1997-98 ABCs Report Card, Volume I - Appendix B
Main ABCs Vol. 1 Page | Foreword | Executive Summary | Program Notes | Schools Not Included | Glossary
This appendix describes how certain statistical and technical computations were implemented. Also, tables listing the parameters and constants used in the K-8 and high school models are provided.
Grade 3 Growth Parameters. The growth parameters at grade three were adjusted to reflect more appropriate data that was available after the 1996-97 implementation of the ABCs. The 1996-97 implementation of the ABCs relied on data for unmatched groups of third graders to set the growth expectations. The 1996-97 implementation of the ABCs was the first year for which matched grade 3 pretest and grade 3 (end-of-grade) posttest data were available. Matched student data provide a more appropriate estimate of the third grade growth parameters than unmatched data; therefore, the parameters for the third grade growth model were recalculated using the matched data from 1996-97. However, for 1997-98 only, the State Board of Education decided to use both sets of third grade growth parameters and apply the set that gives the best outcome for each school. Therefore for this report, the growth was computed using both the "unmatched" or "old," and "matched" or "new" parameters, and whichever computation yielded the greater growth for each school was used. See Table 1 for a list of the parameters and other constants used in the ABCs K-8 Growth Model. For 1997-98-99 and subsequent years, only the "new" parameters will be used.
Grade 7 Writing Standard Deviation. Writing scores are included in the growth composite using an index that compares performance over a three year period. This year, 1997-98, was the first year in which three years of grade 7 writing scores were available. Therefore, this was the first year that grade 7 writing scores could be included in the growth composite. This was also the first year in which the standard deviation of change could be calculated for grade 7 writing. See Table 1 for the grade 7 writing standard deviation of change used in the ABCs K-8 Growth Model.
Algebra I Scores in the Performance Composite. The high school performance composite is the percent of students at or above Level III in Algebra I; Biology; Economic, Legal, and Political Systems; English I; English II; and U.S. History. To make the performance composite representative of the students in each high school, the Algebra I scores for current ninth graders who took Algebra I prior to their ninth grade year are included in the performance composite. Algebra I scores of students at the middle school (grades 6, 7, or 8) during the current school year are not included in the K-8 performance composite or in the high school performance composite in 1997-98. The Algebra I scores of students at the middle school will be computed as part of the growth and performance composites in 1997-98-99 and subsequent years.
Performance Composite Confidence Band. To be identified as a low-performing school, the school must a) fail to make the expected growth/gain composite and b) have a performance composite less than 50% (i.e., a majority of students performing below Level III). The performance composite is the percent of students at or above Level III.
The confidence band, also known as a confidence interval, is a way of taking into account the statistical fluctuations that occur from year to year in small schools or schools with highly variable populations. The confidence interval itself will be narrow or wide depending on the size of the school and the variation in the scores that make up the performance composite. In general, the confidence band is narrower when the number of students is larger, or the scores are more homogeneous; conversely, the confidence band is wider when the number of students is smaller, or the scores are less homogeneous.
What all this means for the performance composite is that it is possible that a potentially low-performing school has a performance composite that is considerably below 50% but that school is not considered low-performing because the confidence interval for that school is wide (i.e., there is less confidence in the value of the performance composite). This situation would likely be true for a school that has few students or has wide variation in test scores. Conversely, it is possible that a school has a performance composite that is fairly close to 50% and is considered low-performing because the confidence interval for that school is very narrow (i.e., there is high confidence in the performance composite). This situation would likely be true for a school that has a large number of students or students all have about the same test score.
As long as the value, 50, lies within or on the boundary of the confidence interval for an observed performance composite, then the performance composite is not significantly less than 50 and hence the school is not classified as low-performing.
Excessive Exemptions. Schools that were identified as having excessive exemptions were notified and asked to justify their exemption rate. These justifications were reviewed by the state agency staff and schools that were determined to be in violation are noted in this report. Exemption of students must be consistent with federal and state guidelines for students with disabilities and students appropriately identified as Limited English Proficient. Exemptions were determined at the school on a case by case basis and they must be documented in Individualized Education Programs (IEPs).
Table 2. Constants Used in the ABCs High School Gain Model
Top of Page | Main ABCs Vol. 1 Page | Foreword | Executive Summary | Program Notes | Schools Not Included | Glossary
|North Carolina Department of
301 N. Wilmington St.
Raleigh, NC 27601