Nocturnal predators are often overlooked in biological control studies, despite evidence that they can make important contributions to insect pest suppression in agroecosystems. Many sampling methods are only employed during the daytime hours due to limitations of time and labour. Additionally, different sampling methods can provide contrasting information about natural enemy community composition and relative abundance. Here, we use Aphis glycines and its arthropod predators as a model system to compare natural enemy community composition described by vacuum samples, direct observations and video observations across the diel cycle in soybean. All sampling methods identified several common taxa. Anthocorids were dominant in vacuum samples and direct observations, and both methods indicated that this taxa may be more active in the afternoon. In contrast, anthocorids were recorded infrequently on video, possibly due to their small size. On video samples, lacewing larvae were the most active taxa during the day and lacewing larvae, spiders, opiliones and carabids were the most active taxa at night. We directly observed 22 predation events on soybean aphid: 17 by anthocorids, two by chrysopid larvae, and one each by a coccinellid, spider and predatory mite. The differences between the sample methods suggest that vacuum samples may represent predator abundance more accurately, while video data may miss small predators, but can be used to better assess relative time spent foraging.