The data on the mature equipments and the proposed design must be compared and significant differences must be identified. An engineering analysis of each of the differences must then be carried out, and the adjustment that should be made to the relevant reliability value must be assessed. Ground rules should be established, wherever possible, for making such assessments and should be stated in the analysis. For example, state of the art might be divided in four categories as follows:

Each of these categories might be given a weighting factor that could then be used to adjust reliability values.

It may not always be possible to define clear-cut categories for all the various data concerned, but it is most important that the reasoning behind all reliability weighting factors is clearly stated in the analysis.

Care must be taken to make due allowance for any differences in design, development and production resources. Clearly this will be more difficult if the mature equipment being used for comparison originated from different sources. This can lead to large errors in the prediction.

Due to the qualitative nature of the data comparison, the results may frequently depend upon the assumptions made. In general, therefore, reliability values should be assessed for both the best case and worst case assumptions.