To increase the accuracy of the visual evaluation method, we studied the significance of the two-sample preference test using the two-visit method, as proposed by Ferris. Advantages of this method include not only its ability to estimate whether the image quality of the sample is statistically the same but also its ability to distinguish the difference between samples more clearly than when using the conventional two-sample preference test or paired-comparison method. However, a disadvantage is that this method involves two observations of the same sample. When using chest images, even if a difference between the samples is not recognizable by the conventional method, a significant difference can be elicited with this method. The data obtained by this method can be evaluated with the paired-comparison method, so some samples may be ranked as well. Therefore, we verified that the two-sample preference test, which Ferris formulated, is useful in visual evaluations.