What are the practical considerations in scoring a test?
Now let’s consider the following aspects of scoring a test:
What is your scoring method?
- A holistic approach (one overall score)?
- An analytic approach (separate scores for separate aspects of the performance)?
- Counting the number correct?
This is related to what you are testing (e.g., it’s harder to count the number correct in a writing prompt), resources (e.g., do you have time to score each writing piece analytically?), and how you intend to use the test (e.g., if you want to provide specific feedback to students, an analytic rubric might be useful).
Who is involved in scoring?
Is the test…
- Machine- or computer-scored?
This is related to the resources you have for the test, especially time, expertise, and budget.
Where is the test scored?
- Local/On site?
- Central/Off site?
This is related to the resources you have for the test, especially time, expertise, and budget. It may also be important when thinking of the effects of testing and use of the results. For example, if the test is used to evaluate a program or to determine whether a student can graduate, you may decide it is better to have it scored centrally by people who do not know the student, who have undergone rigorous rater training, and who have no stake in the outcomes. Or, you may decide it is better to have on-site raters who know the students to score the test.