Be it resolved that:
The Georgia State University Senate urges the Board of Regents of the University System of Georgia to allow the implementation of the letter-grade plus/minus grading system at University System of Georgia institutions.
If the Board of Regents allows the implementation of the letter-grade plus/minus grading system at University System of Georgia institutions, Georgia State University shall begin using the plus/minus grading system as soon as the necessary computer changes can be made.
|A = 4.00||B = 3.00||C = 2.00||D = 1.00|
|A- = 3.67||B- = 2.67||C- = 1.67||F = 0.00|
|B+ = 3.33||C+ = 2.33||D+ = 1.33|
There are several arguments in favor of adding pluses and minuses to GSU letter-grades.
(1) It is unfair to reward students equally when there are large disparities in their accomplishments. For example, under the current grading system, both 81 and 89 yield grades of B, while a 91 earns an A and a 79 a C. The plus/minus system allows this difference to be expressed on grade sheets and thus allows faculty to reward the superior performance of the student who earns an 89.
(2) The simple letter-grade system gives students a strong disincentive to work hard at the end of the term. A student has no reason to work hard at the end of a term if she or he has, say, an 85 average, and knows that unless s/he does extremely badly on the final examination, s/he will get a B for the course. Therefore, there is reason to think that students will work harder and learn more if pluses and minuses are added to the grading system.
(3) As noted below, adding pluses and minuses would make GSU's grades more reliable by reducing grouping error.
(4) Adding pluses and minuses may aid GSU's best students in their quest to enter the most competitive graduate programs. If a student with a 4.0 from GSU is being compared by an admissions officer with with a student with a 4.0 from an institution with pluses and minuses, the admissions officer would have reason to favor the other student over the GSU student. The other student's 4.0 shows that s/he got all As. The GSU student's 4.0 shows only that s/he got all As or A minuses.
At the turn of century, the vast majority of U.S. college and universities used a decimal, 0-100, grading system. From about 1900 until the 1960s, there was a trend to using grading systems that were less and less fine-grained. Thus, at the beginning of the 1960s, most colleges and universities were using the letter-grade system that is currently in use at GSU. During the 1960s, many schools tried to remove a perceived punitive aspect from grades. This led to many different grading systems including the pass/fail system. But this period of experimentation did not fundamentally change the basic use of the letter-grade system. From the 1970s and continuing through today, there has been a move, especially among universities and 4-year colleges, to adding pluses and minuses to the letter-grade system.
|Grading System||2-Yr Colleges
|Letter with +/-||9.5||18.4||33.8||52.8||24.6||41.2||25.1||40.0|
Other institutions have found that adding pluses and minuses to the letter-grade system has several effects.
(1) Fewer students graduate with a 4.0 GPA because the students who have a 4.0 GPA under the simple letter-grade divide themselves into two groups--A- students and A students.
(2) There are fewer students who graduate with a 2.0 GPA because whose who were just barely getting by and whose Cs were C minuses, fail to remain in good standing.
(3) There is no overall GPA effect on the students between 4.0 and 2.0. In other words, on average, there are as many B+ students as B- students. This is an especially important point here at GSU because it means that adding pluses and minuses can be expected to have no effect on the number of students who qualify for HOPE scholarships.
(4) There is some, tentative, evidence that overall average GPA rises more slowly over time at institutions using pluses and minuses. In other words, there is some evidence that pluses and minuses reduce grade inflation.
(5) Adding pluses and minuses causes an increase in the number of grade changes that Registrars have to perform. Interestingly, while some of this effect is, as one would expect, the result of more student appeals, it seems that much of it is simply due to more clerical errors made by professors on grade sheets. In other words, given more bubbles on grade sheet, professors fill in the wrong bubble more often.
Scholars of measurement theory have severely criticized the simple letter-grade grading system and the American Association of Collegiate Registrars and Admissions Officers has urged institutions to move to grading systems, such as the one proposed, which provide more possible grades than the simple letter-grade system.(5) The consensus among scholars of measurement theory is such that Admissions and Standards has been unable to find a scholarly publication which defends the simple letter-grade system, while many scholarly articles attack it.
The fundamental argument against the simple letter-grade system is that, because it allows only 5 possible grades, it allows for too much grouping error. To simplify a bit, grouping error occurs because the mean of the actual grades within an interval will be lower than the mean of the interval. Here is an example. Suppose that a professor is computing grades on a 0-100 system and that everyone who earns a score greater than or equal to 80 and less than 90 receives a B. In effect, all those who are given a B are given the same grade--an 85 (the midpoint between 80 and 90). But, if grades are distributed on a bell curve, there will be more grades below 85 than grades above 85. So if one averaged the numerical scores that the professor used to do the grading, this average would be less than 85. The grouping error is the difference between the grade reported to the Registrar (in this case, 85) and the actual average of those in that grade interval. The effect of grouping error is to inflate the performance of weaker students and reduce the reliability of grading. For any particular given reliability of the grading basis, one can compute the reliability for grading systems with different possible numbers of grades.
|Reliability of Grading Basis||Reliability of Grades for Various Numbers of Possible Grades|
As you can see, if the grading basis is .95 reliable, a move from a grading system which used 5 categories to one which used 10 would improve reliability from .85 to .92. Notice that the reliability of the grade given increases with the number of possible grades no matter how reliable the grading basis is.
In February and March 1997 the University Senate Office conducted a survey of full-time faculty at Georgia State at the behest of the Admissions and Standards Committee. The following choices and results emerged from that effort, in which 849 surveys were distributed to the faculty, and 459 were returned to the Senate Office or the Chair of Admissions and Standards:
Faculty favoring the current letter-grade system 19.6% (n=90)
Faculty favoring a system with the following grades: A,AB,B,BC,C,CD,D,F 12.4% (n=57)
Faculty favoring the letter-grades with pluses and minuses 66.0% (n=303)
Faculty favoring some other, unspecified, system 2% (n=9)*
*8 of these nine which were for some type of grading system finer than the letter-grade system, e.g., decimal on a scale of 100; +/- but with A+ = 4.3, etc.
The return rate of 54.1% was very good, and sentiment was overwhelmingly (80.4%) in favor of a finer grading scale, with only 19.6% of the faculty expressing support for the current system. Among the choices offered, 66.0% preferred the +/- system and 12.4% preferred the intermediate (A, AB, B, BC, ...) system.
Admissions and Standards conducted a survey of 223 GSU students. Members of Admissions and Standards asked students in their classes to fill out the survey so the response rate was 100% but the sampling technique is poor. Students were first asked to rank order four possible grading systems: pass/fail, letter-grades, letter-grades with pluses and minuses and the decimal system (0-100). Here are the results concerning the grading system that was ranked number one.
Pass/fail system 10.8% (n=24)
Letter-grade 48% (n=107)
Letter-grade with +/- 26.9% (n=60)
Decimal 14.3% (n=32)
To determine the strength of student preferences, students were than asked to allocate 100 points between the four grading systems. Here is the average point score for each system.
Letter-grade with +/- 26
(These numbers do not total to 100 because some students did not allocate 100 points.)
1. The information in this paragraph is drawn from American Association of Collegiate Registrars and Admissions Officers, Grades and Grading Practices: Results of the 1992 AACRAO Survey (Washington, D.C., 1994).
2. AACRAO Survey, p. 9 and 11.
3. The information in this paragraph is drawn from C. James Quann, Plus (+) Minus (-) Grading: A Case Study and National Implications (Washington, D.C., 1987), a white paper produced by the American Association of Collegiate Registrars and Admissions Officers.
4. Unless otherwise noted, the information in this paragraph is drawn from the AACRAO Survey and Quann, 1987.
5. AACRAO Survey, p. 25.
6. Robert Ebel, "Marks and Marking Systems," in Measuring Educational Achievement (Prentice-Hall, 1965), p. 423.