Samenvatting

Recent years have shown a growing effort in the evaluation of DNA given activity level propositions. Yet, its implementation remains challenging for forensic scientists due to the complexity of assessing relevant DNA TPPR mechanisms. One particular challenge is the sparsity of relevant experimental data in general and their applicability in the context of the case at hand. This difficulty is most apparent when considering the presence and quantity of prevalent and background DNA as the variables affecting them are highly specific and case-dependent, complicating their assessment based on literature data alone. It would hence be advantageous to assess prevalent and background DNA using information specific to the case, ideally gathered at the crime scene. We introduce the concept of contextual sampling: the targeted collection of additional samples from the surroundings of crime-related items to assign probabilities on prevalence and background based on case-relevant data. We identify several categories of contextual samples and demonstrate how these can inform sampling strategies and be integrated into Bayesian networks. Our findings show how contextual sampling allows for case-specific probability assignment, thereby reducing dependence on less-representative literature data. We also address practical considerations – including sample number, sampling location, and analytical strategy – and discuss limitations such as resource demands and uncertainties tied to small sample sizes. We conclude that contextual samples can be collected in anticipation of relevant alternative scenarios. This will provide forensic scientists with a valuable tool for casework, offering more nuanced and case-specific evaluations given activity level propositions. Further research could address operational protocols to optimize its implementation in forensic practice.