Based on the acoustic experiment conducted on two groups of speakers, this study aims to account for the variation in VOTs of English voiceless stops in four contexts, namelywhen they are placed in word-initial vs. word-medial position and also when they are placed in the onset of the stressed vs. unstressed syllable. The results of the acoustic analysis indicate that the VOT difference between the stressed and the unstressed syllables is statistically significant in both speaker groups regardless of their positions in word, but the difference turned out to be greater in the word-medial position than in the word-initial position. Unlike the metrical factor, however, the VOT difference between the word-initial and the word-medial positions is statistically significant in the unstressed syllable only. In the stressed syllable, the VOT varies as a function of the position in word, and the difference is statistically significant in the Korean speakers, but not in the native speakers of English. This comes to the point that the main effect on the VOT value differs depending on the speakers, namely that it is the prosodic prominence in the case of the Korean speakers, whereas it is the metrical prominence in the case of the native speakers of English that plays a crucial role in defining the VOT value. The results further support the claim that the VOT difference is maximized in the prosodically as well as metrically weak position, and this lends supporting evidence to the phonetic effect of perceptual enhancement in the weak position.