This paper proposes Hierarchical Linear Models as a procedure for investigating reciprocal interaction between individuals and their group context. This interaction is an important component of many theories of organizational behavior. A study of task-performing laboratory groups examined the relationship between individual affect and group cohesion over seven measurement occasions. A series of three HLM analyses was used to estimate the relationship between group cohesion and individual affect over time. The first analysis estimated a cross-level model of the impact of group cohesion of levels of individual affect in groups. The second analysis used the residual from the cross-level analysis as a predictor of subsequent levels of group cohesion for each of the seven time periods. The third analysis used HLM procedures to pool the results of the previous analyses over the seven measurement periods. Results of the HLM procedure provide a method for linking within-group measures to subsequent group level outcomes.