Matrix Training & Applied Behavior Analysis
Articles,  Blog

Matrix Training & Applied Behavior Analysis


(whooshing noise) – Hi, my name is Brianna Snyder and I’m a BCBA with Brett
Dinovi & Associates. In this video today we’re going to talk about matrix training. When you think of a matrix, what is your first thought? Do you think Keanu Reeves in the Matrix or do you think matrix training? If you’re watching this,
fortunately for you, as long as you’re here
for the AVA knowledge, we’re talking about matrix training. This video will cover fifth edition task list
topics G-11 and G-21. Matrix training is a
generative teaching approach in which words are arranged in a matrix so that some phrases are taught and others emerge without direct teaching. Research has shown matrix training to be effective in generating
emergent multi-word tasks and can be used to teach listener skills. Other examples in which
matrix training can be utilized include teaching learners how to spell and engage
in a socio-dramatic play. Matrix training can be used to teach more than just verbal behavior,
but for the purpose of this video we will be
mainly focusing on language. In order to understand matrix training, it is important to touch
on generative learning and recombinative generalization. Generative learning occurs
when previously acquired skills accelerate the acquisition of other skills without direct teaching or
prior history or reinforcement. An example of generative
learning is behavioral cusps. Behavioral cusps are
any change in behavior that allows a learner to
contact new contingencies. A behavioral cusp exposes the learner to new environments, new
reinforcers and punishers, new responses, and new contingencies. For example, when a child begins to emit echoic, tact,
and listener behaviors, new mands and tacts can be acquired without direct teaching or
prior reinforcement history. When new combinations of
verbal behavior emerge, the child is able to
access new contingencies in the environment. When a learner has the ability to successfully utilize these behaviors they will be able to engage
in social interactions with peers, thus
contacting new reinforcers. Recombinative generalization is defined as the differential responding
to novel combinations of stimulus components that
have been included previously in other stimulus contexts. The components of trained
combinations are arranged in new combinations based on
environmental requirements. For example, a learner previously learned to respond, “Green triangle,” in the presence of a green triangle, and, “Orange square,” in the
presence of an orang square. Without any prior learning history, when presented with an orange triangle, the child was able to
respond, “Orange triangle.” If a child is able to accurately respond to orange triangle without
a prior learning history, recombinative generalization has occurred. Another example would
be using matrix training to teach prepositions expressively. A learner previously
learned to place a fork to the left of a plate
when instructed to do so, and a knife to the right of the plate when instructed to do so. Recombinative generalization has occurred if the learner is able to place a fork to the right of the plate without any prior learning history. When targets are arranged in a matrix and you teach along the
diagonal of the matrix skills are acquired
without direct teaching thus this makes for an
efficient teaching strategy especially given that
most of us clinicians are pressed for time. This is not to say that
you should cut corners by any means whatsoever. Training down the diagonal
of the matrix allows for exposure to all stimuli
the learner will encounter in generalization probes. This is just a perfect
example of an effective and efficient way of
teaching multi-word phrases without having to teach every
single phrase separately. Another example can be found in Ax and Sainato’s article
regarding matrix training. They taught preschoolers with
autism to follow instructions to perform action-picture combinations. There were six actions on one axis and six pictures on the other axis. The instructions along
the diagonal of the matrix were trained, and instructions outside of the diagonal were
probed for generalization. This study demonstrated
that matrix training is an efficient and effective approach when teaching language
and literacy skills. Matrices can come in all sizes as long as two concepts are being taught. For example, a two-by-two
matrix can be used to teach colors and shapes. On one axis there will be two colors. On the other axis there
will be two shapes. So, in total there will be
four color-shape combinations. If two out of four
combinations are trained the other two should emerge
without direct teaching. When setting up a matrix training program you must first determine what
it is you’d like to teach. Pick something that is
individualized to the learner and is a functional skill. Once you determine what it is you’re going to teach, decide how big the matrix is going to be. It must be at least a two-by-two matrix meaning there should be
at least two variations of the phrase being taught
on each axis of the matrix. Each axis of the matrix
should be one category. For example, one axis is strictly colors while the other axis strictly shapes. Once the matrix is setup,
initially probe all combinations. When conducting the probes, you should provide reinforcement
for correct answers. If the learner is incorrect, do not provide corrective feedback. You can continue to probe
the other combinations or if there is nothing left to probe move on to something
else within the session. If the learner is continuing to get the responses incorrect, you can still provide
reinforcement for sitting at the table with you,
attending to the stimuli, or for working in general. If the learner responded correctly to a combination during baseline and it was on the diagonal of the matrix you do not have to explicitly
teach this response but it is important to
probe the combinations often when conducting maintenance checks. Now that you have your baseline data teach the first two combinations on the diagonal of the matrix. Teach the responses one at a
time until they are mastered. Once the two combinations are taught and the learner has mastered them, probe for generalization by testing the other combinations in the sub-matrix surrounding the diagonal. If the learner did not
generalize, you will need to go back and directly
teach the combinations that were not generalized. As the matrices get
bigger, it is important to probe more frequently. For every two responses taught from the diagonal of the
matrix, you should probe the responses in the
sub-matrix surrounding it. As always you should continue to do monthly maintenance checks to ensure this skill is maintaining across time. This is a great teaching approach because learners are able
to emit untrained responses that were not previously taught. This is a time saver for all of us who can never seem to
find enough time in a day. Not only is it efficient,
but research demonstrates that this is an effective teaching method. It should be noted that not every learner is going to benefit from
this teaching method. If there are challenging behaviors that are impeding learning,
I would not recommend starting matrix training. Your learner should have
basic attending skills have the ability to stay on task during discrete trial training,
and be able to generalize. If you have a learner who
does not generalize easily this may not be the
teaching method for them. Thank you so much for watching the video. If you’d like to help us disseminate the science, please click
the link below and subscribe. (whooshing noise)

Leave a Reply

Your email address will not be published. Required fields are marked *