6. Mistakes to avoid

6.1 Don't apply a type 1 approach using a statistical distribution from a sector of activity other than the one under study.

As with database frequencies, a statistical breakdown can only be transposed if the sectors of activity are "similar". Indeed, the "weight" of causes may differ from one sector to another, due to operating or process conditions. For example, if a statistical distribution reports that on an offshore installation 10% of small breaches are linked to corrosion, this result may not be representative of an onshore installation (or at least it is not possible to demonstrate this). Applying this coefficient could lead to an erroneous calculation of the "weighted" leak frequency.