For over 100 years, teacher-assigned grades have been found to reflect a mixture of academic and nonacademic factors. Teachers often give grades for classwork and homework, for example, and then count them in the student’s report card grade. Thus, two students who ultimately reach the same level of accomplishment might receive different grades, depending on how much practice each one needed to get there. Sometimes, teachers use point systems for grading that include points for assignment completion, studying, and extra credit, often not realizing that such points inject a measure of behavior, as opposed to achievement, in a grade. Teachers weight achievement, behavior, and effort differently for low-achieving students than they do for high-achieving students. Thus, the amount of information a grade communicates about student learning varies from student to student.
Research on feedback and formative assessment also speaks to grading practices. Students are more able to regulate their learning if they have a clear idea of what they are supposed to be learning and receive ungraded feedback along the way, with opportunities to apply that feedback. Ungraded feedback—which can be in the form of scores or comments but does not punish students for their attempts at learning—is the kind of feedback most students prefer and are most willing to use for improvement. In this case, the grade from final, summative work can reflect what students learned, not how hard they practiced. Reforming grading practices has proved difficult for many years, but the current disruption in education presents a unique opportunity to move closer to grading practices that provide better information about student achievement.
Another line of research focuses on what is graded in the first place. In current practice, examinations, tests, and quizzes carry a lot of weight in determining a grade, increasing in importance from elementary through middle and high school. However, to be meaningful indicators of learning, the information grades convey should come from assessments that are well aligned with the learning standards. For many 21st-century standards that implies the use of more performance assessment. A performance assessment requires students to create a product or demonstrate a process, or both, and it is evaluated by observation based on clearly defined criteria. A performance assessment has two parts: a task or tasks and rubrics or some other presentation of criteria and scoring rules. Performance assessments can: (a) measure complex thinking and application that cannot be meaningfully assessed by traditional on-demand tests, which are often designed to assess rote memory and simple comprehension; (b) make the assessment itself an episode of learning; and (c) signal for teachers what and how to teach, and for students what is important to learn.
Research also suggests that when students perform rich and meaningful tasks—especially when they are asked to explain their reasoning and when they are asked to self-assess—achievement increases, especially for middle- and higher achieving students. This is likely because lower achieving students are often, currently, not given the same opportunities to learn complex thinking as their peers. Providing that opportunity to learn is critical for helping all students with 21st-century learning standards.
Especially during school disruptions, when students have unequal access to technology and learning supports, all students should receive grades that report to what extent they have learned what they were taught. Research on grading and assessment practices provides lessons for the types of approaches schools could use in the year ahead:
1. Replace grading quantities like the extent to which students have completed assignments with grading qualities in student work that reflect students’ current achievement level at the time of the report. Give feedback, not a grade, to formative practice work. Design summative (graded) assessments to find out what students have learned after instruction, using rubrics or other scoring schemes tied to learning rather than to completion or compliance with directions. Alternatively, summative decisions can be made by applying rubrics to a body of work, again focusing on learning rather than completion. Separate measures of attendance, work habits, and other factors from the measures of achievement that feed into students’ grades, to allow reporting them without confounding the achievement grade.
2. Use flexibility in timing the collection of evidence for grading decisions so students are graded on the learning they do, not when they do it. Examples of flexible timing include:
- Use ungraded (formative) practice work until students reach a desired level of achievement, then collect evidence for grading
- Give students choice in formative work, so that they may practice as much or as little as they feel they need before being graded.
- As appropriate, negotiate due-dates. Practices like these differentiate the schedule of learning for different students without changing the substance of what is to be learned, thereby increasing access to standards for more students. Some flexibility in timing is already built into remote learning, at least some of which is asynchronous anyway; the current context offers a golden opportunity to move to more flexible timing in assessing learning.
3. Report current evidence of student achievement. Replace, typically by overwriting an entry in the gradebook, a grade from a prior assessment of a learning goal with a newer grade for the same learning goal, reporting the most current estimate of achievement. Often a series of assignments will show a pattern of gradual improvement, and a grade can report what students know and can do at the end of the sequence. Carefully crafted grading policies about retakes and revisions can deter students who did not succeed the first time from giving up. Such policies should require students to explain what they did to improve and what they learned from doing it.
4. Use multiple measures for assessing and reporting student achievement on any learning goal. A standard usually represents a domain of learning that is broader than any one assessment. Standards-based grading communicates how students are doing in each standard, which is useful for learning as well as reporting. For standards-based grading, base the report card grade for a standard on more than one assessment—and depending on what the standard says, on more than one format. For many standards, sounder grading decisions occur when performance assessment results are combined with or substituted for test results. However, if you are using conventional report cards where several standards are summarized in one subject-area grade, the same principle applies: assess each standard taught with more than one assessment and then aggregate them.
5. Use performance assessments to allow students to demonstrate mastery of complex content. Design performance assessment tasks that are more than just engaging activities and that require extended thinking, planning, and application. Align both the task and the criteria for assessment with the learning goal. In the context of remote learning, tasks must be accessible to students in a variety of home contexts. Design rubrics and evaluation rules to result in individual grades, not group grades, for accurate reporting of individual achievement. Use criteria about learning—not about compliance with directions or with surface features of the work—that are accessible to students for self-assessment as they work. Many resources for educators give examples of how to construct performance assessments; an abbreviated list of resources is given below.
The current disruption in education offers a unique opportunity to increase flexibility in grading practices and enrich assessment methods. Recommendations include the following:
- Assign grades that report students’ current achievement of clear learning goals.
- Use flexible timing for collecting evidence of learning.
- Consider grades as temporary, subject to revision as more recent evidence becomes available.
- Use multiple measures for each learning goal.
- Increase the use of performance assessment.
- Choose performance tasks that require extended thinking, planning, and application of knowledge or demonstration of skills.
- Use criteria and rubrics that indicate achievement of learning goals, not compliance with directions or surface-level features of students’ work.
Endnotes and an extensive resource list can be found in the full practice brief here.