ABSTRACT
P. Thompson argues that the effects of learning loss are possibly much less than others have estimated, and that at least a proportion of those losses may be attributable to other reasons such as labour turnover, and 'aggregation and product mix'. In terms of calibrating any lost learning due to disruption, the Segmentation Technique has the distinct advantage that estimators can use a multi-linear transformation and regression to evaluate all variables. If estimators want to force the issue of additional 'logical' drivers to model their Learning Curve to the actual data, then they can always look for a solution using Microsoft Excel's Solver. Estimators can use Microsoft Excel's Solver to determine an appropriate offset before running a transformed multi-linear regression to establish the statistical significance of the result. As the G. Anderlohr technique for lost learning invokes a Generalised Power Function with an offset x-variable, estimators cannot easily use linear regression in isolation.
