Careful considerations are necessary when there is a need to choose an anchor test form from a list of old test forms for equating under the random groups design. The choice of the anchor form potentially affects the accuracy of equated scores on new test forms. Few guidelines, however, can be found in the literature on choosing the anchor form. Five indices were proposed in this study to aid the selection and their performances were evaluated through a real-data-based simulation. The results shed some light on how to choose an anchor form in practice.