05 Juin 2018

Classification method of the GEIPAN

The classification method used by the GEIPAN (A/B/C/D1/D2) is based on 2 different aspects: the strangeness and the consistency  (see glossary). It requires:


1/The identification of research hypothesis that could explain the full strangeness of the sighting (as perceived by the witness) and the evaluation of their probability

  • Each hypothesis is based on one or several physical (such as stars or plasma) or psychological (such as perception effects, false memories) known phenomena

2/ The characterisation of the perceived strangeness (S, ranging from 0 to 1). It is a distance to what is “known”, measured by the complement to unity: 1 minus the probability of the strongest hypothesis. 

  • Ex: the probability of the strongest hypothesis, “hovering aircraft due to the relative motion of the witnesses, but with no visible wings due to the moon lighting”, is 0.4; the strangeness will then be 0.6
  • If the strangeness is above 0.5, the GEIPAN has no explanation (no hypothesis seems appropriate) 

3/ The characterisation of the consistency of the sighting (C=IxR, ranging between 0 and 1) using:

  • the amount of gathered Information (I) (number of witnesses, number and precision of the responses, presence or not of photographic materials…)
  • the Reliability (R) of the information sources (coherency of the testimony, credibility of the witnesses, existing links between the different witnesses…)

Finally, it requires to apply the following basic and common-sens principle: the more the strangeness is strong (i.e. the less the best hypothesis is probable), the more the consistency must be strong :

  • to validate the “unexplained” when the strangeness is above 0.5. D1 (strange case) or D2 cases (strangest cases).
  • to validate the explanation when the strangeness is below 0.5. A (almost  proved hypothesis) or B cases (probable explanation)
  • otherwise, no conclusion can be drawn from the sighting: C cases. 

The consistency and the strangeness can be drawn on a diagram in order to determine the final classification. For example, the cas having a consistency of 0.6 and a strangeness of 0.4 will be classified as B as shown below:


Note: in french, in our reports "E/Etrangeté" stands for "S/Strangeness" and "F/Fiabilité" stands for "R/Reliability".