Adapted from Effective Grading (Barbara Walvoord and Virginia Johnson Anderson): Grading isn't merely marking red ink on a page, but a process of sending subtle messages to students about what you value as a teacher. Grading is "a context-dependent, complex process that serves multiple roles," including evaluation, communication, motivation, and organization.

We evaluate student work, providing them with feedback about the quality of their work. Grading also communicates to students, as well as advisors, parents, future employers and other important groups that have a stake in student success. Grading can also motivate students, often determining the extent to which a student is participating in a course. And finally, grades can serve as organizational tools of the teacher, bringing closure to concepts or units and shifting focus from one learning module to another.

Before beginning, it's important to dispel three commonly-held assumptions about grading: it can be totally objective, everyone is in total agreement about a grade, and grading is a "one-dimensional student motivation for learning" (Walvoord and Johnson 10). In fact, grading can be an extraordinarily messy, complex, subjective process that will ultimately be ignored by some students.

Step 1: Create Effective Assignments

As you have already discovered from earlier parts of this packet, effective course design leads to effective assignments. Making sure that assessment tools fit learning goals and objectives and are feasible in terms of workload.

Also, make sure assignment and test instructions are clear to students; if they can't decipher the questions, they are unlikely to give you the answers you are expecting. Students may have learned the material but are unable to show you they've learned it.

Step 2: Teach to the Test

If your assignments are created from well-constructed learning objectives and are feasible, then teaching to the test isn't a problem, as you are in essence teaching to the objectives.

Step 3: Determine Criteria and Grading Schema

If you're a graduate teaching assistant serving primarily as a course grader, assignment criteria and grading schemas will be determined by the professor. In some situations, you may be establishing criteria or schema in concert with a professor, or determining grading all on your own if you are the instructor of record for a course.

Step 4: Primary Trait Analysis

To assist you with this process, Primary Trait Analysis can be used as "a way of explicitly stating the teacher's criteria, and it is used in the classroom to make criteria and standards clear to themselves and their students and to guide classroom teaching and learning" (67).

"For three semesters, Anderson asked first-year students in Biology 101 to answer this test question:  'In a coherent paragraph, compare and contrast prokaryotic and eukaryotic cells in four ways.'  Each semester, a large number of students did poorly on this fifteen-point question designed to measure vocabulary usage, concept master, and information integration.  So Anderson ordered a better movie on the cell than she had been using and beefed up her lecture coverage, but still there was no significant improvement.  Then an indignant student in the lab—armed with a pen and a textbook—demanded to know why he only got half credit when all four answers were right.

"Then it dawned on Anderson:  she had let the atrocious 'it-has-a-nucleus' topic sentences, the membrain misspellings, and the scrambled prose of poor answers hide the truth.  The students didn't need more biology—they needed to know how to compare and contrast, which was exactly what Anderson required.  Anderson assumed students understood that the answers to her questions comprised eight components:  four comparisons and four contrasts.  She had been so busy grading the seventy-five-plus papers that she hadn't been analyzing.  The grading process was all wrong.  Anderson saw that she should have constructed explicit criteria for grading rather than rely on her global judgment simply to award overall points.

Anderson recognized that she needed to state her criteria much more explicitly for her students and to intervene in the learning process much sooner.  So she kept the better movie, dumped the extra lecture material she had added, and stated exactly what she expected in a comparison-contrast paragraph.  Then she selected several sample student compare-contrast writings, read them in class (without identifying the student writers by name), and graded them orally, making her new and clearly constructed criteria extremely explicit.  Finally, she made students practice.  She asked them to compare and contrast plant and animal cells in three ways as a 'ticket' to a lab.  Collected at the door of the lab, these writings functioned as tickets normally do—students had to have one in order to enter.  And student test scores did indeed improve" (66).
--Walvoord and Johnson (1998).

To construct a primary trait analysis scale, you must first identify the traits or characteristics that you are looking for in student work. Refer back to your learning objectives for that particular learning module: What expectations do you have for performance?

Next, build a scale for scoring the student's performance on that trait. What would a more polished assignment look like? What are the gradations of performance that you expect on the assignment . . . simply pass-fail, or could students be assessed up to five distinct levels of performance for an assignment?

Finally, evaluate the student's performance based on the traits and your scales. Of course, revision should be expected. You may not be comfortable with your initial results after using the scale, but just revise it until you are satisfied with its performance.

Sample PTA Scale

Anderson asked upper-level biology students to create and conduct a scientific experiment in which they compare two products and report results on the comparison in a scientific report. Anderson employed the primary trait analysis method and began by selecting ten traits she wanted to measure:

She then developed her scale that describes each level's performance. To illustrate, let's examine her scale for Trait #4, the methods and materials section:

Level 5 – contains appropriate, quantifiable, concisely organized information that allows the experiment to be replicated.  All information in the report can be related back to this section.  Identifies sources of data.  Sequences information appropriately.  No wordiness.

Level 4 – as above, but contains unnecessary information or wordiness

Level 3 – experiment could be replicated from the information given.  All information in the report can be related back to this section.  However, fails to identify some data sources or has problematic sequencing.

Level 2 – marginally replicable.  Parts of basic design must be inferred.  Procedures not quantitatively described.  Some information in Results or Conclusions sections cannot be anticipated by reading this section.

To construct a PTA scale, begin with the "traits" that you are evaluating for a particular assignment, such as "thesis," "eye contact with client," "use of color," or "control of variables." Next, for each trait construct a 2-5 point scale, which are descriptive statements that illustrate what performance at each level would look like. For example, a "Level 5" thesis can be described as focused enough to be an appropriate topic for the assignment, clear to the audience, written in a style that suggests the author is a member of the discipline, uses sources to support arguments, and indicates the author has synthesized this research and offers new insights.

After developing your scale, try out the scale with a sample of student work or review with colleagues and revise. You may be unhappy with the results the first time using it, so do not be afraid to make adjustments.