Hemoglobin (Hb), mean cell volume (MCV), and erythrocyte protoporphyrin (EP) are commonly used to screen for iron deficiency (ID), but systematic evaluation of the sensitivity and specificity of these tests is limited. The objective of this study is to determine the sensitivity and specificity of Hb, MCV, and EP measurements in screening for ID in preschool children, non-pregnant women 15-49 years of age, and pregnant women. Data from the National Health and Nutrition Examination Surveys (NHANES) (NHANES 2003-2006: n = 861, children three to five years of age; n = 3112, non-pregnant women 15 to 49 years of age. NHANES 1999-2006: n = 1150, pregnant women) were examined for this purpose. Children or women with blood lead ≥10 µg/dL or C-reactive protein (CRP) >5.0 mg/L were excluded. ID was defined as total body iron stores <0 mg/kg body weight, calculated from the ratio of soluble transferrin receptor (sTfR) to serum ferritin (SF). The receiver operating characteristic (ROC) curve was used to characterize the sensitivity and specificity of Hb, MCV, and EP measurements in screening for ID. In detecting ID in children three to five years of age, EP (Area under the Curve (AUC) 0.80) was superior to Hb (AUC 0.62) (p < 0.01) but not statistically different from MCV (AUC 0.73). In women, EP and Hb were comparable (non-pregnant AUC 0.86 and 0.84, respectively; pregnant 0.77 and 0.74, respectively), and both were better than MCV (non-pregnant AUC 0.80; pregnant 0.70) (p < 0.01). We concluded that the sensitivity and specificity of EP in screening for ID were consistently superior to or at least as effective as those of Hb and MCV in each population examined. For children three to five years of age, EP screening for ID was significantly better than Hb and similar to MCV. For both non-pregnant and pregnant women, the performance of EP and Hb were comparable; both were significantly superior to MCV.