Numerous research found a significant role of emotion in problem solving. However, how emotionsshape collaborative problem solving remains an open question. In this study, we sought to determinerelationships between emotions shared by problem solvers and collaborative processes involved. Twenty-two participants were randomly assigned to same-sex pairs and asked to complete a problem solving task.At any time, participants could voluntarily share their emotions with their partner, using a reporting tool.Their speech was transcribed and categorised into a series of meaningful collaborative categories. Twochi-square tests of independence (due to test constraints, only positive emotions were considered)confirmed a relation of dependency between collaborative processes and positive shared emotions, bothbefore sharing emotions (χ2(96) = 218.41, p<.001) and after (χ2(96) = 131.14, p<.01). Results underlinethat problem solvers do not randomly display positive emotions during collaboration. Rather, certainemotions appear, more than chance, in conjunction with certain collaborative processes. Adjustedstandardized residuals (showing which associations are the source of the statistically significant chi-square tests) were calculated and strongest associations identified. For example, acceptance ofproposition increases while outside activity decreases, both significantly, before the sharing of theemotion interested. We propose three main hypotheses concerning the potential role of shared emotionsin collaborative problem solving, namely sharing one's own mindset about task achievement, clarifyingmessages and initiating new stages in the collaboration. Future perspective, especially the considerationof negative shared emotions and relational processes, are developped to gain a finer understanting intocollaborative problem solving dynamics.