New Longitudinal Study Shows Core Knowledge Boosting Scores, Closing Achievement Gap
A new study indicates that a Core Knowledge education can improve students’ scores on standardized tests and also narrow the achievement gap between advantaged and disadvantaged students.
Researcher Fred Smith tracked the effect of Core Knowledge on the achievement of elementary students in a Virginia school using a quasi-experimental, longitudinal, matched-comparison design. Smith, then a Ph.D. candidate at the University of Virginia’s Curry School of Education, compared students in a Core Knowledge school with students in another school in the same district with a similar demographic make-up.
Smith examined test results on Virginia’s state tests, the Standards of Learning (SOL), and on the national Stanford 9TA test. He also tracked gain scores, the achievement of disadvantaged and advantaged students, and the achievement gaps between advantaged and disadvantaged students at the two schools. Smith found that Core Knowledge had the following positive effects:
- Core Knowledge increased student academic achievement as measured on the Stanford 9TA tests.
- Core Knowledge promoted fairness in schooling by providing equal educational opportunity to disadvantaged as well as advantaged students.
- Core Knowledge helped narrow the achievement gap on the Stanford 9TA test between advantaged and disadvantaged students
- Core Knowledge helped students achieve larger gains on the Stanford 9TA over two-year periods.
Smith tracked two cohorts of students at a Core Knowledge school and two similar cohorts at a comparison school. All of the students in the study remained in the same school from kindergarten to sixth grade.
Results for the Virginia state tests, the Standards of Learning, were mostly inconclusive, with the Core Knowledge cohorts sometimes posting higher mean scores and sometimes failing to do so. In most cases the differences did not qualify as statistically significant. However, results on the well-validated Stanford 9TA tests revealed a large number of statistically significant advantages in favor of the Core Knowledge students.
The cohort of students who began their studies in a Core Knowledge kindergarten in 1994 and remained in the same school continuously through grade 6 posted higher mean scaled scores than the control group in all three subject areas of the 6th grade Stanford 9TA tests: Reading (706.93 vs. 675.85), Math (713.70 vs. 662.25), and Language (659.87 vs. 635.80). The margins of superiority for Reading and Math were deemed statistically significant: Reading p ≤ .029 and Math p ≤ .002. (A p factor, or probability value, is a statistical indicator of the reliability of a finding. The smaller the number, the smaller the likelihood of the results being just a fluke. A p value of .05 means there is a 5% chance that the outcome was only a chance occurrence; a p value of .01 indicates a one percent chance. Scientists generally consider any p value smaller than .05 statistically significant.)
The cohort of students who began their studies in a Core Knowledge school in 1995 and remained in the same school through grade 6 also posted higher mean scaled scores in all subjects tested on 6th grade the Stanford 9TA: Reading (709.91 vs. 670.95), Math (718.67 vs. 680.61), Language (662.13 vs. 645.76). Again, the Reading and Math results were statistically significant (Reading p ≤ .002 and Math p ≤ .014).
Additional analysis showed that the superior performance of Core Knowledge students on the 6th grade Stanford 9TA tests held true for both advantaged students and disadvantaged (free and reduced lunch) students. Advantaged students in the Core Knowledge school outscored advantaged students in the control school in all three areas tested on the 6th grade Stanford 9TA, and for both cohorts examined. Likewise, disadvantaged students in Core Knowledge schools outscored disadvantaged students in the control school in all three areas tested on the 6th grade Stanford 9TA, and for both cohorts examined. The disadvantaged students showed statistically significant advantages in reading (p ≤ .017 for one cohort and p ≤ .030 for the other).
Smith also found statistically significant evidence that the Core Knowledge school was doing a better job than the control school at narrowing the achievement gap between advantaged and disadvantaged students. Smith computed the differences between advantaged and disadvantaged students’ Stanford 9TA scores for sixth grade in Reading, Math, and Language. In the Core Knowledge cohort that began kindergarten in 1994 and was tested several years later in grade 6, advantaged students outperformed disadvantaged students by 7 points in Reading and 3 points in Language. Remarkably enough, however, the disadvantaged students outperformed the advantaged students by 18 points in Math. In the aggregate, disadvantaged students outperformed advantaged students by 8 points. Although the number of students Smith was able to track longitudinally was small — just 31 students in the Core Knowledge school for this cohort — this is nevertheless a tantalizing indication that Core Knowledge can help narrow the gap between the haves and have-nots in society. The results stand out even more sharply when one notes that the combined achievement gap in the control school for the same cohort was a whopping 156 points.
Results for the 1995 cohort were less spectacular but still showed Core Knowledge narrowing the achievement gap more successfully than the control school. For this group of students, the control school showed a smaller achievement gap in 6th grade: only 40 points in the aggregate. But the Core Knowledge school did even better, almost completely erasing the achievement gap among the students it educated for six years. The 12 disadvantaged students in the Core Knowledge school actually outperformed the 25 advantaged students in Reading (707.64 advantaged vs. 714.66 disadvantaged) and Language (661.36 advantaged vs. 663.75 disadvantaged), though they were outperformed in Math (722.04 advantaged vs. 711.66 disadvantaged). Taken as an aggregate, the disadvantaged Core Knowledge students in this cohort lagged behind the advantaged students by only 1 point. Essentially, there was no achievement gap.
Smith also tracked gains from grade to grade. For example he tracked the gains students made from the fourth grade Stanford 9TA to the sixth grade Stanford 9TA. Students in the Core Knowledge school made larger gains on the Stanford 9TA in all 6 cases Smith examined. Many of the gains detected were statistically significant. For example, on the Stanford 9TA, gains in reading and math from grade 4 to grade 6 were deemed highly significant (p ≤ .001).
Smith separated students into advantaged and disadvantaged groups and analyzed the gain scores again. He found that advantaged Core Knowledge students made larger gains than advantaged students in the control group in all six of the Stanford 9TA cases he examined. The advantaged Core Knowledge students exhibited gain-score superiority that was deemed highly significant (p ≤ .001) in all three subjects (Reading, Language, and Math) and for both cohorts tested.
Among disadvantaged students, the Core Knowledge students displayed an equally impressive advantage, posting larger gains than their control group peers in all 6 of the Stanford 9TA cases examined (as well as 10 of 12 Standards of Learning cases). In some cases the differences in gain scores on the Stanford 9TA were staggering. For example, disadvantaged students in the 1994 cohort in the Core Knowledge school posted healthy gains in all three Stanford 9TA subject tested (Reading +61, Math +83, Language +40), while their peers in the control school posted sharp losses in all subjects (Reading -37, Math -33, Language -46). The edge to Core Knowledge was deemed highly significant in all three subjects (p ≤ .001 for Reading, Math, and Language). Such large results in a small sample might seem to suggest a fluke, but the gain-score advantages for a completely different group of students in the 1995 cohort were almost as statistically robust. Once again the Core Knowledge students posted healthy gains in every subject (Reading +97, Math +112, Language +76). Once again students in the control school slipped in all areas (Reading -4, Math -31, and Language -11). And once again the differences in gain scores were deemed highly significant for each subject (Reading p ≤ .001, Math p ≤ .001, Language, p ≤ .002).
Smith’s research provides compelling longitudinal evidence that Core Knowledge can improve academic performance for both advantaged and disadvantaged students, and can help to narrow the achievement gap between these two groups. His findings also suggest that Core Knowledge may have certain latent effects—effects that may not be visible immediately, and may not show up in a one-year study, but begin to appear after several years of exposure to the curriculum and can grow quite large when exposure persists throughout the elementary years.
Smith’s dissertation is available on microfilm and in digital form through Proquest/UMI, 1-800-521-0600.
 |
 |
 |
|
||
 |
|
 |
|
||
Last updated: Fri, May 23 2008
