Continued from Exhibit 4.4.

Principal Howell and Mr. Olsen met Superintendent Sanders in her office the next morning to explain the error and submit revised dropout data for Lincoln High School. The superintendent reviewed the report and applauded the decrease in dropout rates, which reflected what Mr. Howell had assured her was occurring at the school. Despite her satisfaction, she became quite serious as she turned back to them. "So how is it that this critical information about the change in enrollment codes wasn't communicated effectively?"

Mr. Howell stammered for a moment, but Mr. Olsen was ready. "We need 'metadata,' " he replied in a measured tone. "A metadata system would, among other things, include definitions, code lists, and any changes in data items from year to year on the web form the schools use to report their data. Our data providers and users would have one-click access to the up-to-date guidance they need to understand what is being requested, and in what format. I know it sounds like something magical, but it's not. It is a standard practice for data collection, management, and use that is designed specifically to help avoid these types of data quality problems."

Dr. Sanders nodded. "I remember reading about metadata, but thinking that it seemed like an investment in the theoretical. Clearly this episode illustrates that there are practical implications of understanding our data better." Mr. Olsen smiled. "Don't get me wrong, there is an investment in instituting a robust metadata system," he offered cautiously. "Oh, I am sure" the superintendent agreed. "Starting with a rigorous training program! Evidently we haven't mastered the quality assurance procedures we already have in place. But such an investment would be a small price to pay if it means we can feel confident about using data to make sound management, instructional, and planning decisions."

The end