Almost had it
Your characterization of the Education Week methodology (see here) as analyzing "the percentage of 9th graders who completed high school four years later" isn't quite correct. The formula on page 12 of the Ed Week report does, indeed, make use of dropout rates from grade