The Science Behind Better Questions

Archaine uses Item Response Theory (IRT) - a psychometric framework that models the relationship between a test-taker's ability and their probability of answering a question correctly.

1

Define Your Assessment Goals

Tell Archaine about your assessment:

  • Target ability level
  • Number of questions needed
  • Subject matter and difficulty range
  • Desired psychometric properties
2

IRT-Based Question Generation

Our platform generates questions calibrated with three key IRT parameters:

Difficulty (b-parameter)

Determines how hard a question is. Questions are calibrated to match your target ability distribution.

Discrimination (a-parameter)

Measures how well a question distinguishes between high and low ability test-takers. Higher discrimination = better differentiation.

Guessing (c-parameter)

Accounts for the probability of guessing the correct answer. Archaine optimizes this to minimize random success.

3

Quality Assurance

Every generated question undergoes:

  • Psychometric validation - Ensures parameters fall within acceptable ranges
  • Content review - Checks for clarity and relevance
  • Bias detection - Identifies potential fairness issues
4

Download & Deploy

Export your validated questions in your preferred format and deploy with confidence, knowing each item meets psychometric standards.

Ready to Create Better Assessments?