The present study attempted to replicate the previous results of Hu et al. (Learning and Instruction 35:85-93 2015) and further examined the boundary condition of tracing gesture whether its cognitive effect is merely comparable with other attention-guiding means, i.e., textual attention cueing, in two different learning tasks in nature. In two experiments, 11- to 13-year-old students were asked to study the worked examples on angles relationships involving parallel lines (experiment 1) and laws of exponents (experiment 2) either without tracing instruction, with textual attention cueing instruction, or tracing instruction. The tracing group outperformed the other two groups on a subsequent test and reported lower levels of test difficulty in experiment 1 but not in experiment 2, suggesting that the cognitive effect of tracing gesture was beyond guiding attention in the learning from highly visuospatial-based worked examples but not in the materials that are less obviously visuospatial in nature. The present findings suggested that tracing gesture offers a simple yet effective embodied technique that may further enhance the worked example-based learning through its cognitive effect specifically on the task involved with spatial information.