Costs were compared for task shift and reconsideration only. Preparation ability was studied by presenting all transition information at the beginning of a 2-part block or only prior to each part. Results showed pronounced first-trial transition costs. Different factors were associated with stop-start and task-switching requirements. Transition costs were separate from those of basic task performance. Costs were sensitive to global control considerations and were larger for task dimension changes than for attention strategy shifts. Costs involving task dimension change, but not strategy shifts, were reduced with advanced preparation. These results are discussed in relation to contemporary models of control. A new distinction is proposed between activation and execution of control strategies.