home *** CD-ROM | disk | FTP | other *** search
- -------------------------------------------------------------------------
- Information on Grading Curves
- -------------------------------------------------------------------------
-
- Grades v2.70 now uses a grading curve feature in order to give
- the user a better idea of what their grade might be. However,
- implementing a grading curve is rather difficult and the difficulties
- will be explained in this document.
-
- For those unfamiliar with the concept, a grade curve is when your
- letter grade has no direct correlation to your numeric grade. For
- instance if you receive an 86 out of 100 which is usually designated
- a B, with a curve you may receive a B+ or A-.
-
- As nice as this concept is, a grading curve is very
- unpredictable and usually up to a teacher's discretion. Some
- teacher's will follow a set pattern, while other's will curve based
- upon the highest and lowest scores received. At times teachers might
- feel that a general curve is better than a high/low curve. Finally,
- some teacher's might curve for students who are in the low ranges,
- yet not curve for those in the higher ranges. As you can see, there
- are thousands of different curve possibilities. It would be
- irresponsible for a program to tell the user that it could foresee
- what sort of curve a teacher might use, and what the user's grade
- will be based upon any curve.
-
- Then there is the problem of calculating the curve once it has
- been determined which curve is going to be used. Some curves will be
- based upon the median score of the class, and then your score is
- compared to this median score. The most accepted method is to compare
- your score directly to the highest and the lowest score to determine
- where you fit in the class. Again, depending on the curve used your
- score might be compared to only those scores in the lower ranges, or
- maybe to those in the higher ranges.
-
- Because of the problem of determining the curve used, there is
- also a mathematical problem of how to implement the curve. Grades
- v2.70 will now predict your grade using a pre-defined curve
- percentage, and makes NO guarantee that this is the type of curve
- percentage that will be used. When you configure the program, you
- will be asked what curve percentage to use. Unless you find that the
- program is consistently wrong when it comes to the curve, do NOT
- change the default value!
-
- Here is how the program will now determine the different curve
- gradients: First, all Test Scores are totalled and averaged, and
- then a curve is performed on that TOTAL average grade. This may seem
- backwards since a curve is usually applied to each grade
- individually. (Since curves on individual tests differ) However, this
- method will give the program a better chance of accurately
- determining the curve since it is applied over all Test Scores. The
- program is thus assuming that whatever the curve fluctuations are,
- they will converge to an average curve. That is, if the first test
- has a curve of 1% and the next one a curve of 2% and the next a curve
- of 1.5%, the average curve is between 1% and 1.75%.
-
- You will be shown 16 different Curve Values, starting with the
- default value, and then incrementing in single units. If you have
- specified the default Curve Percentage correctly, this should then
- produce a range of an entire letter grade. That is, the grade with no
- curve is a C, and the Curve of 14.25% is a B. Do not expect the
- grades in the last two columns to be correct. They are there to give
- you a better idea of how a curve would affect your grade. Therefore,
- to determine what a likely grade would be, compare the grades from
- the early percentage values to see how many different percentages
- produced the same grade. For example if the following is the display:
-
-
- 0.00% B 3.80% B+ 7.60% A- 11.40% A
- 0.95% B 4.75% B+ 8.55% A- 12.35% A
- 1.90% B+ 5.70% B+ 9.50% A- 13.30% A
- 2.85% B+ 6.65% A- 10.45% A- 14.25% A+
-
-
- Your grade is most likely to be a B+ since this grade showed up the
- most in the early percentage values. While A- showed up many times,
- it was only in the higher percentage points and it is unlikely that
- this will be the grade. Certainly, if there was a small curve you
- would still be likely to receive a B+. If the curve was large, around
- 3-4%, the result would also be a B+.
-
- The above example was produced from ONE Test Score of 84%. If
- one adds more Test Scores, the curve begins to differ. However, 84%
- is a borderline grade, and you can see how a curve would affect it. A
- grade like 80% would not produce the same results.
-
- Finally, we would like to stress that we do NOT guarantee any
- grade produced by these curves, nor should you as a user assume that
- the grades produced by the program are valid! Instead, use the values
- obtained, and the knowledge you have of your own class status, to
- determine if you are doing well in a class. No program will ever help
- you receive a grade you do not deserve.
-
- If you, or a school administrator/teacher, has any suggestions
- for improving the Curve Prediction please let us know. The above
- method was derived after many discussions with Teachers, yet we are
- sure it can still be improved upon.
-
-
- Send Comments/Registrations to:
-
- SupremeSoft
- 105 Deerfield Drive
- Easton, CT 06612
- -------------------------------------------------------------------------
- This document is (c) Copyright 1990 by SupremeSoft and Manu Singh.
- All Rights Reserved.
- -------------------------------------------------------------------------
-
-