When Diversity in Training Improves Dyadic Problem Solving

Students learned how to solve binomial probability problems from either a procedurally based lesson or a conceptually based lesson and then worked in distributed pairs by using a computer‐based chat environment. Cognitively homogeneous dyads (i.e. both members received the same lesson) performed more accurately on standard problems, whereas cognitively diverse dyads (i.e. each member received a different lesson) performed more accurately on transfer problems. The cognitively homogeneous dyads perceived a greater sense of common ground with their partner, but spent a greater proportion of their time communicating about low‐level details (e.g. message verification) whereas the cognitively diverse dyads spent a greater proportion of their time on high‐level discussion (e.g. solution development). Results help to clarify that common training leads to more positive perceptions of collaboration, but only improves performance on problems that are highly similar to those experienced during training, whereas diverse training improves the ability of a dyad to perform well in new situations. Copyright © 2011 John Wiley & Sons, Ltd.