Grading Rubrics

Rubrics are used to communicate expectations and provide feedback to students on their learning. Standards-based rubrics clarify for students and instructors what it means to demonstrate full achievement of a learning objective.

Consider this learning objective for first-year engineering students learning MATLAB (computational tool): Create and evaluate a x-y plot suitable for technical presentation. A demonstration of full achievement of this learning objective would entail: (1) using the plot command, (2) using the hold command or an equivalent method to put more than one data set on the same plot, (3) placing the independent and dependent variables on the correct axis, (4) using appropriate data markers and lines formats and providing a legend, (5) placing an appropriate title on the plot, and (6) placing appropriate labels on the axes.

To design a rubric, careful articulation of what full and various lower levels of achievement entail is a step towards ensuring consistent assessment of student work by one grader or across multiple graders. When a problem can result in many different solutions, one approach to differentiating the levels of achievement is to look at the number of fully achieved elements that have been satisfied.   

Another consideration for the design of a rubric is the number of and meaningful names for the different levels of achievement. Four levels provide sufficient differentiation between levels of achievement. A variety of naming conventions are used:

  • No evidence, Underachieved, Partially achieved, Fully achieved
  • Needs practice and further development or not assessed, Approaching appropriate development, Demonstrates appropriate development, Strong development
  • Basic, Developing, Accomplished, Exemplary
  • Needs Work, Adequate, Excellent
  • Does Not Meet Expectations, Meets Expectations, Exceeds Expectations

To demonstrate how a rubric item might be used, consider this problem to assess a student’s achievement of the plotting learning objective: On the same MATLAB plot, plot the populations of Rabbits versus Time and Wolves versus Time. Plot the rabbit population using a solid blue line and circles as data markers. Plot the wolf population using a dashed red line and x’s as data markers. Include appropriate axis labels and a title.

In the example student’s work, the student only misses which command to use to label the axes. So this student would receive feedback via the rubric that this learning objective is partially achieved.

When learning objectives are repeatedly assessed across multiple assignments, it is possible to track individual student and class performance.  Here the class demonstrated poor achievement of this learning objective on the exam as compared to the problems sets leading up to the exam. This information enables the instructor to investigate the causes of poor performance.  Was it the design of the problem or the rubric? Did students not have enough practice with the things they missed?