Iron-deficiency anemia and thalassemia are among the most common microcytic anemias. Differentiating these anemias by means of hemogram indices is imprecise. Powerful statistical computer programming now enables sensitive discriminant analyses to aid in the diagnosis. Laboratory results from 383 adults were examined retrospectively and grouped according to their original diagnoses: normal (n = 78); β-thalassemia (n = 134); α-thalassemia (n = 106); and iron-deficiency anemia (n = 65). Statistical analysis of results evaluated only RBC indices: RBC count, hemoglobin level, mean corpuscular volume, mean corpuscular hemoglobin, and RBC distribution width. Stepwise multivariate discriminant analysis determined those indices that best differentiated the 4 groups. The Fisher linear discriminant function for each group was calculated and tested casewise. Discriminant analysis identified mean corpuscular hemoglobin, RBC count, mean corpuscular volume, and RBC distribution width as the best set of indices for differentiating the 4 diagnoses. Casewise testing of the calculated Fisher linear discriminant function resulted in mean-weighted sensitivity of 80.4%. The present study demonstrates that a set of linear discriminant functions based on routine hemogram data can effectively differentiate between α-thalassemia, β-thalassemia, and iron-deficiency anemia, with a high degree of accuracy.