Grade point average (GPA) is a historical mistake in two senses. First, it has had an impact on student assessment the world over from elementary school through to university, and in this sense it is historic. Second, it has a very long history, appearing two centuries before the birth of modern-day theories and technologies of quantitative educational assessment; in this sense, it is also historical.
Today, however, we know so much more about educational assessment than the academics of the 18th century, and that there is no reason for continued acceptance of the GPA.
Let's imagine what might have happened in the past. A professor had a pile of students' term papers to assess. He studied them one by one and labelled them as 'Excellent', 'Good', 'Fair', 'Borderline' or 'Poor' according to his expectations based on his academic experience.
From 'Excellent' to 'Poor' there was a decrease in quality, and it was more convenient to label them as grades A, B, C, D and F. These were not convenient either, and were coded as 1st, 2nd, 3rd, 4th and 5th classes to indicate the order of quality. As writing 'st', 'nd', 'rd', and 'th' was clumsy to a busy professor, they were now written as 1, 2, 3, 4 and 5.
Here we see several things taking place in the professor's mind in a short time: qualities were coded in labels; labels were replaced by grades; grades were translated into ordinals; and ordinals were, for convenience, written in shorthand and appeared as cardinal numbers.
In this process of quality>labels>grades>ordinals>cardinals transformation, the first four stages are fine and right; re-coding does not change the meanings or the nature of assessment. But the last stage of equating 'ordinals' (numeric used for ranking and grading) with 'cardinals' (numeric used for enumerating or counting) changes the meaning and nature of measurement.
This is where GPA went wrong. Because ordinals 5(th), 4(th), 3(rd), 2(nd) and 1(st) denoting ranks based on 'subjective qualitative judgement' look exactly like cardinals 5, 4, 3, 2 and 1 based on 'objective quantitative enumeration', they were mistaken as equivalents. In short, ranking on quality became counting of quantity.
Even more fundamental is the issue of when a GPA is derived by summing and dividing a few grades, we make a basic assumption that grades are mutually exchangeable and hence compensatory to one another. This may be argued for if the term papers measure the same kind of knowledge and skills, but strengths and weaknesses in different disciplines may not be compensatory. In a sense, the GPA system is analogous to a multiple-regression system with all grades un-weighted (or rather unity-weighted) which assumes and hence allows compensation across disciplines.
When GPAs are used in the international arena, more confusion is added. Different nations have developed their own grading systems or, more likely, borrowed and then modified others to suit their own needs or beliefs.
This may have been fine in the past when the grades were not used beyond national boundaries. Nowadays, there is a tremendous amount of cross-nation student traffic for different purposes, for inter-institution collaboration, self-selected enrolment and distance education. The capacity for problems, confusion and anxiety to arise are many.
The three cases below illustrate the agonies of students seeking cross-nation admission to study programmes:
Case one: "I did my undergrad at Queen's University in electrical engineering. My cumulative average is in the high 70s, and I ranked about 20 out of 100. My transcript only shows percentage grades, not letter ones. I'm thinking of applying to both US and Canadian law schools. According to LSDAS's guideline, my percentage grade would convert only to a 2-point-something GPA, which is horribly low by US standards. Queen's applied science faculty absolutely refuses to convert my percentage grades to letter grades on my transcript."
Case two: "Please help! I received a conditional offer from UCL via email in December and then got my official letter last week. It looks great, except that the condition is that I graduate from my American law school with a GPA of 3.3/4.0. This is mathematically impossible for me to do at this point! Does anybody know how strict UCL is with their conditions? Mine has currently been sent back to be 're-considered' in light of the GPA problem. My GPA is 3.13 right now, probably will be 3.16 upon graduation."
Case three: "I would like to know if there is any formula to convert DGPA score (scale of 10) to GPA (scale of 4). I have done my undergraduation (sic) from IT-BHU, Varanasi, and we receive DGPA score on a scale of 10. However, most of the US universities ask to submit GPA score on a scale of 4."
Generally, 80 or more goes to Grade A (and its chromatic variants). But beyond this, there is much variation. For instance, a score of 70 is represented by a B in US and Canada, but by a C in Russia, India, China, Japan and Korea. If the student reports only the letter grade C when seeking admission to universities in the US or Canada, they may not be good enough as they will be seen as scoring below B.
Also, a student who scores one grade each of A to D in a US or Canadian university will have a total score of 283.5, if the mid-points of the grades are taken for the respective ranges. The average of this is 70.9 and hence a grade of B. This US-Canadian grade will translate into Russia C, India B+, China C, Japan B and Korea C.
Thus, whether the student is advantaged or disadvantaged when seeking admission across-nations depends on the nation to which he or she intends to go.
Additional difficulties can occur when attempting to align letter grades to GPA scores across national borders. The problem occurs because different countries used different GPA ranges. While the highest GPA is 4.00 in US, it is 4.5 in Korea, 10 in Vietnam and 13 in Denmark, indicating that the same letter grades not only have different values in different nations but there are also different scale lengths as well. And, more interestingly, Germany has the scale reversed, with lower GPAs representing better performance.
Some admission offices provide conversion charts to facilitate re-coding score percentages to GPA and many websites provide calculators of GPA as well. In spite of this, the problem just refuses to go away, as illustrated by the case below:
"I did my engineering at RV College of Engineering, Bangalore, India. It is one of the top ranked colleges in city and country. I was consistently in the top five of my class when it came to academics. But percentage wise I got 74% overall for the four years. Now according to the conversion chart, my GPA is 2.3. With all due respect I am forced to say that the GPA 2.3 is surely not fair in assessing my academic performance. The topper of the college had a percentage of 79.5. So even his GPA, according to the charts, comes to 2.3. So it didn't seem right to me and felt I needed to clear this doubt with you."
In the past, students who sought admission across-nations were usually the academically good ones in their original institutions. The conversion problem therefore might not have bothered admission offices too much. Even if many of these difficulties are known, the situation serves to cause confusion and frustration. Official figures about these matters are rarely in the public domain.
Solving a knotty problem calls for a paradigm shift. The current GPA system is an offshoot of a clever innovation which pre-dated modern knowledge and technology of educational assessment by two centuries. Its popularity has grown due to its assumed efficacy (or rather its administrative convenience) and institutional inertia has built up through years of common practice reinforced by the time, energy and emotion invested in it. The assumptions underlying the paradigm have seldom, if ever, been questioned.
The basic problem of GPA lies with mistaking ordinal numbers as cardinal numbers and applying arithmetic operations to the grades. Because of this, the basic principles of measuring at the ordinal level are consequently violated. This problem is then aggravated by its expansion and the modifications that have been made in many nations and over a long period of practice. The end result is that GPA takes on a range of meanings that make it meaningless and hence difficult and confusing to use.
The same information about grades can be used differently in the changed paradigm of a multiple cut-off model. In the multiple cut-off model, grades are taken to be discrete ordinals, not to be added or divided; the non-equivalence of units of grades and the lack of a ratio relation among grade differences are given due recognition. In short, grades are to be used in their original format with their original meanings. A student's achievement will be looked at collectively as a 'grade profile' (GP) comprising all the relevant grades obtained for a normal programme or a specific purpose (for example, scholarship).
The advantage of so doing is to maintain the integrity of grades as grades and to avoid confusion arising from conversion.
There may be occasions when a student's performance needs be summarised into a single number indicator. For this, a grade index (GI) can be obtained simply. Basically, this is looking for a grade among her/his grades that is a representation of their central tendency.
Basic statistics says that the arithmetic average (or the mean which yields the GPA) may not be a good indicator for such a purpose. The alternatives are the mode (the most common value) and the median (the value that occurs at the mid-point of the number of cases being examined).
Between them, the mode seems to be a better indicator in that, normally, a student usually gets about the same grades for different papers. Thus, a student who gets three Bs out of five or six grades can be referred to as a 'B student'. In addition, an indication of the range of her/his grades can be useful to show consistency. For example, if her/his five or six grades vary from C+ to A−, s/he can then be described as a 'B (A−/C+)', compared with another student whose grades vary from B- to A+ as 'B (A+/B-)'. Such a grade index provides information not only of global performance level but also fluctuation and is more consistent with the statistical practice of indicating the central tendency together with variability.
Concurrent use of GP and GI makes the information much richer than the single number GPA for guidance and feedback purposes. It also provides the teacher as well as the student with better understanding of where s/he should go and how s/he can get there.
It is assumed to be more convenient to use a single number GPA for selection and placement purposes. However, such convenience is obtained at the cost of non-equivalence. We do not realise that the multiple cut-off approach is equally if not more effective.
For instance, different departments require students of different relevant capabilities. By using GP and GI based on the multiple cut-off approach, the numbers of preferred grades (say, at least two A's and one B) can be maintained with the added advantage of having the relevant disciplines specified (say, at least two A's in the sciences and one B in languages).
The usefulness of a global indicator such as the GPA as a predictor is doubtful; more specific information is needed about a student's past achievement and future potential in relevant disciplines.
It is not easy to accept that what we have been so used to can be wrong. This is just one aspect of our human nature. A more positive aspect is the courage to change when change is desired. The GPA system is a case in point. If we realise that it is less than adequate, we need to have the courage and will to change it.
The information need may be the same today as it was two centuries ago, the practice has been passed down and 'refined' over the past two centuries, the system may have served us well up till now. But there is always a day when change is necessary to do a better job when we know we can do it better - better in that we do not have to continue making dubious assumptions and applying improper operations so as to minimise confusion and frustration.
* Dr Kay Cheng Soh is former head of the Centre for Applied Research in Education at the National Institute of Education, part of the Nanyang Technological University in Singapore.
* This is an edited version of a much longer article, "Grade Point Average: What's wrong and what's the alternative?" published in the current edition of the Journal of Higher Education Policy and Management. It is reproduced with permission.
The issue is not just the fact that Grade Point Average (GPA) adds numbers which are not meant to be added. There are also huge differences in grading policies from one country to the other. A French student finishing high school with a GPA or a weighted GPA of 15/20 (75%) is very good. In some more competitive schools the same very good student may well have a GPA of 12/20 (60%). In some subjects like French lit, or philosophy, getting 14/20 is already what Americans would call outstanding. Only in the final exam (the national baccalaureate) will grades be somewhat higher (but even there, the weighted GPA of 16/20 is being in the top 10%, and 18/20 well into the top 1%).
University of Versailles Saint-Quentin-en-Yvelines,
Receive UWN's free weekly e-newsletters