Graphing in Applied Behavior Analysis is how we determine if the treatment being provided to your child is showing a positive change. Graphs also allow us to adjust our interventions unique to the client to provide the best care applicable. When asked for clarification on reliability and validity we can refer to these graphs and feel confident when explaining the reason for our interventions. In this field we take pride in making sure that everything we do is based on reliable and valid empirical research.
Parents Reading Research
When a parent has come to learn their child has been diagnosed with Autism Spectrum Disorder (ASD) or even find their child engaging in a new problem behavior, it is likely they are filled with an abundance of questions. When questions aren’t answered sufficiently or are not answered in a timely manner it is likely that the parent will rely on books. This can be overwhelming for parents, who are buried in research full of numbers and graphs that are incomprehensible. It is important that these graphs are clear to understand for any reader. It can become frustrating when the parent is attempting to find what will best aid their child, but they are unable to find an answer due to faulty or confusing data. As you keep reading, I will point out some key components of graphs that are essential and explain why.
Graphing in an Applied Setting
When working in an in-home setting, the Registered Behavior Technician (RBT) is consistently taking data on the programs given unique to your child’s based on their goals and behaviors. The RBT is trained prior starting the job on how to correctly input data into their system or independently graph the data by hand. This data is then provided on a graph for the Board-Certified Behavior Analyst (BCBA) to examine. If paper data is being utilized, the RBT should spend the last minutes of session correctly drawing in data points and figuring percentages. The BCBA will analyze and adjust interventions when they supervise the RBT or meet with the family.
There are rules for graphing that every person creating and examining graphs must know. There are general rules that everyone learns in school, such as the graph must have an X and Y axis and be labeled. However, there are a few rules that unfortunately are not practiced by many, which are vital details needed for the graph to be displayed correctly. The components I am going to focus on are: making sure the data points are legible for any reader, the use of the proportional construction rule, using an appropriate unit of time, and subjective VS objective language.
The Importance of Data Points
There are many issues with the graph above but here we are going to mainly focus on data points. In this graph we see a positive trend, this is good, that means one of our goals is increasing! Yay! But wait. How do I know how fast the client reached this success? What did their progress look like? Here is where data points come in handy. Let me show you the difference it makes.
I have personally added my own data points to the graph. With this visual I can recognize my client started increasing at a slower rate and later moved up to a more consistent rate. Data points allow the BCBA to recognize possible errors in intervention and can compare graphs across multiple providers to ensure consistency and that our client is effectively generalizing the material appropriately.
Proportional Construction Rule
Now we have a complete graph! Right? Not quite. Now we need to make sure the graph is now proportional. Graphing rules state every graph must 2/3 or 3/ 4 proportional (Cooper, Heron, and Heward 2007) Why is this important? When this rule is not utilized it allows for individuals to either attempt to exaggerate or depress their data. Obviously, that is not ideal if we want to provide the best care for our clients. I have provided visuals below.
The graph above is a correct proportional graph. I have added black lines representing the two-thirds and three-fourths proportion. If you examine the graphs below, I extended the X-axis in one and the Y-axis in the other. As you can see when the x-axis is extended it depresses the data and when the y-axis is extended it exaggerates the data. The creation of these false graphs can allow the researcher to conclude that their study’s treatment data represents a higher significance or mistakenly, examine no significance than the actual results.
Units of Time
In the next graph below, we are going to focus on the term “sessions”. The National Institute of Standards and Technology (2020) explained that units of time are seconds, minutes, hours, days, weeks, years, centuries, etc. Now, let’s examine this graph. Are you able to tell me how long it took for the researchers to collect this data? The answer is no. When looking at this graph I wonder, was it nine 5-minute sessions in one day? Was it one session a week for nine weeks? Was it one session per day for nine days? We all deserve to know! If we were to use this data to incorporate into our interventions, we want to be able to hypothesize when the change in behavior from our intervention should arise.
To think about it another way, imagine your child is engaging in self-injurious behaviors. How quickly do you want the intervention to decrease the behavior? I think you and I can both agree on, as soon as possible. Based on the labels in the graph above we cannot be certain that using the techniques from this study will minimize the behaviors as quickly as we would like due to no clear unit of time represented.
Subjective VS Objective Language
While you are reading these articles, you may come across some subjective language the researchers use to help explain their graphs and results. An example of this is rapidly VS moderately VS gradually increasing or decreasing trends. There is currently no research out there expressing the criteria for what is considered rapid, moderate, or gradual. Therefore, it is subjective. Trends are only represented objectively in four ways: increasing, decreasing, or maintaining also known as zero.
Data points, proportion construction of graphs, unit of time, and objective language are all essential components for a graph to have in order to fully understand the data being provided. Keep these in mind when reading research articles. Some studies may state that their data represents as significant, but you can always double check by looking at how they administered their graph.
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior analysis (2nd ed.). Hoboken, NJ: Pearson Education, Inc.
Graphing and Applied Behavior Analysis . (n.d.). Retrieved fromhttps://institute.centralreach.com/learn/course/graphing-and-applied-behavior-analysis/welcome/welcome-to-graphing-and-applied-behavior-analysis
National Institute of Standards and Technology. (2020, March 31). Unit and Systems of Measurement. Retrieved April 6, 2020, from https://nist.gov/