Recently, as the use of the Internet and the Web becomes universal, there has been a great deal of efforts to use it for education. The advantage of a Web-based lecture is that people can take it anywhere at a low cost anytime. Therefore, its prospect is very bright and its potential customers abound. The demand for java programming education using the Web is increasing. This paper compares and analyzes the learning effects of Web-based and traditional lectures in a java programming course, using statistical inference. Based on the analysis result, it concludes that the learning effects of the Web-based lecture are less effective than those of traditional lecture. Also, it recognizes weak points of Web-based lecture in java programming education and proposes a scheme for enhancing Web-based lectures.