The capabilities of additive manufacturing (AM) open up designers’ solution space and enable them to build designs previously impossible through traditional manufacturing (TM). To leverage this design freedom, designers must emphasize opportunistic design for AM (DfAM), i.e., design techniques that leverage AM capabilities. Additionally, designers must also emphasize restrictive DfAM, i.e., design considerations that account for AM limitations, to ensure that their designs can be successfully built. Therefore, designers must adopt a “dual” design mindset—emphasizing both, opportunistic and restrictive DfAM—when designing for AM. However, to leverage AM capabilities, designers must not only generate creative ideas for AM but also select these creative ideas during the concept selection stage. Design educators must specifically emphasize selecting creative ideas in DfAM, as ideas perceived as infeasible through the traditional design for manufacturing lens may now be feasible with AM. This emphasis could prevent creative but feasible ideas from being discarded due to their perceived infeasibility. While several studies have discussed the role of DfAM in encouraging creative idea generation, there is a need to investigate concept selection in DfAM. In this paper, we investigated the effects of four variations in DfAM education: (1) restrictive, (2) opportunistic, (3) restrictive followed by opportunistic (R-O), and (4) opportunistic followed by restrictive (O-R), on students’ concept selection process. We compared the creativity of the concepts generated by students to the creativity of the concepts they selected. The creativity of designs was measured on four dimensions: (1) uniqueness, (2) usefulness, (3) technical goodness, and (4) overall creativity. We also performed qualitative analyses to gain insight into the rationale provided by students when making their design decisions. From the results, we see that only teams from the restrictive and dual O-R groups selected ideas of higher uniqueness and overall creativity. In contrast, teams from the dual R-O DfAM group selected ideas of lower uniqueness compared with the mean uniqueness of ideas generated. Finally, we see that students trained in opportunistic DfAM emphasized minimizing build material the most, whereas those trained only in restrictive DfAM emphasized minimizing build time. These results highlight the need for DfAM education to encourage AM designers to not just generate creative ideas but also have the courage to select them for the next stage of design.