Home Research Files Reports
Spring 2000

Help Topics

Home

Score Explanations

Percentile Ranks
A percentile rank is the percent of people in the norming sample who have scores less than or equal to a student's score. A student with a reading comprehension score at the 60th percentile scored equal to or better than 60 percent of the students in the norming sample. Stanford 9 is a nationally normed test. This means that the norming sample is representative of a national cross-section of students. The norming sample for Stanford 9 included students from the northeastern, midwestern, southern, and western regions of the U.S. The sample was also representative of the nation in terms of ethnicity, urbanicity, socio-economic status, and students enrolled in private and Catholic schools. Since Stanford 9 is a nationally normed test, its norming sample was representative of the nation, but not necessarily of the state. The composition of this sample can be found in a Frequently Asked Questions (FAQ's) section of the California Department of Education web site (http://www.cde.ca.gov/statetests).

The SABE/2 is designed for students whose primary language is Spanish, and it was normed on a group of Spanish speaking students in bilingual classes in 12 states, including California,with substantial populations of Spanish speaking students,. Because the norming group was not a nationally representative sample, all the normed scores are called "reference" scores rather than "national" scores. Students are compared to the performance of students in the reference group that is similar to them in the same way students who take the Stanford 9 are compared to the performance of the representative national sample.

NPR (National Percentile Rank) of the "Average" Student
For the STAR program, Stanford 9 reports national percentile ranks (NPRs) for individual students, but does not produce group level (e.g., school or district) percentile scores. A mean percentile rank for a group cannot be calculated from percentile ranks for individuals because percentile rank scores cannot be added or averaged across students. The national percentile rank of a hypothetical "average" student can be approximated by using Normal Curve Equivalent (NCE) scores. NCE scores divide the normal curve into 99 equal units. The NPR of the "average" student can be found by find the average NCE score for a group of students, then finding the percentile score that is equivalent to the mean (average) NCE. This mean percentile can be roughly interpreted as the 'NPR (national percentile rank) of the "average" student.'

RPR of the "Average" Student
Because the SABE/2 uses a reference sample of Spanish speaking students rather than a national representative sample of all students, all the percentile scores are compared to the reference group. The Reference Percentile of the "Average" student is calculated for groups of Spanish speaking students just as the National Percentile of the "Average" student is calculated for nationally normed tests such as the Stanford 9. Normal curve equivalent scores were used in the same way for the SABE/2 as for the Stanford 9 to estimate a percentile rank for the "average" student.

% Scoring Above the 75th NPR, % Scoring At or Above the 50th NPR, and % Scoring Above the 25th NPR
These scores correspond to the percent of students in the school, district, county, or state scoring in the top quarter nationally; the percent of students scoring in the top half nationally; and the percent of students scoring in the top three-quarters nationally. The three levels used to create these group scores are the 75th, 50th, and 25th national percentiles, respectively. The percent of students scoring above each level is calculated by counting the number of students scoring at or above a particular level (e.g., 50th percentile) or above a particular level (e.g., 25th percentile and the 75th percentile), dividing by the total number of scores, and converting to a percentage. The percent scoring at or above the 50th percentile is the percent of students in this school, district, county, or state whose scores would place them in the top half of the students tested nationally. The percent scoring above the 75th percentile and above the 25th percentile are the percents of students in this group whose scores would place them in the top quarter and top three-quarters, respectively, of the students tested nationally.

% Scoring Above the 75th RP, % Scoring At or Above the 50th RP, and % Scoring Above the 25th RP
These scores are similar to the scores above, except that the SABE/2 comparison group is a "reference" group of Spanish speaking students in bilingual classes.

Mean Scaled Score
Raw scores identify the number of items answered correctly on a test or sub-test. Raw scores are limited in their measurement precision because of differences among test items. For example, some items are more difficult than others. A scaled score takes item differences into account and is calculated to provide a more precise measure of the knowledge or skills tested. Through this calculation, an increase of one point at one place on the scale is described as being equal to a one- point increase anywhere else on the scale. Scaled scores are particularly useful for reporting changes over time. The Stanford Achievement Test Series provides results in terms of scaled scores for individual students and a mean or average scaled score for groups of students. The SABE/2 also provides scaled scores for individual students and a mean or average scaled score for groups of students. Scaled scores can be compared within the same test, but not between two different tests (e.g. Stanford 9 and SABE/2), nor even two different subjects on the same test, (e.g. Reading and Mathematics). For example, the scale on the Stanford 9 starts at approximately 200 and goes to 900, while the SABE/2 starts at 1 and goes to 999. A scaled score of 500 will have very different meanings on the Stanford 9 5th grade reading test, the Stanford 9 5th grade math test, and the SABE/2 5th grade reading test.