In January’s State of the Union address, President Obama announced that graduation rates were at an all-time high. Official data from the National Center for Education Statistics (NCES) show the national graduation rate hit 81% in 2012-13, up two percentage points in as many years. Considering how valuable a high school diploma is, the increases in graduation rate are very good news. But NPR is suspicious.
In multiple stories last week on All Things Considered and Morning Edition, NPR took a deep and well-rounded look at graduation rates and the (sometimes dubious) means schools are using to get students to graduate high school, or at least to be counted among the graduates. National stories were supplemented by reporting by numerous local affiliates.
One of the reports that had traction was a WBEZ report that Mayor Emanuel cited an all-time high 69% graduation rate for Chicago Public Schools (CPS). The report found flaws in CPS accounting on graduation rates, and a day later CPS responded, indicating that it was aware of the problems and is taking steps to correct them.
NPR’s reporting raises important questions about these all-time high graduation rates. Are they accurate? Are politicians and school officials playing politics with the numbers? Are graduation rates really at an all-time high?
At least for the national statistics, the answers are: No, they are not accurate. No, officials are not playing politics with the numbers. And yes, graduation rates are at an all-time high.
No, they are not accurate.
Graduation rates are complicated statistics, primarily because they are built from so many other numbers. NCES’s official graduation rates are based on EDFacts data. States submit EDFacts data to NCES at three levels: state, district and school. In theory, the school counts should sum to district counts, and district to state counts, resulting in the same graduation rate. In actuality, they do not result in one graduation rate; they result in three.
Published national graduation rates, 79% in 2010-11, 80% in 2011-12, and the all-time high 81% in 2012-13, are based on state-level EDFacts data, meaning each state submits a single statistic. The school-level EDFacts data yield graduation rates that are two percentage points higher, 81% in 2010-11 and 82% in 2011-12 (the 2012-13 district and school level data files are not yet released). The district-level data rates are fall somewhere in between. In a perfect system, the whole would equal the sum of its parts; for graduation rates, the parts sum to more than the whole.
EDFacts state graduation rates show even larger mismatches across levels than then national rates do. The table compares graduation rates based on school-level and state-level data, and the differences between them, for the US and the 5 states with the greatest discrepancies. California shows the biggest difference at 7 percentage points, between 86 and 79%, with Michigan close behind with a 6 point difference.
Why are there different graduation rates based on the level of data the states submit? Because collecting these data is an incredibly complex and error prone endeavor. NCES gathers EDFacts data from more than 50 state-level education agencies. The state agencies gather data from over 11,000 districts, which contain more than over 20,000 high schools, where the actual record keeping is done.
To calculate graduation rates, or specifically the Adjusted Cohort Graduation Rate (ACGR), discussed in depth by NPR, each school counts a cohort of students in grade 9, and for four years must add students to the cohort who come into the school, and remove students who legitimately transfer out, move or die. Four years later, the number of graduates is divided by this adjusted cohort count, to produce the ACGR.
While we know the root of the accuracy problem is grounded at the state level, the effort required to track four year counts in 20,000 schools without some error is incalculable. Some states have consistent data (20 in 2011-12) but others states, like those in the table above, have large disparities that result in inconsistent national rates. As a result, graduation rates that are perfectly accurate and precise are a myth.
No, officials are not playing politics with the numbers.
If officials were playing politics with graduation rates, they would need to first understand that the rates differed and second use them selectively for political gain. The notion that officials even understand that rates differ credits officials with an unlikely amount of cunning. Furthermore, high level officials do not have the capacity to direct and manipulate these statistics. (Legitimate concerns about how schools are raising graduation rates are distinct issues.) The second notion, that politicians use rates selectively, does not bear out. For instance, President Obama cited the published NCES rate in the State of the Union, despite the fact that the district and school level data would have produced higher rates. Mayor Emanuel’s office and CPS responded with similar logic, saying despite known problems, graduation rates are up. Generally, officials are citing the best data they have.
And yes, graduation rates are at an all-time high.
While graduation rates are not perfectly accurate and precise, we know they are going up because the data are collected and reported consistently over time. The independent and nonpartisan NCES excels at these processes to return statistics that are reliable, albeit not perfectly accurate, consistent and impartial. It is impossible to say that the US graduation rate is 81% for 2012-13, because the same family of data consistently yields varying rates based on the state, district, or school level data. However, while the rates are different depending on how the data are reported, the trends in both the national and state rates are the similar using the state-, district- and school-level data. The consistency over time and across data sources provides a solid foundation for knowing graduation rates are at an all-time high, even if we can’t say exactly how high.
from AEI » Latest Content http://ift.tt/1RcpemD
0 التعليقات:
Post a Comment