Acrylamide is commonly found in various foods. Cancer studies in rats have reported increases in tumors of the thyroid, mammary tissues, tunica vaginalis of the testis, and sometimes tumors at other sites. We review relevant studies on acrylamide’s DNA toxicity, tumor formation and the manner of its tumor formation. We find, as do others, that glycidamide (a metabolite of acrylamide) causes point mutations, but acrylamide does not. We also find that thyroid tumors are most consistently sensitive in rats, being evoked in each of four long-term experiments. We evaluate the common manners of this tumor formation in the thyroid, including both mutagenicity and thyroid growth-stimulation. Consistent with the overall weight of the evidence, we conclude that both of these manners or modes of action may be occurring. We conservatively assume that the mutagenic mode of action determines the low-dose–response and we conclude that growth stimulation likely dominates the response at higher doses. Following US EPA guidelines, we determined that the probit model best reflects the overall data set; this model is also preferred because it better reflects the underlying “decoupled” biology of contributions from potentially two modes of action. We use the probit model to identify a health-protective, linear cancer slope factor (SF) of 0.030 (mg/kg-day)−1 for the low area of the dose–response curve associated with possible mutagenicity. We also identify a Reference Dose (RfD) in the range of 0.05–0.02 mg/kg-day for the high area of the dose–response curve associated with the growth stimulation. This latter value can be used to determine the upper range of risk. This dose–response assessment is briefly summarized in light of other related work.