Illustrative example: Four
candidates A, B, C, and D
appeared in the same shift. The
difference in marks between A
and B is exactly the same as the
difference in marks between C
and D.
Will the percentiles also be equally spaced?
No. The percentiles may not be equally spaced.
Why not?
Percentiles indicate the relative position of the
candidate in the shift. Experience suggests that
there are fewer candidates scoring high marks
than those scoring marks in the middle range.
This will be reflected in the percentile scores.
Can this be elaborated with reference to the
example?
Suppose the raw marks of A and B are
respectively 91 and 93 (closer to the top) and
those of C and D are 64 and 66 (closer to the
middle).
There are likely to be fewer candidates with
scores near the top i.e. between A and B and
many more with scores between C and D, even
though in each case the difference in the raw
marks is only 2.
Hence the percentile scores of A and B are likely
to be closer to each other while the percentile
scores of C and D are not so much.
Isn’t the same phenomenon observed for (usual)
ranks which are based on raw scores?
Indeed, it is true that the same phenomenon is
observed for usual ranks based on raw scores –
and for the same reason. Namely, there would be
fewer candidates at the top between candidates A
and B than between candidates C and D even
though the difference between the respective raw
scores is 2 marks each.
Hence the ranks of candidates A and B are closer
to each other than those of candidates C and D.
Does the above analysis indicate that ranks based