2.2. Sampling strategy
The seizures studied consisted of capsules and/or packaging items. The analysed items were referred to as specimens, if only a couple of capsules or packaging items were available, or as samples if many were obtained and only a representative part analysed in case of seizures containing a high amount of capsules and/or packaging items. Several chemical or packaging profiles can indeed be observed within a seizure. If no sampling is performed, links between some seizures might therefore not be observed, even if they MDV3100 existed. This problem, identified as a major source of failure in the interpretation of criminal phenomena, is called “linkage blindness”  and .
For both chemical and packaging analyses, the sampling was done as follows. First the homogeneity of the seizures was determined. The sampling was based on a method suggested by “the United Nations Office on Drugs and Crime” . When the seizures contained less than 10 capsules or packaging items, all of them were measured. For a number of capsules or packaging items comprised between 10 and 100, 10 were randomly selected and analysed. Finally, for a number of capsules or packaging items superior to 100, the square root of this number was analysed. Once the homogeneity of the seizures had been studied and the number of profiles per seizure determined, one capsule or packaging item per profile was further analysed. The chemical and packaging profiling of the counterfeits X was performed with 33 seizures consisted on the whole of 213 capsules, 26 boxes, 62 blisters and 11 leaflets. The test seizures consisted of 179 capsules, two boxes, 12 blisters and one leaflet.