Two is not safer than one
The UK CAA needs to understand that their response to the paradigm that two engines are safer than one, though not true, is a natural reaction by an organisation that lives by established patterns and rules. It is a response that it so aptly described by Joel Barker in the New Business of Paradigms. The culture in large bureaucratic organisations is such that people filter out information that does not fit their view of aviation and often can only see things that they want to see.
The UK CAA has found it very difficult to accept the data on the accident rates Single Engine Turbine Aircraft v. other Light Twin Aircraft. Emotions continue to take precedence over the science and faced with the inevitable problem of not being able to change the data the only remaining alternative is for the UK-CAA to challenge the statistics.
In the light of criticism from the UK CAA that the data set was not a sufficiently large enough sample an alternative statistically rigorous analysis method was sought that could be applied to the available data using a Bayesian analysis, where data from one set is used to predict the outcome of future experiments comparable to the other data sets, but assuming that they have the same value of the Poisson parameter as the original set.
This analysis indicates that the actual Single Engine Turbine Aircraft data set is likely to be significantly better than that for the Twin Engine Turbine Aircraft. This can be stated with even increasing confidence as the Single Engine Turbine Aircraft hours build up without experiencing any power loss related fatal accidents. The data set again confirms that Single Engine Turbine Aircraft have at least as good as, and probably better, an accident record when compared with the Twin Engine Turbine Aircraft.
It is simply a very dangerous situation, when faced by evidence for the need of change, the UK-CAA continues to ignore the evidence. Unfortunately it will probably need a major fatal accident in a light twin aircraft for the UK- CAA to react.