We investigated whether “embodiment” of objects used in a problem-solving task (i.e., whether they have a bodily shape) would have a detrimental effect on learning to solve that problem through practice or through studying video-based modeling examples. A 2x2 design with factors Training (Practice/Example study) and Embodiment (Present/Absent) was used (N = 80). Results showed a large main effect of Training on effort investment in learning and on retention test performance, with Example study leading to higher scores with lower investment of effort during the learning phase than Practice. Numerically, Embodiment seemed to have an effect, with participants practicing/studying the task with embodied objects (plastic animals) performing worse on retention than participants practicing/ studying with non-embodied objects (discs), but this did not reach statistical significance. A new study with more power and an additional control condition is currently being conducted and results are expected to be available well before the conference.