Beware of the Straight and Narrow: What You Need to Know About Non-Linear Contingency Analysis
Part 1
A while back, I talked about how collecting ABC data and simply building an intervention off of it might not be enough.
Well, for months I’ve been wanting to return to the subject.
For those of you who don’t know what it is, non-linear contingency analysis can be a game changer.
At least it’s fun to nerd out about.
Non-linear contingency analysis is an advanced approach in ABA that aims to understand the complex interactions between multiple variables that contribute to, and maintain, a learner’s behavior. It recognizes that relationships between variables and behavior can be interconnected and complex…aka non-linear.
In contrast, traditional contingency analysis—the kind that you an I have been brought up on—primarily focuses on identifying the immediate antecedents and consequences of a behavior, assuming a linear relationship between these two variables and behavior. This traditional analysis simplifies the behavior-environment relationship by examining only one (or a couple) variables at a time. As such, BCBAs build entire behavior plans around the one variable that they think might be impacting the problem behavior.
Stunningly tedious.
For part one, I’ll unpack the main points and differences between the two different approaches which goes as follows:
Non-linear contingency analysis factors in how complex problem behavior can be.
Sometimes there’s just more to problem behavior than we originally anticipated. This approach acknowledges that there are complex interactions between individual, environmental, and social variables, while traditional contingency analysis simplifies or avoids these relationships.
For example, let’s say you have a kiddo who has meltdowns during family outings. The traditional approach identifies the specific triggers and consequences, like going to the park, and designs interventions accordingly (we go to the park more). The non-linear approach acknowledges the complex interplay of factors and it digs a little deeper, such as the child's sensory sensitivities (Too much sunlight?), family dynamics (Loud siblings?) and outing context (Is it cold outside? Have we ever been reinforced at the park?). Interventions may involve addressing multiple aspects, like sensory strategies, teaching siblings to provide support, and gradually increasing exposure to different outing contexts.
Non-linear contingency analysis acknowledges that the relationship between two variables might not always be proportionate.
So what does that mean?
Well, let’s pretend a kiddo experiences severe anxiety in the presence of a large group of peers in a classroom setting. The traditional approach assumes a linear relationship between anxiety level and the number of peers, predicting a proportional increase in anxiety the more peers there are. Interventions may involve gradual exposure to more and more peers. It basically says lots of peers = lots of tantrums. The non-linear approach, however, recognizes that the relationship between anxiety and audience size might not be that proportional; the learner might experience similar anxiety levels in the presence of only one peer—and maybe a familiar peer at that. You see, the nonlinear approach digs a little deeper. As such, your intervention might focus on social skills development, self-soothing behaviors, and re-framing thoughts related to these situations.
Non-linear contingency analysis acknowledges that all of these variables are playing on and off each other, all of the time.
They are interconnected, interwoven, interdependent, inter…well, you see where I’m going. The non-linear approach considers the interconnected nature of various factors contributing to behavior, while traditional analysis typically examines one variable at a time.
Let’s say you have a learner that struggles with eating certain foods. The traditional approach identifies one variable—or one food—that controls the behavior, like broccoli. This then suggests that the intervention should fade the learner to lesser preferred foods. On the other hand, the non-linear approach considers the interconnectedness of many variables. Variables like: your learner’s ability to sit in a chair (and for how long), to complete a task or preferred activity in a seated position, to use tools to gain access to a desired outcome (pencil, paintbrush, fork, etc.), in addition to considering the food item itself as a variable. As such, you are choosing to look at multiple variables simultaneously, and building an intervention off of this multivariate assessment.
In conclusion, non-linear contingency analysis does a little more than ABC data. In the simplest terms, it demands that we look at more than just one variable when analyzing behavior. We know this to be true.
Here’s a thought exercise:
What makes a bad day a “bad day”?
Sometimes a “bad day” comes to us because one bad thing happens, right? But other times, a “bad day” comes about as a result of numerous small things that don’t go right either. In other words, a “bad day” transpires as a result of multiple variables playing off each other in a short amount of time. It’s not the flat tire that makes it all bad. It’s the fact that you might have gotten a flat tire…and you’re coming down with a head cold…and the weather has taken a turn for the worst…and you don’t have a jacket.
So, next time a problem behavior pops up. Take a deep breath and consider the factors (plural) that might be contributing to it!
Stay tuned for more examples next week!
In the mean time, some resources for you:
Our ABA skill database, bxmastery.com, is deep in the throes of a beta test. Check out bxmastery.com and click here to see how you can access hundreds (going on thousands) of skill ideas for free for a limited time.
Bailey and Burch’s SECOND addition of 25 Essential Skills for the Successful Behavior Analyst is officially out. I’m incredibly gracious to the authors for having reached out and given me the privilege to review and endorse the text with Pat Friman. Check it out here. Highest recommendation!
Finally, regarding non-linear contingency analysis: Read this.
Have a clinical question and need it answered? Join our exclusive Facebook Group!