Expressing link values as a LR clarifies the limit between forensic science results and contextual information (prior probabilities and alternative intelligence) which are in principle within the competence of different professionals.
Comparatively, the deter-ministic approach requires a mixing of both forensic and contextual information in order to define the decision threshold [25] but is more straightforward and pragmatic in the sense that interpretation delivers a ‘black or white’ result associated with false positive and false negative rates.
Such a result can be easily handled and communicated, and is therefore appreciated when subsequent timely decisions have to be made.
The disadvantage of the Bayesian approach is the fact that no decision is made, which prevents reasoning a step further in the process (e.g. to form groups and classes among linked profiles that may facilitate the organisation and exploration of information as well as the elicitation of working hypotheses).
Although assessing uncertainty may be important from a purely scientific point of view, it raises many difficulties in terms of information processing and requires sophisticated information management techniques such as fuzzy sets for instance [30].
This can be a drawback when facing large and and/or complex datasets where analysis and subsequent communication necessitate simplifications.