Match-to-sample (MTS) is one of the most studied and reliable procedures in Applied Behavior Analysis. At its core, the task is simple: a child is shown a sample stimulus and asked to select the matching item from a set of comparison stimuli. But behind that simplicity lies a powerful mechanism for teaching concept formation, language, and relational thinking.
The Basic Procedure
In a standard MTS trial, an instructor presents a sample — say, an image of a dog — and places two or three comparison images in front of the child. One comparison matches the sample; the others do not. The child’s task is to select the correct match.
When the child selects correctly, reinforcement follows. When they select incorrectly, no reinforcement is delivered and the trial is repeated. Over many trials across many concepts, children learn not just the specific stimuli presented, but the relationship between stimuli — and that relationship transfers to new items they have never seen before.
Why MTS Is Foundational
MTS works because it trains discrimination without requiring language. A child does not need to know the word “dog” to learn that a picture of a dog matches another picture of a dog. This makes the procedure accessible to children who are pre-verbal or have limited verbal repertoires — precisely the learners who benefit most from structured, adaptive teaching.
Research has consistently shown that MTS training accelerates vocabulary acquisition, supports the development of symbolic thinking, and lays the groundwork for reading and language comprehension. Sidman’s landmark work on stimulus equivalence demonstrated that children trained on MTS relations spontaneously develop untrained relations — a finding that has shaped decades of ABA research and practice.
The Four Levels of MTS
Not all match-to-sample tasks are equal. INTERLAZA organizes MTS training across four progressively more challenging levels:
-
Identity matching (Level 1): The sample and the correct comparison are identical images. This is the simplest form — the child simply finds the picture that looks the same.
-
Non-identical matching (Level 2): The sample and the correct comparison show different instances of the same concept. A photograph of a specific dog matches a pictogram of a different dog. This level trains the abstract concept rather than the specific image.
-
Symbolic matching (Level 3): The sample is a word and the comparison is an image, or vice versa. This level directly trains the connection between language and meaning — the foundation of reading and vocabulary.
-
Auditory matching (Level 4): The sample is a spoken word or sound, and the child selects the corresponding image. This level trains auditory comprehension and is particularly important for children developing their listening vocabulary.
Each level builds on the one before. A child who has mastered Level 2 for a concept has a much stronger foundation for Level 3 than one who has only practiced identity matching.
From Simple Matching to Relational Learning
MTS is not just a rote memorization procedure. When trained systematically across many concepts, children begin to generalize — they apply what they have learned to new instances they have never encountered. This is the hallmark of true concept learning rather than stimulus-specific responding.
INTERLAZA extends MTS into equivalence training and relational frame theory (RFT) exercises, which take the trained relations and probe whether the child has spontaneously derived the untrained relations. A child taught that a picture of a dog matches the word “dog” should also match the word “dog” to a picture of a different dog — without ever being directly taught that pairing. When this happens, it demonstrates genuine symbolic understanding.
How INTERLAZA Implements MTS Digitally
Implementing MTS on a screen introduces challenges that do not exist at a tabletop. INTERLAZA addresses them with several design choices:
-
Errorless learning: Distractors fade in gradually from low opacity to full opacity, giving children an initial period of guaranteed success before the task reaches full difficulty. This reduces frustration and builds momentum.
-
Bayesian Knowledge Tracing (BKT): The app tracks each child’s probability of mastery for every concept in real time. As that probability rises, the number of distractors increases automatically. As it falls, distractors are removed. Difficulty adjusts to each learner without instructor intervention.
-
Adaptive prompt fading: For children who need additional support, the app can highlight the correct answer, display an arrow, or use size differences as prompts. These supports fade systematically as the child demonstrates competence.
Match-to-sample training has a 50-year evidence base. INTERLAZA’s role is to bring that evidence base to a format that instructors can use efficiently and children can engage with joyfully — one trial at a time.