This paper analyzed standardized achievement test items collected from 581 5th grade children in a school district in Korea with the application of a Rasch model computer program. A total of 25 items measuring mostly listening ability were submitted to the Rasch analysis to find item fit, item difficulty, and item coverage of children`s ability. Item infit and item outfit values were calculated and reported. The analysis of misfit items was conducted and plausible explanations were made with reference to misfit items. The second analysis of item difficulty indicated that some cognitive dimensions including memory retention may have been possible causes of item difficulty. The third analysis of comparing between item difficulty and person ability indicated that some test items were so easy that the children were predicted to get most items correct. Most misfit items may have been due to picture drawings that were confusing to the children or due to poor constructions of conversations. Some implications for developing test items were suggested based on the results.