Skip to content
  • There are no suggestions because the search field is empty.

Why Impact Scores Don’t Always Match the Difference Between Raw and Scaled Scores

Learn why impact scores explain a student’s scaled score but won’t equal the difference between the raw assignment score and the scaled score.

Overview

When reviewing student performance, you may notice that the sum of impact scores (shown in the student answers hover details) does not match the difference between a student’s raw assignment score and their scaled score. This is expected behavior.

This article explains why these values differ and how to correctly interpret impact scores.

What Each Score Represents

Raw Assignment Score

This is the straight average of all question-level scores on the assignment. It reflects how many points the student earned out of what was possible—nothing more.

Scaled Score

The scaled score represents a prediction of the student's end-of-course state test. It accounts for:

  • Which standards were assessed

  • How the student performed on each standard

  • How those standards relate to end-of-course mastery expectations

Impact Scores

Impact scores show how each standard influenced the student’s predicted score. Each impact value represents how much that standard pushed the model’s prediction above or below a typical baseline student.

Why Impact Scores Don’t Equal Scaled – Raw

Because the scaled score comes from a model that uses standard-level performance, and not the assignment percentage, there is no mathematical link between:

scaled score – raw assignment score

and

sum of question impact scores

Impact scores explain how the model arrived at the scaled/mastery prediction, not how that prediction compares to the assignment average.

A student may move from 50% → 84% because the model learned that the standards they did well on are highly predictive of end-of-course readiness. This is independent of the raw assignment percentage.

What Impact Scores Do Represent

Impact scores follow this relationship:

baseline mastery prediction
+ combined impact of the standards/questions
= scaled mastery score

This makes impact scores a powerful way to understand why the model believes a student is above, below, or right at the expected baseline.

Takeaway

  • Impact scores explain the scaled score

  • Raw score is separate and not part of the model

  • It is normal—and correct—that the impact score sum does not match the difference between raw and scaled