Key Metrics Dashboard
% Interviews Audited
Displays the proportion of total interviews that have undergone automated or manual bias auditing.
Flag Rate
Indicates the percentage of interviews that have triggered one or more bias-related alerts.
Actions Closed
Shows the number of bias-related cases that have been reviewed and resolved.
Adverse-Impact Matrix (AIM)
Adverse-Impact Matrix Evaluates fairness across recruitment stages by comparing selection outcomes between demographic groups.Controls
- Weighted: Toggle to apply weighted calculations based on group size.
- Show n: Toggle to display sample sizes for each group.
- Suppress <N: Input setting to hide values with insufficient sample sizes.
Legend Description
Explains the meaning of matrix color indicators that represent fairness thresholds for comparing demographic groups.Parity & Drift Timeline
Parity & Drift Timeline A line chart that visualizes fairness trends over time, tracking changes in parity and potential scoring drift.Controls
- Metric Dropdown: Selects the fairness metric to display.
- Time Period Dropdown: Defines whether the chart displays daily, weekly, or monthly intervals.
- Date Range Selector: Allows viewing trends within a specific time window.
- Reference Lines: Indicates target fairness levels and watch thresholds.
Interviews (Audit/Investigation) Table
Interviews Audit Table A review table listing interviews that triggered bias checks, with expandable details for deeper investigation.Table Controls
- Export Button: Allows downloading audit data.
- Search Bar: Search interviews by ID, role, or flags.
- Filter Button: Apply advanced filtering options.
Interview ID
Interview ID
- Identifier assigned to the interview under review.
Role / Stage
Role / Stage
- Specifies the job role and recruitment stage where the interview took place.
Decision + Score
Decision + Score
- Displays the outcome of the interview and the evaluation score assigned.
Bias Score
Bias Score
- Shows the bias risk level detected for this interview.
Flags
Flags
- Lists the bias-related alerts triggered during the evaluation.
Reasoning (Expandable)
Reasoning (Expandable)
- Detailed narrative explaining why the interview was flagged, displayed when expanded.
Protected Attribute
Protected Attribute
- Indicates whether demographic attributes were considered during the audit.
Model / Rubric
Model / Rubric
- Shows which model version and evaluation rubric were used.
Reviewer & Actions
Reviewer & Actions
- Displays reviewer information and available action buttons such as assignment or re-interview.
Timestamp
Timestamp
- Indicates when the interview was evaluated or flagged.