A simple and routine example can be used to illustrate this framework. A burglary was committed in a factory and a DNA specimen was collected at the scene. When checked against the national database, a match was obtained with a profile related to one case included in a series of house burglaries committed two years earlier. In the deep level, this match can be assessed as the presence of a link (deterministic sub-model) and as a very high likelihood ratio in favour of a link (Bayesian sub-model). In the working level, two hypotheses were drawn from the analysis: ‘WHa: there is a link between both traces since they have a common source’; ‘WHb: there is no link between both traces since the result is due to a random match’. Two scenarios may explain WHa: ‘a serial burglar deposited both traces’ or ‘a person who has legitimate access to both locations (house and factory) deposited both traces’. Each working hypothesis guided different decisions and operations. Investigations following WHa showed that the owner of the house was also an employee of the factory.