Public and Private School Students Perform Similarly
in NAEP Study
A national assessment comparing the performance of
public schools to private schools has found that nearly
always, when certain student characteristics are taken
into account, public school students perform just as
well, if not better, than private school students. These
conclusions, presented by the National Center for Education
Statistics, of the U.S. Department of Education, in
Comparing Private Schools and Public Schools Using
Hierarchical Linear Modeling, are based on data
from the 2003 NAEP reading and mathematics exams. This
good news demonstrates that public schools continue
to be a valid source for a solid education.
The Results
The study
examined the differences between public and private
school NAEP scores when adjustments were made for selected
characteristics of students. These characteristics included
gender, race/ethnicity, disability status, and English
Language Learner status. While all the unadjusted NAEP
scores initially indicated higher scores for private
schools, taking the selected student characteristics
into account substantially reduced the difference in
all analyses between 11 and 15 score points. The report
also further divided the category for private schools
into Catholic, Lutheran, and Conservative Christian
subcategories.
Grade 4
In reading, when all private schools were compared
to all public schools, the average private school mean
score was 14.7 points higher than that for public schools.
However, when adjusted for student characteristics,
the difference was statistically insignificant and close
to zero.
In mathematics, the average private school mean score
was 7.8 points higher than the average public school
mean score; again, after adjusting for student characteristics,
the average public school mean score was a statistically
significant 4.5 points higher than that for private
schools.
Grade 8
The average mathematics mean score for eighth graders
in private schools was 12.3 points higher than the average
public school mean score. After adjustment, the difference
was nearly zero and statistically insignificant.
In the case of average eighth grade reading scores,
private schools outperformed public schools, by 18.1
points before adjustment and 7.3 points after adjustment.
The Method
The Hierarchical Linear Modeling technique used in
this study accounted both for multiple covariates simultaneously
(e.g., race/ethnicity and disability status) and for
the clustering of students within schools. Student variables
included gender, race/ethnicity, students with disabilities,
English language learners, computer in the home, eligibility
for free-reduced-price lunch, participation in Title
I, number of books in the home, and number of absences.
Selected school-level variables included teacher experience
and certification, student mobility and absenteeism,
school location, and percentage of students in Title
I, eligible for free/reduced-price lunch, with a disability,
or learning English. This analysis allows for comparisons
between subgroups of public and private school students
who share multiple characteristics, like gender and
parents’ education. The sample included 6,900
public and 530 private schools at grade 4, and 5,500
public and 550 private schools at grade 8.
The report warns, however, that “an overall comparison
of the two types of schools is of modest utility.”
The study addresses national estimates and is based
on observational study, rather than a randomized experiment.
Furthermore, it is impossible to determine how self-selection
in private schools affects the data.
Prepared by Katherine Lu, July 19, 2006
|
