Laboratory data obtained from blood analysis by an electronic cell counter is a routine test along with physical evaluation. We assessed the hematological characteristics of schoolchildren to establish the prevalence of anemia in Istanbul. The study was performed on 1,600 students between the ages of 6-16. Hematological parameters of all the students were measured by an electronic cell counter on blood taken intravenously. Anemia prevalence was primarily determined according to Dallman's criteria. The results showed that if hemoglobin (Hb) was used as the primary criterion to define anemia, 423 students (27.6%) were found anemic; if hematocrit (Hct) was used, 625 students (40.8%) were found anemic and if mean corpuscular volume (MCV) was taken as a criterion to define microcytosis, 121 students (7.9%) remained below the normal level. The difference of anemia prevalence was high (p < 0.001) when two evaluation criteria (Hb or Hct) were compared. Furthermore it is known that a 3% difference occurs when automated and spun Hct values are compared. When a correction was made by adding this 3% to our series' Hct values, the anemic students' number decreased to 418 and the ratio became 27.3%, quite similar to Hb criterion ratio. Population based screening surveys for anemia must be designed and compared cautiously taking into account the method used (electronic counters or gravity based centrifuges), otherwise misleading results can be achieved. We evaluated the data of the survey according to different cut-off values and tried to select the proper one for our conditions.