Invertebrates are essential for ecosystem functioning and have received much attention in biodiversity, conservation and ecological studies. Numerous collection techniques for assessing arthropod diversity exist, but there is little understanding of the consequences of choice of any particular method. This information is essential for project planning and for the interpretation of results. Here we assess sweep netting and suction sampling, two widely used methods for sampling foliage arthropods, and compare them in terms of their efficacy and effort in surveying arthropod biodiversity. Arthropods were collected in paired transects in shrubby fynbos vegetation in the Cape Floristic Region, a biodiversity hotspot. Overall, the vacuum sampling collected significantly more species and individuals than sweep netting, although this was taxon specific. Use of the vacuum sampler significantly increased the time spent in the field per transect and it also collected significantly more dead plant biomass than the sweep net, thus requires a longer period to sort samples afterwards. Sweep netting caused significantly more damage to plants, an important consideration when working in sensitive ecosystems. Assemblage composition of arthropods collected by the two methods were significantly different with only about half of all arthropod species collected shared between the two methods. Our results show that overall, the vacuum sampler outperformed the sweep net in terms of arthropod species richness and abundance and minimising damage to vegetation, but that it takes considerably longer to collect the relevant data. When more representative biodiversity surveying is needed, the use of both techniques is encouraged.