In recent years, Johns Hopkins has begun attracting and admitting an undergraduate student body of higher academic caliber and from a more diverse range of geographic, economic, and cultural background than ever before. For the class entering fall 2019, 98% of admitted students were ranked in the top 10% of their high school class with a mean unweighted academic GPA of 3.92 and the middle 50th percentile achieving SAT composite scores between 1480 and 1550. They hailed from 34 countries. Fifteen percent identified as First Generation College students. This new generation of Hopkins undergraduate students is far from the standard bell curve representation in terms of achievement, aptitude, experience and aspiration.
It is critical that methods of teaching and learning assessment are updated and improved to serve the new generation of Hopkins undergraduates. Assessment of student learning should be individually based and reflective of each student’s performance in achieving the knowledge, skills, and abilities taught in the class. In general, it is not appropriate or effective to impose a normal distribution of grades (often referred to as “grading on a curve”) on exams, assignments, or final grades. Such grading practices arbitrarily limit the number of students who can be identified as having excelled, leading to the creation of a hypercompetitive student culture. We urge that they be eliminated.
The Commission does not advocate watering down or diminishing standards. Instead it encourages the exploration and implementation of more current methods for measuring and recording student learning. It is important that best practices for student assessment be promulgated among all instructional faculty, and expectations regarding assessment be made clear at the school and departmental level to optimally support collaborative learning and creative exploration. Whatever system is used, student performance should be judged and graded relative to a standard of excellence as articulated by the faculty member and the discipline. Faculty should clearly define the knowledge, skills, and abilities that students should have achieved at the end of a course (i.e., course level learning goals) and align assessments to evaluate students against those standards.
The literature documents the effects of curving grades on student competitiveness, and its effect on campus culture, and confirms anecdotal evidence at Hopkins. Setting pre-determined quotas for the number of grades that will be assigned pits students against one another, removing the potential for a more cooperative learning environment. It leads some students to feel that they have less control over their grades and increases their stress and anxiety. The competitive environment fostered by curved grading is one factor contributing to the loss of qualified, talented, underrepresented college students from science fields (Seymour and Hewitt, 1997). Many students attending CUE2 focus groups and Coffee with the Co-Chairs meetings shared similar sentiments. One student reported that she stopped studying with classmates after she realized they were her “direct competition for a final grade.” Another student described the stress caused by his uncertainty, until letter grades were posted, about what grade his 46% class average would receive. He also described his confusion when he discovered that it meant he received an “A.” Moving away from curved grades will promote classroom community by setting the expectation that all students have the opportunity to achieve the highest possible level of excellence and that, if they do, their achievement will be reflected in their grade.
In a CUE2 commissioned qualitative survey of undergraduate faculty who taught a course of 40 or more students in the past two academic years, 28% of respondents (n=135) reported using a grading policy interpreted as relative and contributing to student competition. This includes assigning grades by natural breaks in the distribution or normalizing the distribution.
Grading policies and data concerning the use of the curve at other institutions are elusive. In an informal survey of COFHE peers, asking “At your institution, do the majority of instructors for large (>50 students) sections of UG courses use a curve to determine final course grades? (Answer: yes/no/don’t know)?”, only four schools responded. The responses were highly variable, ranging from “we don’t know” (Stanford, U Penn), to mostly no but yes in many of the Gateway Science courses from Duke, to a firm “no” from MIT.
The Commission recognizes the difficulty in writing exams and assessments that reliably challenge students at the same level each year, and faculty may need to take corrective action when an assessment is judged overly difficult. Some faculty reported adding a standard number of points to all exams in these cases.[1] In any case, faculty should clearly define what students should be able to do at the end of a course and align assessments to evaluate students against those standards. As one faculty member commented, “I compare performance against learning goals and assign grades based on mastery of material.” Alternate practices to curving have been well-documented, and include straight grading, specification grading, and competency-based grading.
Assessment systems in any course have been and will continue to be the purview of the faculty member teaching that course. What should be reviewed for each course is whether 1) the objective and subjective measures of expected performance are well explained to the students and 2) whether the assessment mechanism used does or does not unfairly force a normal distribution of grades. These should be reviewed by the Directors of Undergraduate Studies in each department and instances of relative assessment tools should be brought to the attention of the Vice Deans for Undergraduate Education. Further, the University’s Vice Deans of Education (VDE), a group routinely convened by the Office of the Provost, and the University Council on Learning Assessment (UCLA) should issue a best practices statement regarding student learning assessment methods.
[1] Additional faculty comments from survey supporting elimination of curves: “I compare performance against learning goals and assign grades based on mastery of material.” “I do not curve grades. I do not feel bound to give grades in any proportions; I set my standards.” “I do not curve grades. In my view, if all my students do well, or all do badly, their grades should reflect that fact.” “No, but I might adjust the final score limits slightly if I think the exam problems were unusually hard or there are any confusion about them.”
“I try to gauge the difficulty of my exams so that students have a fair chance to succeed without curving. If everybody does great, everybody should get an A; likewise, everybody should fail if nobody achieves the objectives.”
“No, I do not curve grades… I usually use rubrics to establish grading standards, particularly since I use TAs/graders. Rubrics are made available to the students at the time of the assignment so that they understand what is expected of them.”
“I don’t curve final grades, but I do sometimes curve an individual test grade if I feel that the average was low (<80%-85%). My philosophy is that the students should always know where they stand in class with regards to their final grade. If some magical curve is applied at the end, they never really know.”
“Briefly, I do not curve grades as I like to give points for mastery. That said I do not know how to design an exam with a clear point threshold for mastery for a specific grade in advance. So I rescale grade, same for everyone onto an A, B, C etc.”