When all the responses for an individual item have been scored, the system automatically submits all item scores assigned during scoring, along with their edit records, to a queue to be transmitted to the mainframe. NAEP instrument development staff then initiate a system job to transmit all scoring data to be matched with the original student records on the mainframe. A custom edit program matches the edit records of the scoring files to those of the original edit records on the mainframe. As matches are confirmed, the scores are applied to those individual files. After completion of this stage, all data collected for an individual student are located in one single and complete record (file) identified by the edit record.
An important quality control component of the image-scoring system is the inclusion, for purposes of file identification, of an exact copy of the student edit record, including the Unique Identifying Number (UIN), with every image of a student's response to a constructed-response item. The UIN is assigned to each image as the booklet is scanned. These edit files also remain in the main data files residing on the mainframe computer. Exact matching of scores assigned to constructed-response items and all other data for each individual student is guaranteed, since the UIN for each image is part of every image file.