Dense layers of red abalone shells (“red abalone middens”) have been heavily studied along the California coast for their implications for pre-contact subsistence, paleo-sea surface temperatures, and nearshore ecology. However, shells in these deposits are typically broken which lessens their scientific value. To compensate for this problem, we evaluate four alternative methods for predicting the size of whole abalone shells from fragments. The first was a template made up of the outlines of complete modern and archaeological shells. The others were based on regression formulas that rely on measurements of shell features that potentially track with shell growth: (1) maximum diameter of the last developed aperture, (2) distance between the last two developed apertures, (3) maximum diameter of the muscle scar. We conducted a blind study with archaeological laboratory volunteers to evaluate the accuracy of these methods, and found that the template outperformed the allometric methods, and, of the latter, the muscle scar measurement was the most accurate. Our results provide a framework for future studies to assess variation in abalone size from fragmented shells as a proxy for past climate change and harvesting pressure. They also provide a methodological approach for reconstructing sizes of other shell species found in midden deposits elsewhere.