Page 5

Comparison Groups

“Comparisons” in both the initial and corrected versions of the Duke report are of little value and provide no meaningful context for making determinations about scholarship students relative to their public school peers from similar backgrounds. Immediately following analysis of scholarship student aggregate data, the corrected Duke report includes a “public school comparison” page, showing that public school students in North Carolina are performing above the national public school average on a different test, the National Assessment of Educational Progress (NAEP). This test is administered to a sample of students, both nationally and at the state level. It’s worth noting that the original uncorrected Duke report, which was reported on by the media, compared free and reduced lunch eligible (FRL) public school students in North Carolina only to other FRL public school students nationally—and then summarized this way, “Although it is not an ‘apples-to-apples’ comparison, the most recent data shows that comparable students who remained in public schools are scoring better than the voucher students on national tests.”8 This comparison was never equitable for a variety of reasons. The two groups of NC students (voucher and FRL public school), and the students they were measured against, represented different groups of students. Scholarship students, who are themselves low-income (eligibility for the scholarship is based on having a family income that does not exceed 133% of the amount to quality for free and reduced lunch), were evaluated against a nationally representative sample of students across the country—from both public and private schools, and not just other low-income students. As mentioned above, North Carolina FRL public school students, on the other hand, were not compared in the Duke report to a nationally representative sample of students taking NAEP. They were compared to other FRL public school students nationwide who took NAEP. Here is how the Iowa Testing Programs of the University of Iowa, test developer for the Iowa Assessments (used by a number of North Carolina private schools reporting aggregate data on scholarship students), characterizes their sample for determining NPR:

One metric, National Percentile Rank (NPR), indicates the status or relative rank of a student’s achievement compared with that of a nationally representative sample of students … The most recent NPRs for the Iowa Assessments are based on the standardization studies conducted in 2010-2011 in which nationally representative samples of public and private school students were assessed in all content areas. As a result, information from the Iowa Assessments allows educators and parents to compare an individual student’s or a local group of students’ performance to the most current estimate of national student performance available.9

North Carolina Opportunity Scholarships: Countering a Flawed Duke Report and Setting the Record Straight

4

OSP White Paper  

NORTH CAROLINA OPPORTUNITY SCHOLARSHIPS: Countering a Flawed Duke Report and Setting the Record Straight

OSP White Paper  

NORTH CAROLINA OPPORTUNITY SCHOLARSHIPS: Countering a Flawed Duke Report and Setting the Record Straight