not for grading:
We administer a reading test for a sample representative of the entire population in this case students. We can divide this group to 100 groups on the basis of scores they get in this test and called them as 1st percentile to 100th percentile. For example we can compute a score/point which separate 25 percent of excellent students from the remaining 75 percent. As well as we can compute a point / score which separate 50 percent of good from 50 percent of poor students in this study. This point is called median .We can compute any other percentile for example a point/ score which separate 20 percent of very good students from the other. This norm allows us to compare other students. In 2001 we administered this test for a group of students and we observed only one percent of them get a score which is equal or above of median score or 50th percentile but in 2002 we administered this test for a group of students and we observed 8 percent of them get a score which is equal or above of this score. Note that in the original sample 50 percent could get a score equal or above this score. But in our two samples in 2001 and 2002 only 1 and 8 percent respectively could get this score. So may be this test belongs to other population or other country or may be our groups don’t belong to this population and other norms should be developed for them.