[en] This paper presents a lab-in-the-field experiment with craftsmen working on renovation projects to assess the effect of training programs and incentive scheme on coordination and cooperation. Workers frequently fail to cooperate and coordinate their tasks when not supervised by a project coordinator. This is particularly important in the construction sector where it leads to a lack of final performance in buildings. We introduce two different incentives: a first contract paying craftsmen only according to their individual performance, and a second contract paying a group of three craftsmen with a weak-link payment according to the group’s worst performance. In addition, we test these incentives on two different subject groups: one is composed of craftsmen trained to coordinate their tasks, and the others are not. The results suggest that trained subjects coordinate at significantly higher effort levels than non-trained subjects when facing an individual-based incentive. However, when facing a group-based incentive, non-trained subjects seem to "catch up" trained subjects in terms of coordination level, while these latter subjects do not significantly increase their performance level.
