How PASL Tasks Are Scored

The PASL assessment has three tasks that are centrally scored by ETS-trained raters using rubrics that contain quantitative and qualitative elements of evidence.

When scoring your task, raters review your written commentary, the artifacts that you linked to your written commentary, and video. At least three raters contribute to scoring your assessment.

On your score report you’ll see step scores, task scores, and your overall assessment score.

Step scores

There are a total of 12 steps in this assessment. Step scores are determined using a 4-point rubric. Score levels for each rubric are defined below,

Score	Quantitative and Qualitative Elements of Evidence
Score of 4	Consistent and thorough
Score of 3	Effective and appropriate
Score of 2	Partial and inconsistent
Score of 1	Minimal and ineffective
Score of 0	Blank, insufficient evidence, no required artifacts linked to written commentary

When multiple raters score a task, the step scores are averaged.

The rubric documents contain the four task-specific rubrics used during scoring to evaluate the evidence you provided for each step.

Steps that are determined to be nonscorable receive a score of 0. See Understanding Your Scores.

Task scores and overall assessment score

Your task scores are the total of the step scores for each task. The score for Task 3 is multiplied by 2 to reflect the double weighting of the task. Tasks that aren’t submitted receive a score of 0.

Your overall assessment score is the total of your three task scores.

Fair, unbiased, and accurate scoring

All raters are required to:

participate in a rigorous training program that includes demonstration of understanding of the standards, task directions, rubrics, and more
demonstrate mastery of the scoring process through multiple practice sessions conducted by experts who are trained in qualifying raters
take and pass a certification test verifying their mastery of accurate scoring processes every time they score

All identifying information provided by you at registration is removed from responses to prevent raters from knowing your identity. When scoring resubmitted tasks, raters don’t have access to any prior scores.

Scoring follows a conditional double scoring model in that each task will be single scored, but if a total task score is within a defined proximity of reaching the overall passing requirement, at least one additional rater will independently score the task.

Scoring leaders are assigned to supervise all scoring activity. As part of this work, scoring leaders also conduct random back readings of tasks that were scored. If an error in scoring is detected, the error will be corrected.