Gagné, R.M. and Smith, E.C. Jr. (1962) A study of the effects of verbalization on problem solving. Journal of Experimental Psychology, Vol. 63, no. 1, pp. 12-18.
Gagné and Smith present some studies they have made on the effects of verbalization on solving succeeding similar problems, and compare their results on the studies of others. They start their article by presenting some studies in which it is the experimenter that verbalizes the strategy for the problem before the subject starts the task. It is not very surprising, that in this kind of circumstances, those subjects that found the strategy on their own, performed better in the succeeding tasks than the ones that had the strategy given.
For a usability tester the studies in which the subjects were asked to verbalize their strategies in their own words before continueing to the next tasks were more interesting. Gagné and Smith divided the subjects into four groups representing combinations of two treatment variables: a requirement to verbalize a reason for each move the subject makes while solving the problem, and an instruction to search for a general strategy to solve similar problems. In this blog, I shall concentrate on the first variable, since it had profoundly more effect on the performance with the succeeding problems.
As the subjects who were asked to verbalize their every move tried to solve their first problems, they were slower than the silent ones even to the point of slight annoyance. But as the subjects got to the more demanding problems with which everyone acted silently, the ones that had verbalized their moves, performed much better, having profoundly less excess moves than the ones that solved the first problems silently. Searching for general strategies did not have a significant effect on the succeeding performance.
Gagné and Smith nicely summarize their experience: It would appear that requiring verbalization somehow forced the subjects to think.