This article discusses visual techniques for detecting test items that would be optimal to be selected to the final compilation on the one hand and, on the other hand, to out-select those items that would lower the quality of the compilation. Some classic visual tools are discussed, first, in a practical manner in diagnosing the logical, illogical, and anomalous patterns in item responses, and second, a new way of illustrating the behavior of the items, cut-off curve, is introduced. All the illustrations in the article are accessible even without specific software packages; they are made by using basic spreadsheet software.