This study attempts to establish the ability of a panel of five judges with varied mathematics background to distinguish between two types of mathematical tests by separating their component items when they are presented in a mixed pool of aptitude and achievement tests. Typically, the two tests show high correlation. The judges showed about 70% success in their judgments, which points to a clear distinction between the tests. In examining the items which were consistently misclassified, the general conclusions were that certain achievement items exhibit aptitude-item characteristics, and vice-versa. The use of more formal specifications might be useful in controlling the correlation between the tests. However, as mathematics curricula introduce formerly advanced topics at earlier times, such specifications will probably become increasingly difficult to establish. (Author/AL)