A practical and accurate method of ranking scores

In the ranking of large numbers of scores, many persons continue to follow a more or less random procedure, which not only consumes an unwarranted amount of time, but also results in frequent errors. This practice, which will appear on a little inquiry to be not at all infrequent, is especially unfortunate in view of the fact that recently there has been prepared a set of tables which greatly reduces the work of calculation in determining the coefficient of correlation by the rank method.l The low cost and great convenience of these tables make it literally true that no worker in the psychology of education can afford to be without them. The purpose of this article is to increase their helpfulness by offering suggestions toward a simple, accurate, and expeditious method of ranking scores-a process that in every case must be completed before the tables are brought into use. There is no difficulty in rankiig by mere inspection a small number of scores or measurements; but when the number goes above about fifteen, it becomes desirable to follow a more accurate and expeditious method. Most persons who have had much experience with arranging scores into ranks have doubtless adopted some sort of rational procedure. Let us suppose that one must rank the following grades or scores: 39, 60, 53, 47, 38, 42, 53, 55, 62, 45, 53, 48, 50, 59, 35, 49, 37, 52, 48, 54, 43, 58, 55, 37, 55, 62, 37, 45, 55, 62, 49, 55, and 47. It is possible to look over these numbers and note that they range from somewhere in the 30's to the 60's, and then to search out rather quickly the numbers in the 60's and rank them; then the 50's, the 40's, and finally the 30's. But as the scores to be ranked increase in number, this method becomes more and more troublesome, and the possibilities of errors increase very rapidly. If an error is made, it may escape notice till the process is completed, and one then notices that the final rank number does not agree with the number of scores in the series. The correction of the error is by this time about as difficult as would be the re-ranking of the entire series

Econometrics · Mathematics education · Ranking (information retrieval · Statistics · Computer Science · Education and Islamic Studies · Mathematics · Psychology · Artificial Intelligence