Md. Afsar Ali, Lecturer, Department of B.Ed.

Our educational evaluation generally follows a number of sequential steps, such as, writing the test items i.e. questions, moderation of the questions, administration of the question paper or taking the examination and evaluation of the answer scripts. From writing the test items to the evaluation of answer scripts, every step is influenced by personal liking or disliking or personal opinion of the subject entrusted with the duty. This factor of personal influence on the final outcome is known as subjectivity. So, subjectivity dominates our educational evaluation resuld. In broader sense, it is true that where there is subjectivity, there is less accuracy or in other ward, less reliability in the final resuld. So, now, we can say that our educational evaluation result are not completely reliable or cent percent accurate. Let us clear this situation by taking one example, Mr. X, an examinee, secured 60% marks in an examination, where Mr. Y, another examinee, secured 59% marks. In the eye of our conventional system of evaluation, Mr. X, will be awarded first class/division and Mr. Y second class/division. But their difference in achievement is only 1% and on the basis of this one percent difference in achievement we are making a difference of heaven and hell in awarding class/division to the concerned candidates! this type of conclusion, which is based on very small difference in numerical value of measurement between any two subjects, can only be taken when one can ensure objectivity/accuracy in every step of the whole measuring system. But as mentioned above, all the steps in our conventional educational measuring system are very much influenced by subjectivity. Hence, the results obtained from this system of measurement are not too reliable/accurate to award different class division to the candidates just on the basis of one percent achievement different. There is standard error of educational measurement due to the above subjectivity factor which ranges from 3 to 14% in different subjects. Let us apply this standard error on the numerical achievement scores of Mr. X and Mr. Y above. Mr. X, who scored 60% marks, his true score may be anything between (60+-3)% or 57% to 63% [If we take the minimum value of standard error (3%)in educational measurement]. True score value of Mr. Y would range from (59 + -3) or 56% to 62%. So, in any case, the true score of Mr. X might have been 5% and that of Mr. Y 62%. This resuld now, demands awarding in the reverse way to the concerned candidates, i.e., Mr.Y should be awarded first class/division and Mr. X second class/division.

This contradicting state of judgment can be avoided to a large extent if both the above candidates are awarded a single position in the scale of evaluation, named as “Grade”, say grade B.

In most of our public examinations, the current practice is to measure candidates performance by assigning numerical marks to an answer script. The examiners award numerical scorers while assessing the answer scripts. The scores generally ranges from 0 to 100; this is called hundred and one point scale. So, the total range of achievement of a candidate in a paper of three hours carrying 100 marks is finely divided into 101 score points. In case of grading, the most commonly suggested and used scales are 5 point and 7 point scales of achievement. That is, in this case, the total range of achievement is divided only into 5 or 7 score points. Hence there is a vast difference between any two adjacent scale values. So, this later scale is more convenient and ensures less error from the part of an examiner in choosing one scale value in it while assigning letter grade A, B, C etc. At the same time, it is quite difficult task to choose confidently only scale value from the hundred and one such values in the conventional numerical marking system (in 101 points scale). Because, here, there is very minute difference between any two adjacent scale values.

In short it can be said that in the marking system, the achievement scale demands high level of objectivity or accuracy from the part of an examiner in choosing only one point from 101score points in the scale, whereas, the other steps of the same examination system are greatly influenced by subjectivity! This examiner, who is an human being, is bound to be subjective at least to some extent. So, the numerical scores obtained through our examination system can’t be cent percent accurate.

From this point of view also one can that the grading system is superior to the conventional marking system.

A monograph on grading in universities (1976) gives the following of grading over numerical marks:

  • Examiners often indulge in evaluation rather than measurement, distort results by giving an unusually great number of pass marks and non at all of 1 mark less or 1 mark more.
  • Examiners in different subjects use different ranges of the 101 point scale.
  • Marks in different subjects are added up which can’t be, because different subject’s marks bear different units of measurement.
  • Grading system which is mainly judgment based on measurement is more uniform.
  • In grading system, comparability amongst various disciplines and within various subjects in a particular discipline is possible.
  • If all universities adopt the same  point scale (say), then students can move from one university to another much more easily.
  • Direct grading takes the emphasis away from marks.

Before concluding this topic I would like to quote the comments of the Radhakrishnan Commission (1948- '49) and U.G.C. regarding examination reform. Radhakrishnan commission opined that if a single reform is to be introduced in the field of higher education it is system of examination. The U.G.C. said: "we are fully convinced that if we are to suggest any single reform in university education, it should be that of examinations".