Thinking or acting? Mental workload and subjective preferences for a command code and a direct manipulation interaction style

Subjects practiced drawing a figure on a computer screen by means of two interaction styles: Command codes and direct manipulation. These two interaction styles demanded different cognitive resources of the user and different times to perform the task. After practice, subjects performed multiple trials in three experimental conditions: in a time‐limit condition, with office noise, and in a neutral working condition. The drawing task was performed concurrently with one of three secondary tasks, each tapping different resources. In Experiment 1, subjects were divided into two equal size groups. Each group performed the task with only one of the two learned interaction styles. Secondary and primary task performance indicated no difference in workload between the two interaction styles. Only at the most demanding secondary task did the use of command codes result in higher workload. In Experiment 2, subjects performed the drawing task in each trial using their preferred interaction style. Consistent individual preferences for the interaction styles and a flexible use of the styles according to working conditions emerged with improved performance.