This article aimed to clear that the systematic bias according to the length of work time exists between estimated work time by self-reported questions used mainly in measuring and calculating the length of work time because of strong points of easy in measuring and so on and diary work time by time-diary method used because of the strong point that can measure more accurate lifetime by recording various activities of respondents during 24 hours in the stream. As the result that analyze the data from Lifetime Use Survey in 2004, the result like the contradiction came that the tendency of overstating work time is rising according as estimated work time increases via estimated work time, whereas the tendency of understating work time is rising according as diary work time increases via diary work time. The reason that the opposite results come despite the data from the same survey is that random errors act in the opposite directions by regression to the mean. Therefore, we cannot emphasize that a man working long hours tends to exaggerate his work hours by the result via estimated work time. That is, the fact that the systematic bias by the increase of work time does not exist is confirmed, and therefore, it is also impossible to raise questions about the accuracy of the data through estimated work time by self-reported questions from the evidence of the existence of that bias.