This study revises slightly the findings of my November 2001 report, High School Graduation Rates in the United States. In that study, I used an easily replicable method to estimate the percentage of public high school students receiving a high school diploma in the nation, each state and many of the nationâ€™s largest public school districts. The same method was also used to estimate these rates for major racial and ethnic groups in each state and each of the districts examined.
I recently discovered an error in the calculations that were used to estimate the overall national and state rates. In my methodology, I estimate the graduation rate by dividing the number of public high school diplomas awarded in 1998, which is available from the National Center for Education Statistics, by an estimate of the number of students who would have received diplomas that year if graduation rates were 100 percent. I arrive at this latter number by taking the number of students enrolled in public schools in 8th grade in 1993 (also available from the NCES) and adjusting it for the percentage change in the overall student population between 1993 and 1998. The error stemmed from the inadvertent use of the percentage change in the overall population rather than overall student population between those years.
Recalculating the national rate to correct for this error, I now find that estimated national public school graduation rate in 1998 was 71 percent, slightly lower than the 74 percent originally reported. Since the overall thrust of my report was that public schools graduation rates are much lower than is commonly reported, this recalculation does not change the original reportâ€™s conclusion.
Estimated graduation rates for each state were also recalculated. These changes may be found in Table 1, at the back of this report.
The mistaken calculation occurred only for the overall state, and hence the overall national, graduation rates. The calculation was done correctly for each of the local school districts, the state-level racial and ethnic results, and the district-level racial and ethnic results.
In general, the differences between the new and previously reported numbers are modest. In fact, the two sets of numbers are correlated at .94. If the two sets were identical the correlation would be 1.0. The changes tend to be small because in most states the total population and the total student population grew at similar rates. In those states where the student population grew at a rate very different from the total state population, however, the changes could be larger.
As long the report was being revised I took the opportunity to correct a previously reported data entry error for Jefferson County, Kentucky. I also re-examined the entire data set for any other data entry errors and added information from Arizona that arrived too late to be included in the original report. No data entry errors were found in the state results but a few errors were found in the district numbers. None of the corrections change reported graduation rates by more than one or two percentage points except for a larger error for Virginia Beach, Virginia, where the overall graduation rate was lowered by 11% because of a data entry error, and Saint Paul, where graduation rates were previously understated. All district-level results stated in this report reflect these corrections. All tables in the current appendix refelct these changes.
Of the hundreds of numbers entered and the scores of calculations made I am pleased to have found relatively modest errors, but am chagrined to have found any errors at all. I, the Manhattan Institute, and the Black Alliance for Educational Options will continue to strive to provide the highest quality research.