ABOUT THE GRADING PROCESS (SPRING, 2005)

 

Course letter grades will be determined from the top down by the overall Course Score (CS), calculated from the Normalized Test Score (NTS), the Normalized Lab Score (NLS), and the Normalized Homework Score (NHS), as follows:

 
(CS) = 0.60 (NTS) + 0.30 (NLS) + 0.10 (NHS)

Here the normalized test score, NTS, is the normalized value (See Normalization below) of the adjusted test score, ATS, which in turn is equal to the sum of the following scores for best four of the following five test hours: the (normalized) final exam score, weighted double, and the (normalized) scores of the three hourly tests, as described under EXAM POLICY. In other words, the lowest (normalized) scored test-hour is dropped for every student, and the resulting sum, (labeled here ATS) is renormalized into NTS before being included into the Course Score, CS, with the weight, 0.60, specified above.

 

The Laboratory Score, NLS, is computed from the adjusted raw lab score, ALS, obtained from the raw sum of the semester’s lab report grades, RLS, on the basis of “80% of the Maximum” process described below.
The Homework Score, NHS, is similarly obtained from the adjusted HW score, AHW, obtained from the raw sum, RHS, of the semester’s HW scores by “80% of the Maximum” process described in below. Occasional in-class quizzes related to the homework material may also be given from time to time. Their grades will be added into the raw HW score, RHS, and treated in the same way as the HW grades.


Course Letter Grades

 

Students whose Course Scores lie in the top 25% will receive an A. Students whose Course Scores lie in the top 50% will receive at least a B.  The A/B break-point will be set where a gap occurs in the course scores which is large enough to distinguish the performance of the lowest-scoring A student from that of the highest-scoring B student. Therefore, in practice, more than 25% of the students will likely get A’s. Likewise the precise B/C break-point will be set by such a gap, so that in practice more than 50% of the students will receive A's and B'

To estimate letter grade equivalents from normalized scores, note that about 50% of the population falls below the average normalized score of 70. That average is therefore near the B/C letter grade breakpoint. Furthermore, a normalized score equal to 90=(Avg + S.D)=(70 +20) will typically place a student in the top 1/6= 16.7% of the group, quite comfortably within the top 25% who are promised A grades. In practise, no letter grades are computed (apart from the Early Warning

grades after the first exam) until the end of the course, and then they are defined by the corse score defined above. (This paragraph was added Fri 4/8/05 for clarification.)

Course Requirements

 
Students who do not complete the course requirements will receive an F. Failure to complete all of the Labs and submit all the lab reports, missing the Final Exam, and/or missing two or more hourly exams each constitutes a failure to complete the course requirements. Generally students who do complete the course requirements earn a course score sufficient for a D.  Regarding the C-D breakpoint, we shall apply a prejudice in favor of C by giving D's only to students whose course scores are separated by a gap from the smooth distribution of the rest of the class.  Thus despite our prejudice for C over D, a substantial gap between your score and the low side of the continuous part of the class distribution may be dangerous to your C.

“80% of the Maximum” is Enough

 
The “80% of the Maximum” process for determining Lab and HW components of the Total Course Score is based on the proposition that Lab and Homework are learning experiences, and not exams, and that if they meet a certain pre-set standard, they should carry no grade penalty. We consider the achievement of “80% of the Maximum” possible total score to be “good enough”. In addition, we believe that “80% of the Maximum” is within the reach of every student who is willing to do the required work.

Therefore every student who achieves 80% of the Maximum possible Homework (or Lab) score will receive the same highest (=100) Adjusted Raw HW, AHW, or Adjusted Raw Lab,AL, score. Students who achieve less than “80%of the Maximum” will receive a raw score equal to the percentage of 80% which they achieve. These raw scores will then be normalized into NHS and NLS distributions with an Average of 70 and a standard Deviation of ±20 (just as the adjusted test scores, ATS, are normalized), to yield the Normalized Lab and Normalized HW scores, NLS and NHS, used to compute the Course Score, CS, with the above 60-30-10 weighting given in the above formula.
  Be Sure to Achieve the “80% of Maximum” Level
Beware: We advise everyone to make sure that he/she achieves the highest possible Adjusted HW and Adjusted Lab score, not just because it guarantees them the highest normalized HW and Lab scores, but because the failure to do so may seriously damage their NHS and NLS component scores. The reason is that the normalization of a distribution in which most of the grades lie at some maximum value can carry the few lower-than-maximum scores to quite low values, as discussed further below. The effect is drastic, but it can be avoided with due care, and it is the flip side of the decision to treat everyone equally who meets a certain specified threshold.
Normalization

Before any two grade components are added, they shall always be Normalized so that their distribution has an average of 70 and a standard deviation of 20. Thus if a certain (e.g. your own Exam I, or your adjusted lab score, ALS, in the formula above) grade has a raw (i.e., unnormalized) value, R, and comes from a class-wide distribution which has an average, A, and a Standard Deviation, D, the corresponding normalized grade is:


N = 70 + (R-A)*(20/D).

 

This normalization process provides a fair mechanism for dropping the "lowest" of several exam scores, even when one exam may have been much more difficult (i.e., had a lower class average) than the other exams: the normalized scores' distributions for all tests have, by construction, the same average (70) and the same standard deviation (20). Note that the normalization formula can never alter the relative ranking of any student with respect the others in the class: a higher value of R always yields a higher value of N.
We repeat the warning issued already above: if in the original distribution, nearly everyone has the highest possible score, as we expect to be the case for the raw HW and raw Lab scores because of the "80% is good enough" rule, then the few people who fail to meet that threshold my see their normalized score diminished significantly by the normalization calculation. Indeed, the normalized score can even become negative, although when it does so, we shall intervene and replace the negative score by a zero. This is the flip side of the promise that if you meet the minimal 80% standard, you will earn the maximum credit for HW and Lab: if you do not satisfy this easily achievable threshold, you may wipe out much or all of your credit for the HW and/or Lab segments of the course.