To eliminate problems in determining parameter end points, a maximum likelihood methodology is extended to multi-lot, multi-parameter data. The present methodology makes more efficient use of the available data than previous methods. It is easily computerized, does not require interpolation to standard stress and bias conditions and reveals problems with data which might otherwise yield misleading solutions. A typical example is solved to show how the present method compares with past procedure. The extension of the methodology to more complex situations will be discussed.