Recent interest in studying the nonlinear effects of intelligence, surveillance, and reconnaissance in combat models has prompted researchers to employ vertical aggregation in object-oriented simulations. Traditional horizontal aggregation falls short for its inability to provide accurate means for nonlinear functions. Averaging a group of objects that exhibit nonlinear behavior provides a linear approximation to the mean, which is not necessarily the expected value of the underlying nonlinear function. Vertical aggregation explicitly models individual objects, thus preserving their nonlinear behaviors. In this research, a validation procedure is derived to study the aptness of vertical aggregation methods. Validation is carried out by comparison with a control, considered model truth, since it contains no vertical aggregation. Response surfaces are mapped for the control and the hypothesized model. Family confidence intervals are used to test the hypothesis that the difference between the two is zero. An illustrative example is presented using a homogeneous combat scenario embellished with experimental factors. Metamodels are derived using the method of least squares and validated prior to drawing inferences. Simultaneous inferences are drawn between the ith regression coefficient of two models. The results suggest fascinating avenues for further study.