Context for and purpose of the resolution
There have been multiple Faculty Senate resolutions about whether and where to publicize median grades – whether to include them on student transcripts, as is the current practice, and/or to publish them on the Office of the University Registrar (OUR) website, which was once the practice, but was discontinued. (The resolutions are discussed in detail below.) The Faculty Senate originally sought, and the first resolution called for, publication in both realms, with the expectation that these postings would encourage students to take courses with low median grades, which the Faculty Senate thought students might not otherwise do. However, the Faculty Senate learned, based upon a published research study of the OUR posting practice, that students used the OUR postings to shun instead of select courses with low median grades. Obviously, this is the exact opposite of what was intended.
Additionally, as explained later, according to the research study, the OUR website postings not only impacted students in terms of course selection, but it also impacted their relative standing compared to median grades. The study also indicated, and the Faculty Senate believed, that the OUR website postings contributed to grade inflation.
Ultimately, the Faculty Senate voted to discontinue the OUR website postings, but it did not address the transcript postings, which remained.
After the OUR median grade postings stopped, as a direct response to that action, and continuing to the present day, students have resorted to informal means, to self-help, to replicate those postings by crowdsourcing their own median-grade spreadsheets on Reddit (and, likely, elsewhere). From both student reports and common sense, we can expect that these spreadsheets are producing similar results to the OUR website postings (i.e., promoting the very grade-driven course selection that the various median grades resolutions were designed to prevent). Indeed, that is the raison d’être of the spreadsheets.
It is hoped and expected that if median grades are no longer included on transcripts, students will stop creating such spreadsheets. The data with which to do so would be limited, if not eliminated. Hopefully, too, students’ motivation to create public listings of median grades would be diminished as they would not have to worry about having their performance relative to other students recorded on transcripts.
In addition to the above concerns related to strategic course selection, the sponsors believe there are other compelling reasons not to include median grades on transcripts. Promoting median grades orients students comparatively towards each other instead of towards their own learning. This is not productive and amplifies competition. Additionally, students report that having median grades on transcripts can be demoralizing and stressful, even for those students who did well, because they feel that their good or excellent grades are devalued by high median scores. Shouldn’t letter grades speak for themselves? Presumably, those grades accurately reflect a professor’s evaluation of student learning. Indeed, the Cornell University Grading System, adopted by the Faculty Senate, defines grades according to achievement-based criteria, focusing on mastery of the subject. Lastly, the median grades do not provide comparative data with other institutions because other institutions do not include median grades on transcripts. Cornell’s anomalous approach may, thus, disadvantage our students.
Previous Faculty Senate resolutions about median grades
The first resolution on median grades was passed in 1997. That resolution[1] called for median grades to be included on transcripts and posted on the OUR website. The resolution’s goals were to provide students with “a more accurate idea of their performance, and they will be assured that users of transcripts will also have this knowledge.” This, in turn, “may encourage students to take courses in which the median grade is relatively low.” Additionally, “[o]utside users of the transcript will have more information on which to base their assessment of a student’s performance . . . .” Apparently, it was also hoped that the posting of median grades on the OUR website would reign in grade inflation by publicly “shaming” faculty who gave inflated grades. (See the minutes of a May 2009 Faculty Senate meeting at which the Faculty Senate discussed a resolution to remove median grades from the OUR website.)
Due to technological obstacles, median grades did not appear on transcripts until 2008, although they were posted on the OUR website commencing 1998.
In 2006, the Faculty Senate Educational Policy Committee (EPC) introduced a resolution to remove median grades from the OUR website until “such time as median grades can be posted on student transcripts . . . .” The EPC believed that the dual postings – on the OUR website and student transcripts – were inextricably linked. The resolution failed.
In 2009, one year after the posting of median grades on transcripts had begun for students matriculating in 2008, the Faculty Senate EPC again introduced a resolution to remove median grades from the OUR website on the basis that the posting of median grades was “used by students to select courses that give high grades” and was “contributing to grade inflation.” The resolution was unsuccessful in part because members of the Faculty Senate wanted more time to assess the success of its original plan, which had only recently been effectuated, to post median grades on both transcripts and the OUR website.
In May 2011, the EPC again introduced a resolution to end the practice of posting median grades on the OUR website. That resolution passed. The resolution did not address the practice of including median grades on transcripts, and the process was kept intact. Again, the Faculty Senate’s goal was to discourage strategic course selection and combat grade inflation, now recognizing that it was the accessibility and public nature of the data, not its inaccessibility and private nature, as initially thought in 1997, that was problematic. Specifically, according to the 2011 resolution:
Public publishing of the median grades for all courses on the Cornell website is used by students to select courses that give high grades . . . and this practice is contributing to grade inflation at Cornell, [and] the practice of posting median grades on the web is being exploited by external websites to match median grades to specific professors allowing students to choose those courses or sections with higher median grades . . . .
The research study of the effects of the OUR website median grades postings
The Faculty Senate’s 2011 resolution was heavily influenced by a published study conducted by two Cornell professors, Talia Bar and Vrinda Kadiyali, and colleagues from outside Cornell showing that posting median grades on the OUR website resulted in students taking courses with high median grades, not low ones; the availability of this “grade information online induced students to select leniently graded courses – or in other words, to opt out of courses they would have selected absent considerations of grades” (Bar, T., Vrinda K., & Asaf Z. 2009, p. 107).[2] The authors found that the impact of the OUR median grade postings was most significant for “advanced annual courses” (Bar, T., Vrinda K., & Asaf Z. 2009, p. 101).
To study the effect of the median grade information on course selection (and grade inflation), the authors
utilize[d] a large dataset of grades assigned in undergraduate level courses in the College of Arts and Sciences between the spring semester of 1990 and the fall semester of 2004. We restrict[ed] our analysis to courses that were taught at least once pre–policy change (1990–1997) and at least once post–policy change (1998– 2004) (Bar, T., Vrinda K., & Asaf Z. 2009, p. 95).[3]
The authors found that “[c]omparing the pre- and post-policy change periods . . . , the share of courses with a median in the A range increased by about 16 percent, while the share of students enrolled in such courses increased by more than 42 percent (Bar, T., Vrinda K., & Asaf Z. 2009, p. 98).
The authors tested and found support for their assumption that students had access to the online median grade information and used it to select courses. In 2006, the authors
conducted a survey of around 500 students from Cornell’s Department of Economics. Close to 60 percent of the students indicated that they heard about the website and visited it, and roughly 80 percent indicated that they sometimes use the information provided in the website to select courses or have heard about other students doing so (Bar, T., Vrinda K., & Asaf Z. 2009, p. 98).
Additionally, the authors looked at the time periods when students visited the OUR website median grade postings:
We obtained two and a half years’ worth of daily data on the number of visits (hits) to the website and examined whether it was visited more frequently during periods of course registration, when students are more likely to need grade information for course selection. There are two periods of course registration in every semester at Cornell: in the first (“add/ drop” period) students enroll in courses offered in the current semester; in the second (“pre-enrollment” period) students pre-enroll in courses offered in the next semester. We found large and highly statistically significant differences in the number of hits between the periods: the daily number of visits was roughly twice as large in registration periods than in nonregistration periods. The rise in hits during registration periods indicates that students are aware of these posted median grades and are presumably using them to guide their enrollment choices (Bar, T., Vrinda K., & Asaf Z. 2009, p. 98).
The authors also found a skewed approach to strategic course selection, which in turn affects students’ relative standing vis-à-vis median grades:
The introduction of information on grades increases the utility of some students but decreases the utility of others. For example, a student who chooses the strictly graded course . . . is made worse off . . . because the [publishing of median grades] drives . . . students [with lower SAT scores] out of the strictly graded course and makes this student look worse compared to the remaining students. Conversely, a student who chooses the leniently graded course . . . is made better off, because the leniently graded class will attract more . . . students [with lower SAT scores], and thus the student’s grade relative to the median will look better (Bar, T., Vrinda K., & Asaf Z. 2009, p. 107).[4]
Lastly, regarding the impact of the median grade postings on grade inflation, the authors found that grade inflation, which “existed before the policy change . . . accelerated after the new policy was implemented” (Bar, T., Vrinda K., & Asaf Z. 2009, p. 105); “from 1998 until 2004, the period in which the policy of online reporting of median grades was in effect, the mean grade has steadily increased” (Bar, T., Vrinda K., & Asaf Z. 2009, p. 94).[5]
Cornell is anomalous in its practice of recording median grades on transcripts
According to the Cornell University Registrar, of the 71 member institutions of the Association of American Universities, Cornell is one of only four institutions that record median grades on transcripts. The other institutions are McGill University; University of California San Diego; and Dartmouth, which is the only other Ivy League institution to engage in this practice. Significantly, though, according to the Cornell University Registrar, Dartmouth is leaning towards eliminating its practice; they expect to take up this issue during the 2023-2024 academic year.
Students replicate the OUR postings by creating their own median grades spreadsheets
According to a recent review of Cornell Reddit entries, Cornell students are reconstructing the OUR website median grade listings by crowdsourcing spreadsheets on Reddit that contain median grades listed by course. Students began this self-help practice in direct response to the university discontinuing the OUR postings while retaining the transcript postings. The inaugural Reddit entry about these spreadsheets explained:
Remember when Cornell published Median Grade Reports? Well they stopped after Fall 2010, but started putting median grades on our transcripts. Our plan is to create our own Median Grade Reports for Spring 2011 and Fall 2012 based on our transcripts, but we need your help. If everybody follows these steps, we can have full official Median Grade Reports by next week, maybe in time for Add/Drop.
The posting included instructions for ordering transcripts and making entries in the spreadsheet.
A review of Reddit in March 2023 revealed spreadsheets containing median grades from 2018 through January 2022. On January 8, 2023, a spreadsheet was uploaded to Reddit with instructions for crowdsourcing median grade data for the fall 2022 semester.[6] There are also Reddit entries pertaining to specific courses. As of March 2023, the last such entry was on January 15, 2023.
Students’ perspectives about Cornell’s current approach of posting median grades on transcripts
In 2018, the Student Assembly (S.A.) sponsored a resolution calling on the “Faculty Senate to investigate the possibility of reinstating the semesterly reporting of median grades.” The S.A. resolution was premised in large part on the idea that because median grades are recorded on transcripts, they must additionally be published on the OUR website, “establishing a concept known as ‘Truth in Grading.’”[7] To this point, the resolution noted that “median grades are recorded on student transcripts for graduate schools and employers to view, but students are not allowed to view this data until after they have completed the course.” The Faculty Senate did not take up the S.A. Resolution.
The Cornell Daily Sun has published a few articles about median grades on transcripts. In 2011, after the university discontinued the OUR website postings, the editorial board wrote a column protesting the continued practice of including median grades on transcripts. The column made the point that including median grades on transcripts will not result in students taking courses with low medians because students prioritize their GPA:
A student’s raw G.P.A. is still the dominant — and sometimes the only — method that employers and graduate schools use to evaluate students’ academic performance . . . . Students understand this. Few students would consistently choose to take the most difficult courses at the expense of their G.P.A.s, even if those classes’ median grades were listed on their transcripts.
The column also made the point that including median grades on transcripts can harm students’ standing vis-à-vis graduate/professional schools and employers:
Students who consistently challenge themselves with more difficult courses, but underperform the median, should not be classified as below average academically based solely on a raw median grade. Conversely, students interested in majors and classes that offer consistently higher median grades should not be penalized by a transcript that sends a message of indolence and diminishes their academic achievements. Whether low or high, median grades on transcripts open the door to interpretations that can negatively affect Cornell’s students in the job market.
In 2020, in a similar column, a student author reiterated many of the same points from the 2011 piece: that median grades “disadvantage students during graduate school admissions and job or internship searches,” and median grades lead to misinterpretations of students’ achievements:
Students who enjoy classes or choose majors with higher averages are labeled as ‘taking the easy way out,’ devaluing their accomplishments. On the other hand, the transcripts of those who take risks but fall short of the median present the distorted image that these students will underperform outside of college as well.
Lastly, the article noted that “many classes have multiple sections with different professors or teaching assistants. Under such circumstances, there cannot exist a standardized level of challenge, even within a single course . . . . ”
In a recent article from February 2023, students were interviewed about the inclusion of median grades on transcript. One student stated:
[It’s] stressful to have [median grades] on transcripts . . . . It’s not a common practice, and it is pretty odd and weird. If you do really well in a class and get an A because you worked really hard, then that A shows up on your transcript. . . . But then you see that the median grade is also an A and instead of having that good feeling of having gotten an A, you are stressed and start going down that rabbit hole.[8]
Another student “disagreed with including median grades on transcripts . . . . Cornell places too great an emphasis on individual grades and [median grades on transcripts] take away from the learning experience.”
One “pre-med” student said they like having median grades on transcripts.
The article concluded that “[a]lthough the policy has been in effect for over a decade, students still feel wary about the inclusion of median grades [on transcripts].”
In January 2023, an associate vice provost (AVP) from the Office of the Vice Provost for Undergraduate Education met with undergraduate students to learn their perspectives about median grades on transcripts. The AVP met with students who serve as facilitators for the Intergroup Dialogue Project. There were 22 students from colleges/schools across Cornell. There was one first-year student; the rest were sophomores, juniors, or seniors. Of the 22 students, 17 of them knew, before the AVP informed them, that median grades are recorded on transcripts. Several of them learned this fact through reading student entries on Cornell Reddit (either spreadsheets or student complaints). Others learned by googling something like “median grades and Cornell” or by word of mouth from friends or teammates. Some learned this information as early as their first year. One student paid $3.00 to a commercial website (she could not recall which one) to obtain median grades for Cornell courses. Students reported that they ordered transcripts during their time at Cornell, so they learned their median grades before they graduated.
While a few students thought it a good idea to have median grades on transcripts, most did not, for multiple reasons. Several students reported that they took courses with multiple sections for which the instructors gave markedly different ranges of grades/median grades, yet the disparate median grades were aggregated for the purpose of establishing for the entire course a single median that was included on transcripts. Students from the sections with lower medians felt disadvantaged. One student said that they were in a gateway course that had a high median grade, but many students cheated, and this cheating produced the high median. The student, who had not cheated, performed below the median. This student and other students who did not cheat were upset by the high median because they believed it was a product of dishonesty. Students said that median grades on transcripts “exacerbated comparisons to other students”; made the students “feel pitted against each other”; and were a “disincentive for learning.” One student said that students compared median grades across Cornell colleges/schools to compare which “programs were more rigorous.” The student thought this effect “toxic.” Several students said that they felt “devalued” or “bad” when a median grade was higher than their grade in a course.
Uniformly, the students did not think that including median grades on transcripts encouraged them to take courses with low medians. They said the opposite was true, and about a half of the students volunteered that they had relied on median grade information to select courses with high medians. The students did not want to take courses with low median grades because they did not want to “sacrifice” their GPA; “the GPA is the bottom line.”
Recommendation and Rationale
Undoubtedly, publicizing median grades results in students avoiding courses with low median grades. Such strategic course selection is antithetical to the Faculty Senate’s intent behind its various median grade resolutions and Cornell’s fundamental principle of learning for the sake of learning, and it can have an adverse effect on students’ educational experiences. While the Faculty Senate sought to address the problem of strategic course selection by discontinuing the practice of posting median grades on the OUR website, the posting on transcripts results in the same behavior. Moreover, the practice of posting median grades on transcripts also promotes competition with other students over learning and can be discouraging for students. And according to the research study discussed above, publicizing median grades may even promote grade inflation and affect students’ standing relative to course medians.
As the S.A. resolution noted, if median grades are posted on transcripts, students want advance notice or warning of what those median grades are likely to be. As the Reddit entries demonstrate, if Cornell does not provide this notice, students will recreate it themselves (although surely less accurately), for example, through crowdsourced spreadsheets. It is uncertain whether students would want to retain this entrepreneurial practice if median grades were not listed on transcripts, but without the data readily available from transcripts, they could not so easily reconstruct median grades and produce the spreadsheets. While there is no doubt that such information is and has always been shared by word of mouth to the extent it is available, it is doubtful that grapevine reporting would be sufficient to build comprehensive listings such as the Reddit spreadsheets. Without such public listings, strategic course selection is more difficult. According to the research study of the OUR website postings, word-of-mouth sharing does not have the impact of public postings.
Equally important, even without any public postings such as the Reddit spreadsheets, including median grades on transcripts is in and of itself is detrimental. It reinforces competition, discourages students from taking academic risks, and can be demoralizing. It is not surprising that including median grades on transcripts has these effects. Median grades on transcripts convey to students (and readers of transcripts) that a student’s standing vis-à-vis other students is an important measure of a student’s achievement, perhaps more important than a course’s letter grade itself. Performance as compared to one’s peers is, thus, prioritized over mastery of learning. By making students’ relative standing so significant, the practice encourages students to orient themselves towards this comparison as a rational calculation of their worth and a rational approach for how to conduct their education (e.g., avoiding courses with low medians). According to the 2020 Mental Health Review Final Report, Cornell students’ excessive focus on comparisons/competition with other students is damaging and, as the above excerpts reveal, students are disheartened by it. It is understood that even without median grades on transcripts, many students will still seek to preserve their GPAs, investigate instructors’ grading practices, and take courses with instructors who award higher grades. But median grades on transcripts add a competitive layer and, surely, further fuel strategic course selection.
There is no reason why letter grades cannot speak for themselves, as they are meant to represent a student’s level of achievement (see the Cornell University Grading System, a criteria-referenced grading system). Presumably, the A-earning student has performed at A-level competency, and it should not be necessary to report on a transcript whether that student was one among many or one of a few. The same is true for the C-earning student.
Other comparable institutions do not include median grades on transcripts, so median grade comparisons are not possible, making it hard to see how readers of transcripts would find Cornell’s median grade information useful. If there was any hope that publishing median grades would show graduate/professional schools and employers that median grades and, hence, grade inflation at Cornell are lower than at other institutions (assuming that is the case), again, due to the lack of comparative data, this goal cannot be achieved. It may be the case that because Cornell is singular in its inclusion of medians on transcripts, our students are disadvantaged.
Posting median grades on transcripts could also produce inequities. If median grades are available on commercial websites, as noted in the 2011 Faculty Senate resolution, and recently confirmed by one of the students with whom the AVP met, it is quite possible that students with means to pay for fees charged by commercial websites have better access to median grade data than students of less means do (the Reddit spreadsheets are not always complete).[9] While one student reported paying only $3.00 on a commercial website for a median grade report, it is unknown if such a modest price is still available or the norm.
Other inequities result from public postings of median grades such as through the Reddit spreadsheets. According to the research study, publicizing median grades on the OUR website had a disproportionate impact on course selection by and, hence, the educational experience of students with SAT scores that were lower relative to other Cornell students. The published median grades influenced such students to avoid courses with low medians more than it did students with higher SAT scores who expected to earn high grades regardless of whether the instructor graded leniently or not.[10] These selection processes, in turn, skewed medians.
With Cornell’s current test-optional and score-free undergraduate admissions testing policies, the analysis based on SAT scores could not be replicated now. Regardless, there are factors other than SAT scores that affect a student’s confidence with academic competition at Cornell. Extrapolating from the study, assuming that academic confidence in general is what affects course selection, one can expect that based on their level of academic confidence, students will be disparately impacted by the Reddit postings of median grades – students with less confidence may avoid courses with identified low median grades while students with more confidence may be less likely to avoid courses known to be graded more strictly. Median grades across all courses would thus be skewed by these self-selection processes.
Finally, the S.A. may have a valid point that if median grades are included on transcripts and, hence, available to graduate/professional schools and employers, then students should be forewarned of what course median grades have previously been. Even if that is true, it would be considerable backtracking and extremely detrimental to resume the OUR website median grade postings. From the perspective of an all or nothing approach, only nothing is a viable option.
Accordingly, for the reasons stated above, pursuant to the companion resolution, the university should cease posting median grades on transcripts.
[1] The resolution is too old to be available on the Faculty Senate archive webpages, but it was attached to a 2011 resolution, which is hyperlinked here.
[2] Bar, T., Vrinda K., & Asaf Z. (2009). Grade information and grade inflation: The Cornell experiment. Journal of Economic Perspectives, 23 (3): 93-108. https://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.23.3.93
As previously mentioned, due to technological obstacles, median grades were not recorded on transcripts until 2008. Accordingly, the study looked only at the effects of the OUR website postings.
[3] The authors did one sample that included all courses and another sample that was restricted to advanced (400-level) courses, which are, typically, electives and do not include first-year students who are less knowledgeable about the OUR median grades website and have less latitude in selecting courses.
[4] This would not apply or apply to the same extent to courses for which enrollment is required. The authors’ analysis suggests that in courses that are graded on curves, the curved grades would be similarly skewed.
[5] The authors discussed various potential causal factors in addition to the posting of median grades.
[6] For example, a spreadsheet on Reddit for spring 2018 contained median grades for 114 courses; a spreadsheet for spring 2019 included 352 courses and had instructions for ordering transcripts to find out course median grades; a spreadsheet for fall 2020 had 108 courses and instructions for ordering transcripts; and a composite spreadsheet for fall 2021, spring 2022, and fall 2022 had median grades for 937 courses.
[7] The S.A. resolution states that both Dartmouth and University of Michigan engage in such “Truth in Grading.” According to a senior associate registrar who has been at the University of Michigan for well over twenty years, the University of Michigan does not include and has never included median grades on transcripts.
[8] This concern that including median grades on transcripts is a source of unproductive stress came up during student focus groups conducted by the Academic Policies and Practices Working Group and the Grading Policies and Practices Working Group. These groups were convened by Vice Provost for Undergraduate Education Lisa Nishii in 2021 to consider recommendations from the 2020 Mental Health Review Final Report.
[9] The Reddit postings, which resulted from Cornell’s policy of posting median grades on transcripts, make this data readily available to commercial websites.
[10] Obviously, this analysis does not apply when course enrollment is required. The authors’ analysis suggests that in courses with optional enrollment that are graded on curves, the curved grades could be similarly skewed.
