An experiment was conducted to determine the effect of spatial audio on memory, focal assurance, perceived comprehension and listener preferences during desktop conferences. Nineteen participants listened to six, pre-recorded, desktop conferences. Each conference was presented using either non-spatial audio, co-located spatial audio, or scaled spatial audio, and during half of the conferences, static visual representations of the conferees were present. In the co-located condition, each conferees voice originated from directly above their image on the screen, and in the scaled spatial audio condition, the spatial separation between conferee voices was increased beyond the visual separation. Results showed that spatial audio improved all measures, increasing memory, focal assurance, and perceived comprehension. In addition, participants preferred spatial audio to non-spatial audio. No strong differences were found in the visual conditions, or between the co-located spatial condition and the scaled spatial conditions.