Effects of Compton scatter has been investigated for quantitative single-photon emission computed tomography. Using a technique proposed by Ichihara and Ogawa (triple-energy window method) as a gold standard, effects of correcting the scatter has been evaluated for 5 different phantom configurations, as well as a clinical cerebral blood flow study with use of 123I-IMP. Without the scatter correction but with a correction for attenuation, a conventional reconstruction provided a non-uniform distribution for uniform phantom configurations, and the quantitative pixel counts being highly dependent on size and shape of the objects. On the other hand, a uniform distribution and quantitative pixel counts that were independent of size or shape of the object were obtained by applying the scatter correction using a conventional attenuation correction technique with use of a theoretical mu value of 0.146 cm-1, thus suggesting importance of the scatter correction. Similar results were observed without the scatter correction, if an empirical value of mu of 0.07 cm-1 was used in the attenuation correction. However, this procedure was found to decrease the image contrast between the high count and low count regions. Significant underestimation of 20-30% was caused in cortical gray matter regions, and overestimation of 20% in the white matter regions. In practical clinical studies, magnitude of the error is still unknown, and is probably dependent on shape, size and radioactivity distribution of the object. A further systematic study is required in order to investigate significance of the scatter correction in real clinical studies.