Percentile Ranks

Percentile ranks are a commonly used form of norm-referenced score transformation. They are reported for most ability, aptitude, and achievement tests as well as for attitude, personality, and interest inventories.

Percentile ranks are expressed as whole-number values that range between 1 and 99 to indicate the position or rank of the respondent's score in a reference group of test takers. For example, if 100 individuals completed a test, a percentile rank of 40 would indicate that the respondent's score was equal to or higher than 40 of the 100 individuals in the norm group. In the case of nationally standardized tests with larger norm groups, a percentile rank of 40 would indicate that the respondent had earned a score equal to or higher than 40% of the people in the norm group.

Percentile ranks have one very serious disadvantage—they are not equal units of measurement. Because they are not equal units, score differences in the middle of the distribution appear larger than score differences in the tails of the distribution. For this reason, one must be very careful when interpreting percentile ranks.

In general, approximately 68% of people earn percentile ranks between 16 and 84. Percentile ranks in this range are considered average or typical. In a normal distribution, these percentile ranks include the area between one standard deviation above and one standard deviation below the mean of the distribution. Percentile ranks in the tails of the distribution represent large differences in scores. For instance, the difference between percentile ranks of 95 and 99 or between 10 and 15 represent much larger differences than do the differences between, for example, 40 and 45 or 70 and 75.

Test profiles attempt to show visually that the differences in the middle of the distribution are closer together by compacting the distribution in the middle range and expanding the range in the tails of the distribution. Particular attention should be given to the graphic on the profile for achievement tests when interpreting percentile ranks to students and their parents.

Understanding this characteristic of percentile ranks is very important in comparing scores on a standardized achievement test. Students who earn percentile ranks of 40 in reading and 60 in mathematics [Page 380]are scoring in the average or typical range for both subject areas. The student and his or her parents should not be led to believe that the student's achievement in mathematics is higher or better than the achievement in reading. Conversely, students who earn percentile ranks of 5 in reading and 15 in mathematics have performed much better in mathematics than in reading, although their overall achievement is lower than most students of their age or grade in school. The same differences apply for the upper end of the distribution. Students who earn percentile ranks of 97 in mathematics and 90 in reading have higher levels of achievement in mathematics than they do in reading.