Current methods of auditory-perceptual evaluation of disordered voice quality using ordinal or interval scales have limited reliability and precision to quantify the magnitude of change in response to disorder progression or treatment. Matching tasks with synthetic comparison sounds have emerged as a more effective alternative, reducing biases and providing adjustable comparative values to represent perceived magnitude of change in quality. Previous studies have focused on the individual dimensions of breathiness, roughness, and strain separately and indicate covarying voice qualities may impact listener judgments. Here we investigate a three-dimensional matching (3-DMA) task where three variables, each representing a major quality dimension, can be adjusted concurrently. The comparison sound was a low-pass filtered sawtooth waveform (f o = 151 Hz) mixed with noise. Independent variables for matching were the signal-to-noise ratio for breathiness, amplitude modulation depth for roughness, and bandpass filter gain for strain. Listeners used sliders to adjust each variable, matching all three dimensions of 26 natural voice samples, selected to contain wide variation in all three dimensions. Preliminary results from five listeners demonstrated good reliability, similar to that observed in previous single-variable matching tasks for individual voice qualities. The results support the feasibility of the 3-DMA task. [Work supported by NIH R01DC009029.]