The present study aims to demonstrate how the estimation of vocabulary size might be affected by two neglected factors in vocabulary size tests. The first factor is randomization of question sequence, as opposed to the traditional high-to-low frequency sequencing. The second factor is learners' confidence in choosing the correct meaning for a given target word. A new online vocabulary size test was developed for the purpose of the study with the two factors in mind. The results of the test revealed that (1) randomizing question sequences did not have significant effects on the score of the vocabulary size test and (2) even though the learners who had a mastery level of 8000 words showed higher confidence in high frequency words than the learners with a smaller vocabulary, such confidence faded as early as 4000 frequency level of JACET 8000. The findings are discussed in detail in terms of the scale or the length of vocabulary size tests as well as the need for incorporating confidence in the estimation of vocabulary size. [For full proceedings, see ED565044.]