The purpose of this review of literature was to identify research findings for designing assessments in singing accuracy. The aim was to specify the test construction variables that directly affect test performance to guide future design in singing accuracy assessment for research and classroom uses. Three pitch-matching tasks--single pitch, interval pitch, and patterns--plus song-singing have each been shown to be discriminators of singing accuracy. Teachers and researchers should consider incorporating more than one task type when assessing children, since performance on one song or pitch-matching task cannot easily be generalized to others. Item-level and task-level difficulty must be taken into consideration when designing singing tests. In addition to task type, the variables of singing voice development, contextual presentation, model characteristics, range, and text must be a part of a priori decision making.