The purpose of the current study was (a) to assess the effectiveness of the lexical-set (LS) and the semantically-unrelated (SU) vocabulary instruction, separately and relative to each other, and (b) to assess the differential effects of the two methods for students of lower and upper English proficiency levels. Two intact EFL classes were assigned randomly to the LS or the SU treatment. Each treatment group then received special instruction in vocabulary. The students' gains of target words were measured in terms of vocabulary breadth (VB) and vocabulary depth (VD) knowledge, using the Vocabulary Knowledge Scale (VKS). The study provided some initial evidence to suggest that although students in both instructional methods significantly gained in their knowledge of target words with regard to their VB and VD knowledge, both upper and lower level LS students reached a significantly higher level than their peer SU ones. The upper level students, however, made greater gains than their peer lower level ones. (Contains 7 figures and 9 notes.)