The quality of fish census data generated by REEF volunteers of various experience levels is examined and compared with data generated by experts. The similarity in species reported, the similarity in abundance category recorded, and the statistical power to detect change are used in the comparison. Species composition and structure is comparable between the skill levels. When sighted, most species are reported with similar abundance categories. The ability to detect change in many species is better for the nonexpert data, a result of higher survey effort for the nonexperts. Species that are cryptic or rare have low power (change had to be large before it could be detected) regardless of the skill level, but are generally better using expert data.