By the Duke Talent Identification Program
It’s no secret that standardized testing is increasingly common. Each year new tests appear, and it may seem that more school time is spent on testing than on teaching and learning. Yet test scores themselves are often difficult to interpret.
In brief, achievement tests measure mastery of course content. The term standardized means that the test itself and the test conditions are the same for everyone. Standardized features include time limits, test directions, and question types. Although specific questions vary from one version of a test to another, the overall format stays the same. By providing the same conditions for all test takers, standardization increases the chances that score differences reflect actual differences in what students know or can do, rather than the influence of unrelated factors.
Two Categories of Achievement Tests
Depending on how students’ performance is evaluated, achievement tests may be called norm referenced or criterion referenced. In a norm-referenced test (NRT), individual performance is compared to the performance of a large group of test takers, known as the norm group. As might be expected, individual NRT results are most meaningful when the norm group is very similar to the individual tested.
Test scores can provide important feedback about your child’s abilities.
NRT performance can be difficult to understand, because statisticians use many terms to compare scores. Each term reflects a slightly different way of showing how an individual score compares to the scores made by members of the norm group. You may encounter terms like percentiles, deciles, stanines, orquartiles. Of these, the percentile score is the most meaningful and most readily understood. Percentiles tell how many students out of 100 made a score the same as or lower than your child’s. For example, the 96th percentile indicates a performance as good as or better than 96 out of 100 test takers in the norm group.
Criterion-referenced tests (CRT) are more straightforward. They evaluate mastery of specific knowledge or skills, the criteria against which individual performance is judged. A student either knows the material, labeled “passing,” “mastery,” “meets standard,” or the like, or does not know it. Moreover, a student’s performance on a CRT is independent of others’ performances. On a CRT everyone in a group may pass by demonstrating mastery of the material, whereas on an NRT higher and lower scores are ranked against each other, regardless of how well or poorly the group as a whole has performed.
Common Sources of Confusions
There are important differences between a percentile and a percentage. A percentile score (or rank) represents a norm-referenced comparison of one score with others, while a percentage simply expresses how many questions were answered correctly on a scale from 0 to 100. A score of 70 percent correct might be equivalent to a 40th percentile rank or a 90th percentile, depending on the test. Conversely, different percentage scores may equate to the same percentile rank. Well-designed NRTs commonly use a nationally representative norm group that numbers many thousands of individuals, together with sophisticated statistical techniques for evaluating each potential question before it is chosen for a test.
Grade-equivalent scores are also easily misunderstood. This type of score reports performance as a grade level rather than as a relative ranking or absolute level of performance. For example, it may seem that an 8th-grader who receives a grade-equivalent reading score of 11.1 reads as well as an 11th-grader, but that is incorrect. The test did not include 11th-grade reading, but it did contain readings at the 8th-grade level. The correct interpretation is that this student reads 8th-grade material as well as an 11th-grader would read it.
If you still are not sure you understand a score report, contact your child’s school. Most schools have one or more persons trained to interpret test scores, and most school districts also have testing specialists. Be sure to consult the test company’s Web site as well, since all of the larger companies offer free online parent guides to interpreting their tests.
When the Results Are In
Test scores can provide important feedback about your child’s abilities, which can help you identify his or her strengths or weaknesses. You can use this information to plan future learning, for instance, by enriching areas of strong interest and ability and bolstering areas that may be lagging. Test scores may also provide useful documentation in support of gifted program advocacy efforts.
Although there are many aspects of life that tests do not address, they are powerful tools when used properly. Take advantage of them!
—Michael Matthews, PhD
Michael Matthews is a research fellow with the Duke University Talent Identification Program. He received his PhD in educational psychology with a concentration in gifted and creative education from the University of Georgia.