Using Rubrics to Evaluate Written Work

Students need to know how instructors assign grades on essay exams, term papers, or lab write-ups for the assessment exercises to be useful learning experiences. Instructors must be able to explain how they arrived at the grade they have assigned to a piece of written work in order to avoid concerns about capricious grading. The use of a grading rubric by instructors addresses both of these demands.

A grading rubric is nothing more than an instructor's explicit statement of his or her expectations for student responses at various levels of achievement. How does an “A” lab report differ from a “B” or a “C” report? What does a “B-“ paper have that a “C+” paper doesn’t?  When developing grading rubrics, you must think through each expectation and identify clear and observable elements of content, style, and quality that can be recognized in student responses. It’s generally good practice for instructors to distribute the grading rubric along with the written assignment so as to make their expectations clear and to inform students of the grading process. Although some criticize the sharing of a rubric as “giving the answer away” to the students, actually it is simply helping the students know what you expect. After all, students are not mind readers.

Analytic and holistic rubrics

The two types of grading rubrics, analytic and holistic, both require you to specify those elements that are expected to be present in a response of superior quality. Using an analytic rubric, instructors assign a score for each element and then compile these scores into a single grade. Using a holistic rubric, you still have a list of key elements that must be present in the work and you still must be able to describe varying levels of achievement, but ultimately you assign a single grade or score based on your overall impression of the quality of the student response in meeting those predetermined criteria.

Analytic rubric

When you create an analytic rubric to grade written work, you can vary the maximum number of points for each element. This is helpful when you are grading for several features but you don't want, for example, grammar to count as much as content. To create an analytic rubric you need to do the following:

  1. List all of the elements on which the written responses will be evaluated.
  2. Decide the maximum number of points to give each element.
  3. Write descriptions for each point range for each element of the writing.
  4. Determine the number of total points for different grades (e.g., 45 out of a total of 50 points represents an “A” response, 40-44 points is a “B” response, etc.).

For example, you decide to grade essay responses on the following elements and assign differing point totals for each (in parentheses):

  • Clearly stated thesis sentence (5 points)
  • Expression and organization of ideas (15)
  • Use of appropriate supporting examples (12)
  • Grammar (5)
  • Spelling (3)
  • Effective closing paragraph (5)

You then write a description for each point range for each element of the writing. For example, to define a “clearly stated thesis statement,” perhaps you decide to assign the following:

5 points: clearly stated thesis sentence, point of view is maintained throughout essay

4 points: stated thesis sentence but it is not maintained throughout essay

3 points: the main idea is implicit in the writing but never made explicit

2 points: the main idea needs to be guessed at

1 point: there is no thesis sentence to guide reader; student jumps into topic

0 points: student made no effort to complete task

If you cannot figure out enough appropriate descriptors, you may need to change your point scale. For an element with very few points, like (e), you might assign two points if 75 percent or more of the words are spelled correctly, one point if between 50 and 74 percent of the words are spelled correctly, and zero points if less than half of the words are spelled correctly.

For a category with many points, like (b), you can create a range like this:

15 points: unambiguous expression of ideas, smooth-flowing organization and coherence

12 points: clear expression of ideas, basically good organization and coherence

9 points: good expression of ideas, occasional lack of organization and coherence interrupt essay

6 points: broadly stated ideas, poor organization and some incoherence leave reader confused and struggling but still with a sense of the gist of the essay

3 points: ideas are not clear, organization is muddled, coherence is lacking throughout

0 points: student made no effort to complete task

Obviously, when using the above scoring method you cannot give a score of 11 or a score of 5. If a student achieves a 9, it means the student has not demonstrated enough ability to achieve a 12. To finish this example, a student would be given a total essay score of 45 (out of 50 points) if they received: 5 points for (a), 12 for (b), 4 for (c), 15 for (d), 4 for (e), and 5 for (f).

Holistic rubric

One familiar holistic rubric is the five-point scale of A, B, C, D, and F. Most holistic scales do not have more than 10 points because it becomes difficult to differentiate among the levels. If you have never created a holistic scale before, try beginning with only five levels where the lowest score (e.g., a zero or a F) indicates a lack of achievement and the highest score (e.g., a 4 or an A) indicates exceptional achievement. Note that the lowest score does not necessarily mean that the student literally knows nothing. Rather, it means that the student failed to reach even your minimum expectations for the task.

To create a holistic scoring rubric you need to do the following:

  • Decide your range of grades or scores (e.g., A-F, 0-4).
  • List all of the elements on which the written responses will be evaluated.
  • Write a description of each element for each score level. Each score range indicates varying levels of expertise.

For example, if you decide to grade essay questions on the basis of organization, use of transition markers, effectiveness in conveying one's meaning, use of vocabulary, and mechanics (spelling, grammar) you will need a description of each of these elements. The grading of the organization element may look something like the following:

A:  Contains clear introduction, development of ideas, and conclusion.

B:  There is some logic but parts are not fully developed.

C:  It is extremely simple or disorganized.

D:  It is disjointed, rambling.

F:  It is unclear.

Descriptions need to be developed for all elements so grades can be calculated for each. Often, students' overall grades are calculated by taking the “eyeball average” of all of the element grades. Thus, the overall grade is the one that best describes the student's work as a whole. It becomes a bit of a juggling act when a student is strong on one element but weak on another. For example, if a student is great with ideas but poor in organizing them, you still must choose just one grade that best reflects your evaluation of the response as a whole. The decision is subjective, and in this case you will most likely choose a rating in-between the two scores. This level of subjectivity is the hallmark of holistic rubrics; by more precisely specifying how the overall grade should be determined by the scores on the individual elements, one changes a holistic rubric to an analytic one.

Choosing between holistic and analytic rubrics

Choose to use a holistic rubric if your time for grading is quite limited, if students will not receive back their papers (e.g., a final term paper) and thus detailed feedback is unnecessary, or if the written assignment is intended to assess creativity.

Choose to use an analytic rubric if the purpose of the assignment is to assess specific aspects of the content that you have been teaching, to elicit specific features of an approach to writing or solving problems, or to give students diagnostic feedback on their performance.

You can read more about using grading rubrics, or what Walvoord and Johnson call “Primary Trait Analyses” in their book Effective Grading (1998, Chapter 5).

Resources

  • Stevens, D. D., & Levi, A. J. (2004). Introduction to Rubrics. Sterling, VA: Stylus.
  • Walvoord, B. E., & Anderson, V. J. (1998). Effective grading: A tool for learning and assessment. San Francisco: Jossey-Bass.