SoulTraceSoulTrace
Adaptive Testing

Adaptive Question Selection

Also called: adaptive item selection, next-question selection

Adaptive question selection is the rule an adaptive assessment uses to choose the next item from its bank. The algorithm evaluates what is currently uncertain and selects a question expected to improve the estimate while respecting constraints such as content coverage, item exposure, fairness, and maximum test length.

Reviewed July 14, 2026 · 2 min read

Table of Contents

When faced with a complex decision, I prioritize a methodical approach over intuitive leaps.

Neutral
Strongly DisagreeStrongly Agree
/en/new-test?utm_source=glossary&utm_medium=internal&utm_campaign=glossary-assessment&utm_content=adaptive-question-selection-after-definition-cta-assessment-test

The selection loop

An adaptive assessment repeats three operations:

  1. update the current estimate from all answers so far;
  2. score eligible questions according to a selection objective;
  3. ask the best permitted question and repeat.

Eligibility matters. A useful algorithm excludes already seen items, limits near-duplicates, and enforces coverage across the intended construct.

Selection objectives

Item response theory systems often maximize information near the current trait estimate. Bayesian classifiers may select the question with the greatest expected information gain—the largest predicted reduction in uncertainty across possible profiles.

Other objectives can minimize expected error, separate the two most plausible outcomes, or balance several dimensions. These methods are related but not interchangeable.

Why the “best” item needs constraints

Pure optimization can repeatedly select similar items because they are statistically efficient. That risks measuring a narrow slice of the construct and making the experience repetitive. Content balancing ensures the result still represents the definition being claimed.

Exposure controls also prevent a small set of highly informative items from appearing for nearly everyone. Fairness review checks whether selection behaves differently across groups for irrelevant reasons.

A personality example

If two possible profiles differ mainly on assertiveness, a question about assertive behavior may add more information than another question about organization. Once that uncertainty is resolved, the algorithm can move to another dimension. Adaptation changes the route, not the underlying standard of evidence.

Go deeper: How Soultrace chooses each question

Sources

Related terms

Computerized Adaptive TestingComputerized adaptive testing (CAT) is an assessment method in which software selects each next question using the respondent's previous answers and a statistical model. The goal is to ask items that are informative for the current estimate, often reaching useful precision with fewer questions than a fixed-form test.Information GainInformation gain measures how much an observation is expected to reduce uncertainty. In adaptive testing, the system can predict the possible answers to each eligible question, estimate the posterior uncertainty after each answer, and choose the question with the greatest average reduction. The actual gain is known only after a response is observed.EntropyEntropy is a mathematical summary of uncertainty in a probability distribution. In an adaptive personality assessment, entropy is higher when probability is spread across several plausible results and lower when it is concentrated on fewer possibilities. A question can be valuable when its expected answers would reduce entropy.