• Introduction
• Key Findings and Recommendations

The National Center for Education Statistics (NCES) has been collecting counts of public school dropouts through its Common Core of Data (CCD) survey since the 1991-92 school year. However, not all states report dropout data in strict agreement with the CCD definition (table A), with the result that data from these nonconforming states have been withheld from publication. This situation has led NCES to explore the feasibility of adjusting nonstandard dropout reports to make them comparable with those from states using the standard CCD definition.

The desire for comparable dropout statistics has been accompanied by considerable interest in developing a standard high school completion statistic based upon data available from the CCD. Between 1997 and 1999, staff from NCES and state education agencies worked with analysts from Westat (a private research firm) to develop a methodology for adjusting nonconforming dropout data and to test a proposed high school completion rate.

The analyses presented in this report found that the major types of nonstandard dropout reporting practices have statistically significant, but different, effects on the size of state dropout rates. The most common variant practice uses a reporting calendar that effectively takes a "snapshot" count of dropouts at the conclusion of the school year rather than at the beginning of the next year (table B). This typically leads to a small net increase in the number of dropouts reported, when compared with the CCD reporting guidelines. The effects of how summer dropouts (those who complete one school year but fail to enroll for the next) are reported and whether students moving to adult education GED classes are considered dropouts (as required by the CCD) were stronger.

As a result, the report recommends that data from states using an alternative reporting calendar be published, without adjustment, with data from the states that conform to the CCD reporting calendar, and a footnote be used to identify states using an alternative reporting calendar. This would add 12 states to the number whose CCD dropout data are reported by NCES. Because the effects of the other two variations are stronger and more variable than the first, it is recommended that NCES continue to withhold publication of data from the states that follow these variations. There were 10 such states in 1995.

It appears from the analysis of dropout and completion data that the CCD can support a useful high school completion rate. This rate is the proportion of students who leave high school (grades 9 through 12) with a diploma or other credential to the total number of students who leave (as completers or dropouts). High school equivalency recipients are excluded from the completer group because these data are not reported at the school district level, as are dropout and other high school graduation and completion counts.

The report recommends that a method using multiple years of dropout data be used in preference to a synthetic, or reconstructive, rate based on a single year of information. The recommended method is less affected by single-year or one-time changes in dropout rates and more accurately estimates the proportion of public school students who leave high school successfully.

Table A.—Number of states reporting dropout data to the CCD: 1991 to 1995

Table B.—Types of nonconforming practices