A major problem that has dogged research into body image disturbance in anorexia nervosa concerns the diversity of measurement procedures employed in studies of body size estimation. Many researchers believe that results obtained with different procedures cannot meaningfully be compared with each other, because the methods used measure different aspects of the body image. A meta-analysis of 33 body size estimation studies was conducted to investigate whether methodological differences do indeed influence outcome. This analysis revealed a general overestimation of body size among anorexia patients. Although a difference in mean effect size was found between studies using Body Part and Whole Body methods, this difference was absent when only the most commonly used Body Part method, the visual size estimation procedure, was considered. Among Whole Body methods, silhouette methods could not be distinguished from the rest in terms of effect size. It was concluded that Whole Body methods and the visual size estimation procedure assess correlated aspects of the body image.