assessment systems logo

Computerized Adaptive Testing:
Smarter, Shorter Tests with AI

Build tests that personally adapt to each student, item by item through modern AI & machine learning.  Tests are shorter, more accurate, more secure, and more engaging.

Benefits of computerized adaptive testing

Faster tests

CAT provide 50%-90% reduction in test length – a big deal if you are paying for seat time.

Higher motivation

CAT improves motivation by only giving students items at the right difficulty for them.

Increased security

CAT tests are far harder to copy off your neighbor, cheat, or steal.

More accurate

CAT tests produce scores with greater precision than conventional tests.

What is Computerized Adaptive Testing?

Individualized

CAT adapts uniquely to each examinee, in difficulty and/or number of items.

Science-based

CAT comes from decades of research and academic publications.

Fair

Examinees have equivalently precise scores – a more fair than ensuring the same set of items

Machine Learning

CAT is based on a family of complex mathematical models called item response theory.

Easily publish adaptive tests with no coding!

Import your items into our item banking platform.  Run the item response theory calibration and CAT simulation studies for validity.  Then select your CAT options in our test publishing interface.  No need to write any code!  Our approach to Rapid Assessment Development is core to how our software is built.

You have full control over all other aspects of publishing as well, from time limits to customized messages to score reports.

Adaptive testing algorithm
adaptive testing platform

Reduce Test Length by 50% or More; Improve Security

Quickly publish your adaptive tests with a range of security options, from unique passwords to time/date windows to lockdown browser to remote proctoring.

Our easy-to-use examinee interface means that there is no learning curve for test-takers and they can focus on what is important: showing what they know. Our newest assessment platform, Assess.ai, gets you up and running quickly.

Trust our experts in adaptive testing

Want to learn more about our experience?  Our Co-Founder and the father of CAT, Prof. David J. Weiss, provides a history and introduction to CAT in the video here. Prof. Weiss directed the Psychometrics PhD program at the University of Minnesota for 4 decades; during the 1970s and 1980s, the US Department of Defense funded the seminal research on CAT that was occurring there.

Scholarly research is so important to us that our co founders also founded the International Association for Computerized Adaptive Testing (www.iacat.org).  We’d also love to help your organization migrate to this modern assessment paradigm.

multistage testing

What is computerized adaptive testing?

Computerized adaptive testing (CAT) is a simple algorithm but with complex calculations. Items are intelligently selected one at a time, matched to examinee ability. If you get a question wrong, an easier item is presented; if you get a question right, a harder item is shown. But those calculations are based on item response theory – if you just subjectively rank item difficulty, your test is just a legal landmine.

The test can also be variable-length, which means an intelligent stopping rule is implemented to maximize efficiency and fairness. Sub algorithms can govern content constraints, exposure controls, enemy items, and more.

What do I need for adaptive testing?

Before you can even consider a move to adaptive testing, you should confirm that you meet the minimum requirements:

computerized adaptive testing report
fasttest test statistics tif sem

How is an adaptive test developed?

It’s not easy to build an adaptive test.  Our experts have decades of experience and can help you get there as quickly as possible, with relevant validity documentation.

We follow the validity framework that our cofounders established years ago, which you can view at the link below.


1 Feasibility, applicability, and planning studies
Monte Carlo simulation ; business case evaluation

2 Develop item bank content or utilize existing bank
Item writing and review

3 Pretest and calibrate item bank
Pretesting; item analysis

4 Determine specifications for final CAT
Post-hoc or hybrid simulations

5 Publish live CAT
Publishing and distribution; software development