Introduction to Adaptive Assessments

What is an Adaptive Assessment?

An adaptive assessment is a test that changes based on how students respond. Instead of giving every student the same questions in the same order, adaptive assessments adjust to each learner's needs in real-time.

Think of it like a conversation with a student. If they answer correctly, you might ask a more challenging follow-up question. If they struggle, you might offer a hint or ask a simpler question to build their confidence. Adaptive assessments automate this responsive teaching approach.

Why Adaptive Assessments Matter

Traditional assessments treat all students the same way. Adaptive assessments recognize that:

Students learn at different paces - Some need more support, others need greater challenges
Immediate feedback helps learning - Students benefit from hints and guidance during the assessment
Efficient testing saves time - Students don't waste time on questions that are too easy or impossibly hard
Reduced test anxiety - Students work at their level rather than feeling overwhelmed

Types of Adaptive Assessments

1. Progressive Hints (Within a Question)

The assessment provides scaffolding within a single question based on student responses.

Example: Math Problem

A student attempts to solve: "What is 15% of 80?"

First attempt: Student enters wrong answer
System response: "Remember, 'of' means multiply. Try converting the percentage to a decimal first."
Second attempt: Student tries again with the hint
System response: If still incorrect: "Start by converting 15% to 0.15. What is 0.15 × 80?"

This is called an adaptive item - a single question that provides progressive support.

2. Branching Paths (Between Questions)

The assessment routes students to different questions based on their performance.

Example: Reading Comprehension

Question 1: Basic comprehension question
If correct: Student proceeds to inference questions
If incorrect: Student gets vocabulary support questions first
Question 2a: (Advanced path) Asks students to analyze author's purpose
Question 2b: (Support path) Reviews key vocabulary, then returns to comprehension

This is called test-level branching - the assessment chooses which questions to ask.

Benefits for Teachers

Better Diagnostic Information

Adaptive assessments reveal:

Where students struggle - You see not just wrong answers, but what hints they needed
How students think - The path they took shows their problem-solving approach
Prerequisite gaps - Branching paths identify foundational skills that need review

Differentiation at Scale

You can serve diverse learners without creating multiple versions:

Advanced students stay engaged with challenging problems
Struggling students receive support without feeling singled out
English Language Learners can access vocabulary help when needed

Reduced Grading Time

Many adaptive assessments can provide:

Automatic scoring with detailed feedback
Partial credit for showing work across attempts
Reports showing which students need intervention

Real-World Applications

Formative Assessment

Use adaptive items during instruction:

Check for understanding with questions that provide immediate hints
Build confidence by guiding students to correct answers
Identify misconceptions by tracking which hints students need

Summative Assessment

Use branching tests for final evaluations:

Efficient testing - Students only see questions at their level
Reduce guessing - Harder to game the system when questions adapt
Fair assessment - Students demonstrate mastery at their current level

Practice and Review

Create adaptive practice problems:

Homework support - Built-in tutoring through progressive hints
Test preparation - Students work through problems with scaffolding
Self-paced learning - Students can retry with different hints

Getting Started

Start Simple

Begin with a single adaptive item:

Choose a problem where students commonly need support
Identify 2-3 hints that would help struggling students
Create feedback for correct answers to extend learning

Plan Your Paths

For branching assessments:

Identify foundational concepts that students must master first
Create parallel questions at different difficulty levels
Map decision points: "If correct, then..." and "If incorrect, then..."

Use Technology Standards

Look for assessment tools that support:

QTI (Question and Test Interoperability) - Industry standard for adaptive assessments
Export capabilities - Ensure you own your assessment content
Learning management integration - Works with your existing tools

Common Questions

"Won't students just game the system by answering wrong to get hints?"

This is actually a feature, not a bug! If a student needs hints to learn, they should get them. However:

You can score by number of attempts (first try = full credit, with hints = partial credit)
The goal of formative assessment is learning, not just correct answers
Adaptive systems track all student actions for teacher review

"How do I ensure all students cover the same material?"

You don't need to! The goal is mastery, not uniformity:

Define learning objectives, not question sequences
Students reach the same goals through different paths
Use summative assessments to verify all students meet standards

"Is this more work to create?"

Initially, yes. But:

One adaptive assessment can replace multiple versions
Reusable components save time in the long run
Better data reduces time spent diagnosing problems later
Modern tools make creation easier than building from scratch

Next Steps

Ready to explore adaptive assessments further? Consider:

Try an example - Experience an adaptive assessment as a learner
Review QTI standards - Learn how adaptive content is structured
Start small - Convert one existing assessment to include hints
Collaborate - Share adaptive items with colleagues

Remember: The best adaptive assessment is one that helps your students learn. Start with that goal, and let the technology support your teaching expertise.

Resources

QTI Overview - Technical standards for adaptive assessments
Creating Adaptive Items Tutorial - Step-by-step guide
Branching Logic Examples - Real assessment paths
Assessment Design Principles - Best practices for adaptive testing

How QTI Manages Adaptive Responses

QTI handles adaptive responses in two main ways: adaptive items (within a single question) and test-level branching (routing between different questions). Here's how each works:

1. Adaptive Items (Within-Item Branching)

Adaptive items allow an item to be scored adaptively over a sequence of attempts, enabling the candidate to alter their answer following feedback or to be posed additional questions based on their current answer.

How it works: - The adaptive attribute is set to true on the qti-assessment-item element - For adaptive items, outcome variables retain their values across multiple attempts and are only updated by subsequent response processing, whereas for non-adaptive items each attempt is independent - Feedback blocks (qti-feedback-block) control what content is shown based on outcome variable values - The qti-end-attempt-interaction allows candidates to submit attempts and see feedback

Example structure:

<qti-assessment-item adaptive="true" ...>
    <qti-outcome-declaration identifier="FEEDBACK" .../>
    <qti-outcome-declaration identifier="SCORE" .../>

    <qti-item-body>
        <p>Initial question...</p>
        <qti-choice-interaction .../>

        <!-- Feedback shown conditionally -->
        <qti-feedback-block identifier="HINT1" 
            outcome-identifier="FEEDBACK" 
            show-hide="show">
            <p>Here's a hint...</p>
            <!-- Can contain additional interactions! -->
            <qti-choice-interaction .../>
        </qti-feedback-block>
    </qti-item-body>

    <qti-response-processing>
        <!-- Sets FEEDBACK variable based on response -->
    </qti-response-processing>
</qti-assessment-item>

2. Test-Level Branching (Between-Item Routing)

QTI supports limited adaptivity through the use of pre-conditions and branching, allowing items in the predetermined sequence to be conditionally skipped or branched over during delivery based on context and prior responses.

Branch Rules (qti-branch-rule): - Defined at the section or test level - The branchRule element contains a rule evaluated to determine which item or section should be presented next based on conditions - Can jump to specific items based on outcome variables from previous items

Pre-Conditions (qti-pre-condition): - The preCondition element sets the conditions that need to be met for an assessmentItem or assessmentSection to be displayed - In nonlinear mode, pre-conditions are ignored

Example test structure with branching:

<qti-assessment-test>
    <qti-test-part navigation-mode="linear">
        <qti-assessment-section>
            <qti-assessment-item-ref identifier="Q1" href="question1.xml"/>

            <!-- Branch rule after Q1 -->
            <qti-branch-rule target="Q3">
                <qti-base-value base-type="boolean">true</qti-base-value>
                <qti-variable identifier="Q1.SCORE"/>
                <qti-base-value base-type="float">1.0</qti-base-value>
                <qti-match/>
            </qti-branch-rule>

            <qti-assessment-item-ref identifier="Q2" href="question2.xml">
                <!-- Pre-condition: only show if Q1 score < 1.0 -->
                <qti-pre-condition>
                    <qti-lt>
                        <qti-variable identifier="Q1.SCORE"/>
                        <qti-base-value base-type="float">1.0</qti-base-value>
                    </qti-lt>
                </qti-pre-condition>
            </qti-assessment-item-ref>

            <qti-assessment-item-ref identifier="Q3" href="question3.xml"/>
        </qti-assessment-section>
    </qti-test-part>
</qti-assessment-test>

3. Modern Adaptive Testing (QTI 3.0+)

QTI 3.0 introduces the qti-adaptive-selection element which delegates item selection and ordering to external CAT (Computer Adaptive Testing) engines during delivery, rather than using static selection with preconditions and branch rules.

Key Differences

Feature	Adaptive Items	Test Branching
Scope	Within one item	Between items
Storage	Outcome variables in item	Test-level outcome processing
Use case	Progressive hints, multi-step problems	Remediation paths, difficulty adjustment
Linearity	Multiple attempts on same item	Different items in sequence

The correct/incorrect responses themselves are stored in response variables, while the branching logic uses outcome variables (like SCORE or custom feedback identifiers) to determine the next step.