Grading on a curve
From Wikipedia, the free encyclopedia
In education, grading on a curve (also known as curved grading or simply curving) is a statistical method of assigning grades designed to yield a pre-determined distribution of grades among the students in a class. The "curve" in question is supposed to be the "bell curve", the graphical representation of the probability density of the normal distribution (also called the Gaussian distribution), but this method of grading does not actually make use of any specific frequency distribution such as the bell-shaped normal distribution.
This grading method can be described as proceeding in three steps. First, numeric scores (or possibly scores on a sufficiently fine-grained ordinal scale) are assigned to the students. The actual values are unimportant as long as the ordering of the scores corresponds to the ordering of how good the students are. In the second step these scores are converted to percentiles (or some other system of quantiles). Finally, the percentile values are transformed to grades according to a division of the percentile scale into intervals, where the interval width of each grade indicates the desired relative frequency for that grade.
For example, if there are three grades, P, Q and R, where P is reserved for the top 10% of students, Q for the next 20%, and R for the remaining 70%, then scores in the percentile interval from 0% to 70% get grade R, scores from 71% to 90% get grade Q, and scores from 91% to 100% get grade P.
The grading method can thus be tuned to determine the frequency distribution of the grades in advance, and if the intervals are already fixed at the beginning of a course, then so is the number of students who will receive each grade.
[edit] Benefits and shortcomings
Curved grading is essentially normative; grades are awarded based on performance relative to all other individuals rather than performance relative to the difficulty of a specific course. There must always be at least one student who has a lower score than all others, even if that score is quite high when evaluated against specific performance criteria or standards. Conversely, if all students perform poorly relative to a larger population, even the highest graded students may be failing to meet standards. Thus, curved grading makes it difficult to compare groups of students to one another.
On the other hand, the curve ameliorates the problem of deciding grades that fall very near a grade margin. Clustering of marks establish where the margin should be placed.
Another downfall of grading on a curve is the influence of study groups. When individuals study in a group, it can carry the peak of the curve with the average of the group. Thus, while the idea of a curve being used to determine individual performance is sought, it's interrupted by community performance.
Some professors compensate for this by assigning the student two grades, one based on a curve and another based on the traditional 90%=A, 80%=B, 70%=C and will award the student the higher of the two grades. In some purely elective courses, the professor may simply opt to grade each student individually on the traditional scale and only curve grades up if necessary. Courses that feature independent projects by students will often be graded solely on the merits of each project against a pre-set rubric, thus encouraging students to have fun with their work rather than worrying about competition. It is not uncommon to see grades in these courses feature 50-60% A's and few if any C's, D's and F's.
[edit] See also
[edit] External Links
- A Brief Note about Grade Statistics or How the Curve is Computed http://www.astronomy.ohio-state.edu/~pogge/Ast162/Quizzes/curve.html
- How to create a bell curve in Excel http://support.microsoft.com/kb/213930
