This study intends to identify the lexical features that present particular difficulties to English native speakers learning Korean; to classify those lexical errors in terms of their type and frequency; and to provide possible explanations for the cause of those lexical problems. The data used in this study come from 141 written examination papers administered at three universities. The subjects selected in this study are 71 second-and third-year students, who are native speakers of English or are believed to have English as their first language. In all, 305 lexical errors have been identified for analysis. Among the 11 error categories identified in this study, wrong word choice caused an overwhelmingly high percentage of errors. Other error categories with high frequency were: confusion by semantic similarity, overgeneralization and literal translation. More than half of the total errors involved nouns, followed by verbs and adjectives. And a large number of the errors produced were intralingual. The study concludes with discussions about teaching and research implications.