Roller dung beetles play a pivotal role in the nutrient distribution in soil and secondary dispersal of seeds. Dung beetles are sampled either using a dung-baited pitfall trap or an exposed dung pile on the ground. While the former method is useful for a rapid survey of dung beetles, information on the ecology and behaviour of dung beetles can be lost, which the latter method provides, but underestimates species diversity due to its inefficiency in trapping rollers. Efficiency of a new method for sampling rollers—installing guarding pitfall traps around dung piles—is assessed in three habitats—contiguous tropical rainforests, fragmented forests, and disturbed used home gardens—and two diel periods—day and night. Five guarding pitfall traps were installed at 50 cm radius around dung piles. About 98% of the total rollers were sampled in pitfall traps. The habitats were similar when the roller catches of only dung piles—conventional approach—were analyzed, but were different when the rollers of guarding pitfall traps were considered. The roller abundance was negatively affected by forest fragmentation and land-use change. About 98% of the rollers were collected at daytime. Using guarding pitfall traps around dung piles is highly recommended for dung beetle diversity studies.