Abstract Adherence to a vegetarian diet has been hypothesized to be a factor in the onset and maintenance of disordered eating behavior; however, evidence to support this assumption has been largely mixed. The two studies presented here sought to address the causes of inconsistent findings in previous research, including: small samples of true vegetarians, lack of appropriate operational definitions of “vegetarianism”, and uncertainty about the appropriateness of existing assessments of eating behaviors for semi-vegetarians. Study 1 assessed eating behaviors in the largest samples of confirmed true vegetarians and vegans surveyed to date, and compared them to semi-vegetarians and omnivores. Semi-vegetarians reported the highest levels of eating-related pathology; true vegetarians and vegans appeared to be healthiest in regards to weight and eating. Study 2 examined differences between semi-vegetarians and omnivores in terms of restraint and disordered eating and found little evidence for more eating-related pathology in semi-vegetarians, compared to omnivores. Semi-vegetarians’ higher scores on traditional assessments of eating behaviors appeared artificially inflated by ratings of items assessing avoidance of specific food items which should be considered normative in the context of a vegetarian diet. Findings shed light on the sources of inconsistencies in prior research on eating behaviors in vegetarians and suggest that semi-vegetarianism – as opposed to true vegetarianism or veganism – is the most likely related to disordered eating.