[A conceptual framework is suggested for integrating fixed effects and random effects models into one framework. In that framework, the pertinent distribution is a convolution of two distributions; one is a degenerate distribution. A method is suggested and analyzed for separating between the two distributions when the second distribution is normal.]