The molecular detection of predation is a fast growing field, allowing highly specific and sensitive detection of prey DNA within the gut contents or faeces of a predator. Like all molecular methods, this technique is prone to potential sources of error that can result in both false positive and false negative results. Here, we test the hypothesis that the use of suction samplers to collect predators from the field for later molecular analysis of predation will lead to high numbers of false positive results. We show that, contrary to previous published work, the use of suction samplers resulted in previously starved predators testing positive for aphid and collembolan DNA, either as a results of ectopic contamination or active predation in the collecting cup/bag. The contradictory evidence for false positive results, across different sampling protocols, sampling devices and different predator-prey systems, highlights the need for experimentation prior to mass field collections of predators to find techniques that minimise the risk of false positives.