MLA-C01 ML Model Development Practice Question

A data scientist wants to use SageMaker Clarify to analyze bias during training of a binary classification model. Which TWO types of bias metrics can SageMaker Clarify compute? (Select TWO.)

⚠ Common exam trap

MLA-C01 often tests the confusion between bias metrics and explainability metrics, leading candidates to select feature importance or SHAP values as bias metrics.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Post-training bias metrics (e.g., Difference in Positive Proportions, AD)

SageMaker Clarify computes bias metrics in two distinct phases of the ML lifecycle, and both are valid answers here. Option B is correct because post-training bias metrics evaluate the model's predictions after training, using measures such as Difference in Positive Proportions in Predicted Labels (DPPL), Accuracy Difference (AD), and Disparate Impact, which quantify whether outcomes differ across groups. Option D is correct because pre-training bias metrics analyze the training data before any model is built, using measures such as Class Imbalance (CI), Difference in Proportions of Labels (DPL), and Label Imbalance to detect skew in the dataset itself. Option A is not a bias metric category — feature importance (e.g., via SHAP) explains which features drive predictions, not whether outcomes are biased. Option C is likewise an explainability technique, not a bias metric, and Clarify reports SHAP values under its explainability analysis. Option E is a standard model evaluation artifact for classification performance, not a bias metric computed by Clarify.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Feature importance
Why it's wrong here
Feature importance ranks attributes by their influence on predictions, an explainability output rather than a bias metric. It is tempting because biased models often show skewed importance, but Clarify's bias metrics quantify outcome disparity across facets, while importance belongs to the explainability report.
✓
Post-training bias metrics (e.g., Difference in Positive Proportions, AD)
Why this is correct
Post-training bias metrics are computed from model predictions on a dataset, comparing predicted labels across facets; measures such as Difference in Positive Proportions in Predicted Labels and Accuracy Difference quantify bias introduced by the trained model itself.
✗
SHAP values
Why it's wrong here
SHAP values explain individual feature contributions to predictions, which Clarify produces for explainability, not for bias detection. It is tempting because both features appear in the same Clarify job, yet bias metrics such as disparate impact and conditional demographic disparity are computed separately from SHAP-based explanations.
✓
Pre-training bias metrics (e.g., Class Imbalance, DPL)
Why this is correct
Pre-training bias metrics are computed from the dataset and its facet labels before any model exists; measures such as Class Imbalance and Difference in Proportions of Labels reveal imbalances in the training data that could bias the binary classifier.
✗
Confusion matrix
Why it's wrong here
Confusion matrix counts are performance outputs, not bias metrics; Clarify reports them under model quality, not bias. It is tempting because class-wise counts underpin metrics like disparate impact, but Clarify computes those bias metrics directly from label and prediction data rather than exposing the matrix itself as a bias type.

Go deeper

MLA-C01 ML Model Development Practice Question

Why each option matters

Option-by-option breakdown

Related to this question