Assignments were altered to determine if interspersing additional briefer and easier mathematics problems enhanced students' accuracy on longer, more difficult target problems across two types of mathematics assignments. Students (N = 52) from three fifth-grade classes completed six math assignments incorporating two task demands and three interspersal ratios. The interspersal ratios were no interspersal, one interspersed problem per three target problems (1:3), and one interspersal problem per one target problem (1:1). For the written (low-attention) assignments, students completed problems via paper and pencil. For the cognitive (high-attention) assignments, the experimenter read problems aloud and students had to compute mathematics problems in their heads, using paper and pencil only to record answers. Results showed that for the cognitive assignments, target problem accuracy was significantly higher for the 1:3 ratio relative to the other two ratios. On the written assignments, target problem accuracy levels were significantly higher on the 1:1 ratio assignments relative to the no interspersal assignments. Discussion focuses on applied and theoretical implications of the current results. (Contains 3 tables.)