Adaptive Testing and Computer Adaptive Testing (CAT) Flashcards

In a Computer Adaptive Test (CAT), how is the difficulty of the next item typically determined?

Answer: Based on the examinee's response to the previous item

CAT algorithms select the next item based on the estimated ability level derived from the examinee's responses to prior items.

Which psychometric model is most commonly used as the foundation for Computer Adaptive Testing?

Answer: Item Response Theory (IRT)

IRT provides item-level ability estimates and item characteristic curves that allow CAT algorithms to adaptively select items and estimate examinee ability in real time.

What is the primary advantage of a CAT over a fixed-form test of the same length?

Answer: Greater measurement precision across all ability levels

CAT targets items near each examinee's ability level, yielding more precise ability estimates than fixed-form tests that include many items too easy or too hard for a given examinee.

Which stopping rule in CAT terminates the test when the standard error of the ability estimate falls below a specified threshold?

Answer: Variable-length stopping rule

Variable-length stopping rules end the test once measurement precision reaches a desired level, meaning each examinee may answer a different number of items.

Item exposure control in CAT is implemented primarily to:

Answer: Prevent overuse of items that could compromise test security

Exposure control limits how frequently any single item is administered so that popular items are not memorized and shared by examinees, protecting test security.

In IRT-based CAT, the Maximum Information (MI) item selection algorithm selects the next item that:

Answer: Provides the most information at the current ability estimate (theta)

The MI algorithm selects the item whose Fisher information function value is highest at the current theta estimate, maximizing measurement precision at that ability level.

Which of the following is a key challenge when implementing content balancing constraints in a CAT?

Answer: Ensuring blueprint specifications are met while still selecting near-optimal items

Content balancing requires selecting items that satisfy blueprint content constraints, which may force the algorithm to choose a slightly less informative item, creating a trade-off between optimality and validity.