Interbehavioral Psychology

Competency Transfer Model

Varela & Quintana (1995)

A systematic taxonomy for classifying and measuring how far a person can transfer what they have learned to new situations. Rooted in Kantor's interbehavioral field theory, this model provides a precise framework for evaluating the scope and limits of competency transfer — independent of any specific teaching procedure.

Important distinction

The Competency Transfer Model is not a Match-to-Sample (MTS) model. It is a general taxonomy about competency transfer that applies across any learning context. Varela used MTS as an evaluation procedure to measure transfer — the taxonomy itself describes what changes between training and testing, not how training is delivered.

Core Concepts

The four factors

Every learning situation can be described along four independent factors. When moving from a training situation to a transfer test, each factor is either held Constant (K) or made Variable (Var). The combination of K and Var across these factors determines the transfer level.

Instance

Most concrete

The specific stimulus objects used in the task. If training uses a red circle and a blue square, those are the instances. When Instance is Variable, the transfer test uses entirely different objects (e.g., a green triangle and a yellow star) while everything else stays the same.

Modality

The sensory and perceptual properties of stimuli — color, form, size, spatial arrangement, or sensory channel (visual, auditory, tactile). When Modality is Variable, the transfer test changes these properties: training with color-based stimuli might test with size-based stimuli, or visual stimuli might switch to auditory.

Relation

The matching criterion that defines what counts as a correct response: identity (=), similarity (≈), or difference (≠). Relation is subordinate to Dimension — it operates within the domain that Dimension defines. When Relation is Variable, the matching rule changes: training on identity matching might test similarity or difference matching.

Dimension

Most abstract

The domain of functionality that defines what property is relevant for the task: geometric (shape, form), semantic (meaning, category), numeric (quantity), chromatic (color), or spatial. Dimension is the most abstract factor. When Dimension is Variable, the entire domain changes: training in the geometric domain might test in the semantic domain.

Constant (K)

The factor remains the same between training and the transfer test. No change on this dimension.

Variable (Var)

The factor changes between training and the transfer test. The learner must generalize across this change.

Procedure

How the evaluation works

The model follows a structured sequence: establish a baseline through training, then systematically vary factors to measure how far transfer extends.

Define the training situation

Fix all four factors: choose specific instances, a sensory modality, a matching relation (identity, similarity, or difference), and a functional dimension (geometric, semantic, etc.). This fully specifies the baseline task.

Train to mastery criterion

The learner practices with corrective feedback until reaching a stable performance criterion, typically ≥90% accuracy across two consecutive sessions. This ensures the baseline skill is solidly established before transfer is tested.

Apply the transfer test

Present new trials where one or more factors have changed — without corrective feedback. This is where Varela's classification enters: which factors are Constant and which are Variable determines the transfer level being tested. The absence of feedback is critical — you are measuring what the learner already knows, not teaching new skills.

Interpret results

Compare transfer performance against the training baseline. High performance on a transfer level means the learner has generalized across those changing factors. Low performance identifies exactly which factor combinations are difficult — giving the instructor a precise map of where to focus intervention.

Classification

The 15-level transfer matrix

By combining the four factors as Constant or Variable, Varela and Quintana identified 15 distinct transfer levels. There is no level where all four factors remain constant — if nothing changes, there is no transfer. The levels are ordered from one variable factor (easiest) to all four variable (hardest).

Level	Name	Instance	Modality	Relation	Dimension	Factors changed
1	Extra-Instance	V	K	K	K	1
2	Extra-Modal	K	V	K	K	1
3	Extra-Relational	K	K	V	K	1
4	Extra-Dimensional	K	K	K	V	1
Two-factor combinations
5	Extra-Instance-Modal	V	V	K	K	2
6	Extra-Instance-Relational	V	K	V	K	2
7	Extra-Instance-Dimensional	V	K	K	V	2
8	Extra-Modal-Relational	K	V	V	K	2
9	Extra-Modal-Dimensional	K	V	K	V	2
10	Extra-Relational-Dimensional	K	K	V	V	2
Three-factor combinations
11	Extra-Instance-Modal-Relational	V	V	V	K	3
12	Extra-Instance-Modal-Dimensional	V	V	K	V	3
13	Extra-Instance-Relational-Dimensional	V	K	V	V	3
14	Extra-Modal-Relational-Dimensional	K	V	V	V	3
All four factors
15	Extra-Total	V	V	V	V	4

Reading the table: Each row is a transfer test configuration. K means the factor stays the same as in training. V means the factor changes. Level 15 (Extra-Total) is the most demanding: everything changes between training and test.

Application

Clinical use

Research following Varela's original taxonomy has produced several findings that guide how to sequence and optimize transfer testing in clinical practice.

Pre-training with variability is the most determinant factor

Varela and Barrón (2006) demonstrated that exposing learners to varied examples during the training phase — before any transfer test — is the single most powerful predictor of successful transfer. When training includes diverse instances and modalities, learners are significantly more likely to succeed on higher transfer levels.

Ascending order of tests facilitates performance

Tena, Hickman, and Moreno (2007) found that presenting transfer tests in ascending order (from Level 1 upward) produces better results than random or descending order. Each successful transfer builds confidence and behavioral momentum for the next level.

Relation order matters: Identity → Similarity → Difference

Within levels that vary the Relation factor, the progression from identity matching to similarity matching to difference matching follows a natural difficulty gradient. Identity is the easiest relation, difference the most demanding.

Recommended evaluation protocol

Base training — Establish mastery on a fixed training situation with corrective feedback.

Varied pre-training — Expose the learner to multiple instances, modalities, and examples within the training domain.

Transfer tests — Apply levels in ascending order (L1 through L15), without feedback.

Profile analysis — Map which levels the learner passes and fails to identify the transfer ceiling.

Focused intervention — Target the specific factor combinations where transfer breaks down.

Progression — Re-test to measure growth, advancing through the 15 levels as competency develops.

Bibliography

References

Kantor, J. R. (1924). Principles of Psychology (Vol. I). Bloomington, IN: Principia Press.

Ribes, E., & López, F. (1985). Teoría de la conducta: Un análisis de campo y paramétrico. México: Trillas.

Varela, J., & Quintana, C. (1995). Comportamiento inteligente y su transferencia. Revista Mexicana de Análisis de la Conducta, 21(1), 47–66.

Varela, J., & Barrón, A. (2006). Efectos del entrenamiento variado sobre la transferencia en una tarea de discriminación condicional. Acta Comportamentalia, 14(2), 129–152.

Tena, O., Hickman, H., & Moreno, D. (2007). Efectos del orden de presentación de pruebas de transferencia sobre la ejecución en tareas de igualación de la muestra. Acta Comportamentalia, 15(2), 159–179.

Try it in INTERLAZA

INTERLAZA's Varela Transfer Module turns this 15-level taxonomy into a clinical tool. Select which factors vary, and the engine generates the correct trial structure automatically.

Open INTERLAZA