We present a model for prey survivorship varying with average space size/prey width (Sp/Py) and total cover within an area (Ct/At), which are habitat complexity indices. The model predicts that prey survivorship is maximized at intermediate Sp/Py values, such that prey can fit through the spaces in a habitat, but their larger predators cannot. The model also predicts that prey survivorship increases with increasing cover (Ct/At), which interferes with predators' ability to detect prey. We deployed high-, medium-, and low-complexity artificial vegetation treatments with pit traps to determine if desert beetles respond to differences in habitat complexity consistent with our model's predictions for prey survivorship. We also deployed pit traps in natural vegetation and open sand to determine beetle microhabitat preference. The median number of both large and small beetles was higher in natural vegetation compared with open sand. The median number of large beetles was significantly higher in medium-complexity artificial treatments compared with both low- and high-complexity treatments. Prionotheca coronata, a common species of large beetle, was excluded from the high complexity treatments, as Sp/Py was <1 for them. This demonstrates that high-complexity habitats may exclude larger fauna, reducing community diversity. There was no difference in the mean number of small beetles captured in the different artificial complexity treatments. The results for large beetles are consistent with our model. Small beetles' distribution may be "predator-independent" since they did not respond to differences in complexity.