Robert J. Marzano
Proficiency scoring is more precise and less time-consuming than the traditional method of assigning points to test items.
When teachers construct tests, they typically assign different points to different items, depending on the level of difficulty. For example, they might assign 1 point to easier items that require students to recall content, 5 points to more difficult items that require students to explain principles, and 10 points to complex items that ask students to apply knowledge.
The items worth 1 point are easy to score—they're either right or wrong—so they receive either 1 point or no points. But it gets more complicated with items worth 5 or 10 points. Although it's still easy to assign a score of 0 if an item is completely incorrect or 5 or 10 if an item is completely correct, how does a teacher differentiate among all the scores in the middle?
Teachers usually attempt to construct scoring schemes using an incremental approach. For example, for a 10-point item, a teacher might reason that a student got the majority of the content correct and assign a score of 7. However, after scoring a number of tests, the teacher might realize that he or she has begun assigning a score of 9 points for other students' responses to that same item even though the responses have about the same level of accuracy. The teacher would then deliberate as to whose score to change: Should the 9s be 7s or the 7s be 9s? This back-and-forth deliberation becomes even more complicated for items assigned more than 10 points.
Proficiency Scoring
These deliberations make scoring tests with multipoint items time-consuming, frustrating, and imprecise. Proficiency scoring is a much more efficient and accurate approach.
Establish Levels of Proficiency
Proficiency scoring begins by writing items (or selecting them from an item bank) that reflect three levels of proficiency: Basic refers to simple content that is foundational to understanding more complex elements, proficient refers to complex content that is the desired outcome of instruction, and advanced refers to tasks that require students to go beyond what was addressed in class.
Basic, proficient, and advanced content are usually articulated as learning goals. For example, consider the following three levels of learning goals for the topic of heredity:
- Basic: Students will be able to recognize or recall accurate statements about and isolated examples of heritable and nonheritable traits.
- Proficient: Students will be able to differentiate heritable traits from nonheritable traits in real-world scenarios.
- Advanced: Students will be able to explain how heritable and nonheritable traits affect one another.
When writing or selecting items, the teacher might use selected-response items for basic content, such as, "Circle the traits you can develop over time: shoe size, gender, knowledge of history, fear of snakes." For proficient content, the teacher might use short constructed-response items, such as, "Name three traits you like about yourself and explain whether each is heritable or nonheritable." For advanced content, the teacher might use constructed-response items that focus on students generating and defending claims, such as, "Identify one heritable trait you have and support your contention that it has had a more positive (or negative) effect on your life than a nonheritable trait has."
Code Each Response
Rather than assigning points, the teacher codes each response item to reflect one of four levels of accuracy: completely correct (CC); incorrect (I); high partial accuracy (HP); and low partial accuracy (LP).
This narrows the range of decisions a teacher has to make. Instead of determining whether to assign 8 or 9 points as opposed to 6 or 7 points for a student's response to a 10-point item, the teacher must determine where the student's response falls within a continuum of four categories only. In the case of a student being in that 6–7 or 8–9 range, the teacher would probably code the item as HP—high partial understanding.
Compute the Score
Computing a final score when using proficiency scoring is also quite different from the traditional approach. Instead of adding up the scores across all items on the test, the teacher examines the pattern of responses across the three levels of proficiency.
For example, if a student has consistently received CC (completely correct) or HP (high partial accuracy) scores on the basic items but has not received consistently high scores on the proficient and advanced items, the student has demonstrated the basic level of competence on the assessment. If a student has a strong pattern of accurate responses on both the basic items and proficient items, but not on the advanced items, the student has demonstrated the proficient level on the assessment, and so on.
A Possible Glitch
But what if students score CC and HP on the proficient and advanced items, but not on the basic ones? What final score should the teacher assign?
If the teacher has designed or selected the items so that understanding of basic items is necessary to correctly answer the proficient items and understanding of proficient items is necessary to answer advance items, then this shouldn't happen.
However, in the event that it does, the teacher might assign a tentative score of basic to the test as a whole but change that score to proficient when students demonstrate awareness of the mistakes they made on the basic items. The students might submit an explanation of the errors along with a more detailed explanation of the correct answer or discuss with the teacher why they missed the basic items. There are a number of ways that students can demonstrate competence for items initially missed, all of which require them to take some responsibility for raising their scores.1
A Better Way
Proficiency scoring is more precise and less time-consuming than the traditional method of assigning points to test items. In addition, it clarifies for students the type of content—basic, proficient, or advanced—that they must master to improve.