Bias check is a job in Item Bank Health Check (IBHC) that scans your Collection of Items for language that may be biased, non-inclusive, or culturally insensitive. For each issue identified, the job provides actionable suggestions for how to revise the content, supporting a more equitable and inclusive assessment experience for all learners.

This job also supports the integration of custom Guidelines, enabling your organization to define and apply its own standards for inclusive language alongside Learnosity’s built-in Guidelines.

When to use this job

The Bias check job is particularly effective when you need to:

• Audit existing content: Review large Item banks to identify and address potential equity or inclusion concerns.
• Implement DEI policies: Align your assessment materials with new or updated institutional diversity, equity, and inclusion standards.
• Standardize language: Consistently apply organizational communication Guidelines across a diverse Collection of assessment Items.
• Support diverse populations: Prepare content for learners from varied cultural, socioeconomic, gender, or ability backgrounds.

Configuring Guidelines

Having navigated to Guidelines in the IBHC menu, you can click Create new Guideline and select Bias check in the menu. From this page, you can:

• Enter a title for the Guidelines in the title field
• Select the Item banks to which this set of Guidelines will apply
• Select up to 15 Guideline categories from the existing set or create your own custom categories
• Edit any Guideline category with the refinement options by clicking the button on the right of the category's row

Find further information on Guidelines here.

Running the job

• Open the Learnosity Console and navigate to Item Bank Health Check.
• Create a Collection or select an existing one with the Item bank you want to analyze.
• Click on Create and select Bias check from the menu of available job types.
• Select the Guidelines from the menu of Guidelines configured.
• Initiate the process by clicking Run job.

Reviewing results

Following the analysis of the Collection, a review space lists each Item where potential issues were identified. For each flagged Item, it:

• Lists all categories from the selected Guidelines and whether they passed or failed the check.
• Highlights the specific text that was edited.
• Displays the reason a category failed the check and why that specific edit was made (e.g., gender-coded language, stereotyped framing).
• An option to accept a suggestion, edit the text manually, or dismiss the flag if you determine it is not applicable.

We recommend exercising editorial judgment to ensure that revisions align with your pedagogical style. While the AI identifies concerns, human review is recommended to determine if language is appropriate for the specific subject matter context.