In actual statistical surveys, the inconsistencies within the record are often occurred due to incorrect response. The users may be confused and statistical agencies may have a problem of reliability on statistical data in this case. It is needed to detect and correct the unconvinced record without any special reasons. However, it is not simple to determine which item should be corrected in every failed record. In this paper we briefly introduce Fellegi-Holt method, apply to a business survey, and then discuss the problems for this trial editing.