Predicting team output using indices at group level

The present study explores the usefulness of dyadic quantification of group characteristics to predict team work performance. After reviewing the literature regarding team member characteristics predicting group performance, percentages of explained variance between 3% and 18% were found. These studies have followed an individualistic approach to measure group characteristics (e. g., mean and variance), based on aggregation. The aim of the present work was testing whether by means of dyadic measures group output prediction percentage could be increased. The basis of dyadic measures is data obtained from an interdependent pairs of individuals. Specifically, the present research was intended to develop a new dyadic index to measure personality dissimilarity in groups and to explore whether dyadic measurements allow improving groups' outcome predictions compared to individualistic methods. By means of linear regression, 49.5 % of group performance variance was explained using the skew-symmetry and the proposed dissimilarity index in personality as predictors. These results support the usefulness of the dyadic approach for predicting group outcomes.