In this paper, we investigate some of the most common probability/possibili
ty transformations found in the literature. Some meaningful criteria, on wh
ich the performance of these transformations are evaluated and compared, wi
ll be proposed. It concerns respectively the specificity order, i.e., the s
pecificity Value of the resulting distributions, the strength of the prefer
ence preservation, or in other words, how is the behavior of the result whe
n initial points are close to each other with some strength or according to
some function. The third one is related to the result of the transformatio
n where initial distribution tends towards the ignorance state. Finally, we
study the influence of incomplete data (subnormal distribution) to the res
ult of each transformation. This analysis will be illustrated with some the
oretical foundation and followed with some numerical examples.