This study aims to examine a) whether vocabulary knowledge, captured in the Vocabulary Levels Test (VLT), is related to the performance on the five types of reading comprehension items tested in TOEFL, i.e., Guessing Vocabulary, Main Idea, Inference, Reference, and Stated Detail; and b) whether EFL learners with different levels of vocabulary knowledge (low, middle, and high) differ in their performance on these items. 213 participants were tested in this study and their scores were analyzed using stepwise multiple regression tests. Results show that performance in three of the TOEFL test items (Guessing Vocabulary, Stated Detail, and Main Idea) correlate to the overall performance in VLT. For participants at the highest proficiency level, only one test item (Guessing Vocabulary) was found to correlate with their performance in VLT. No correlation was found for low level and middle level groups. This study showed that the performance in the type of test items used is influenced by vocabulary knowledge and vocabulary proficiency level. It suggests that an indiscriminate inclusion of different types of vocabulary tasks in any given lesson irrespective of their vocabulary knowledge level, might not work effectively for learners. (Contains 1 figure and 7 tables.)