ESDU 91018 illustrates certain aspects of statistical analysis introduced in ESDU 91017 in relation to cruise specific air range. Before issuing guarantees to a customer, flight tests were carried out around the guarantee point to verify the specific air range that could be claimed without a risk of triggering a penalty. Because the first sample of data obtained was small it is analysed using Student's t distribution but is found, at the chosen confidence level, to indicate an unacceptable performance. A further sample of values was therefore measured with a view to improving the prediction accuracy. Data from the two samples are first checked with the F Variance Ratio Test to establish that they may be coalesced. The samples are then combined and Student's t distribution applied again. The results are still unsatisfactory but the comment is made that one point is clearly outside the general dispersion of the data. The deviation from the pooled mean is noted but it is explained that if the result cannot be faulted on sound engineering grounds it must stand and the conclusions drawn must be accepted.