n82% of multimodal input constructions involved draw and speak, while point and speech represented only 14%. nOf draw and speak constructions, 42% were simultaneous, 32% sequential, and 12% were compound (i.e. draw graphic, then point to it while speaking).