The Coral Reef Assessment and Monitoring Program (CRAMP) of Hawai'i was established in 1998 to monitor long-term changes in coral reef benthic communities around the state. Development of the methodology involved analysis of results from previous monitoring programs in Hawai'i to determine precision and statistical power of various methods to detect change. Additional field trials were conducted to examine factors such as repeatability, appropriate transect length, number of transects, number of samples per transect, cover estimation techniques, observer variation, as well as time and financial constraints. Benthic monitoring methods used previously in Hawai'i generally showed low statistical power for detecting change due to low precision and small sample size. Field trials indicated that repeatability of conventional techniques using transects or quadrats had high variation and consequently low statistical power unless efforts were made to reposition the sampling units with greater precision. Longer transects (e.g., 25 and 50 m) had higher variability than shorter transects (e.g., 10 m), suggesting that smaller sampling units were more appropriate for the habitats sampled. Variability among observers analyzing the same data was low in comparison with other sources of error. Visual estimation techniques showed low initial cost but were inefficient per survey. Digital video required the highest initial monetary investment but yielded the greatest quantity of data per survey with sufficient quality. The cost effectiveness of the digital video method compared with other techniques increased with more surveys and in more remote situations where logistical expenses were incurred. A within-habitat stratified random sampling design was implemented for the CRAMP design. Fixed transects were chosen to reduce temporal variance and allowed efficient resurveying under the high-wave-energy field conditions typically found in Hawai'i. The method was designed to detect an absolute change of 10% in benthic cover with high statistical power using 50 points per frame, 20-30 frames per transect, and 8-10 transects per depth. Fixed photoquadrats with high precision and high resolution were included in the design to allow detailed monitoring of coral/algal growth, recruitment, and mortality.