Identifying (Chapters 4 and 5) and selecting (Chapter 6) determinants and sub-determinants and identifying (Chapter 7) and selecting (Chapter 8) behavior change principles (BCPs) usually means that an overwhelming amount of information has been collected and processed, and many decisions have been made. Therefore, it can often happen that one feels a bit overwhelmed and loses overview. At this point, it can be useful to tie everything together in one visualisation.
Fortunately, a tool exists to do exactly that: the acyclic behavior change diagram (ABCD). ABCDs consist of two components. The diagram itself is one component: a visual representation of the most important structural and causal assumptions underlying an intervention. This diagram is generated based on the other component: the ABCD matrix. This matrix is just a table with seven columns, where each row represents one causal-structural chain (see Section 3.2). As such, the seven columns of the ABCD matrix correspond to the seven links of the causal-structural chain.
From left to right, the ABCD matrix columns are:
- Behavior Change Principles (e.g. methods of behavior change or behavior change techniques, BCTs);
- Conditions for effectiveness (conditions that must be satisfied for the BCP to successfully engage underlying evolutionary learning principles, ELPs);
- Applications (concrete, more or less tangible intervention products that implement the BCPs);
- Sub-determinants (specific aspects of the human psychology that are targeted by the applications);
- Determinants (overarching constructs of similar of functionally similar sub-determinants);
- Sub-behaviors (specific behaviors, each predicted by different sub-determinants);
- Target behavior The ultimate target behavior.
Because every row represents one causal-structural chain, a full ABCD matrix contains many of the important assumptions underlying an intervention. Because the matrix has a standardized format, it can easily be shared and read into a variety of programs. In addition to helping intervention developers to get their assumptions clear, and preventing them from forgetting things, ABCD matrices remove the need for coding a lot of the intervention content and logic. In fact, the ABCD matrix clearly shows
The matrix format allows editing the ABCD matrix in spreadsheet software, enabling use of real-time collaboration suites such as Collabora, Google Sheets or Office 365. In addition, the ABCD matrix is the input for the acyclic behavior change diagram itself.
Although this ABCD matrix has the benefit of being machine-readable and easy to edit, dozens of sub-determinants are often targeted. Getting an overview of the entire underlying logic model can therefore be hard. In addition, once familiar with the ABCD matrix, they are a useful tool, but intervention developers will often need to communicate with other parties, such as the producers of the intervention such as advertising agencies, who will not be familiar with behavior change and intervention development tools.
Acyclic behavior change diagrams address this. They are standardized diagrams that are generated from an ABCD matrix. Although ABCDs are not machine readable, they are more human-readable than ABCD matrices. ABCDs visually represent the seven columns of the ABCD matrix, but with cells with the same content merged to be represented by the same element. In other words, assuming the final column has the same target behavior specified, only one element will appear representing the target behavior. The only exception is formed by the determinants, which are only merged within the corresponding sub-behaviors. After all, although the determinants that determine two sub-behaviors may have the same name, they represent different things in reality.
By visually merging duplicated elements of the structural-causal chains in the ABCD matrix, it is easier to get an overview of the logic model represented in the ABCD matrix (and underlying the intervention). Although ABCDs can get rather big, they’re pretty much the simplest overview of why an intervention will work (or not). As such, ABCDs are very useful when communicating with colleagues, members of an intervention planning group, or other parties, such as the executive intervention producers (e.g. advertising agencies).
For example, imagine that an intervention developer is developing an intervention for target behavior Following ecstasy dosing recommendations, and they distinguished sub-behaviors ‘Decide to follow the dosing recommendations’ & ‘In advance, with the groups of friends, discuss everybody’s planned dose.’. Their determinant studies yielded a determinant structure, and based on the CIBER plots, they selected sub-determinants: ‘If I use a high dose of ecstasy, I will feel less connected to others.’, ‘If I use a high dose of ecstasy, I will feel more isolated.’, ‘If I use a high dose of ecstasy, I will remember less’, ‘Most people approve of avoiding a high dose of ecstasy.’ & ‘I can explain why I want to follow the dosing recommendations.’, falling under determinants ‘Attitude’, ‘Perceived norm’ & ‘Perceived behavioral control’.
Based on this information, they selected the behavior change principles ‘Persuasive communication’, ‘Information about others’ approval’ & ‘Modeling (vicarious learning)’. They intend to implement these in applications ‘An infographic shows how the effects of ecstasy change as the dose increases.’, ‘Show the Party Panel result that illustrates that most people want to dose relatively low (compared to the strength of available ecstasy pills).’ & ‘A comic with examples of how to discuss the dose you plan to take.’, and in doing so, will strive to satisfy the following conditions for effectiveness that they identified: ‘Messages must be relevant and not deviate too much from existing beliefs; can be stimulated with surprise and repetition; contains arguments.’, ‘Others do indeed approve of the target behavior.’ & ‘The message recipient must identify with the model; the model has to be a coping model, struggling with the behavior, not a mastery model; the model must be positively reinforced.’.
Presented like this, obtaining an overview of this information is hard. In addition, many intervention developers do not report this information like this. If they would, extracting the exact behavior change principles they used, and the exact sub-determinants they target, would be relatively easy. Often, however, this information is not explicitly listed in articles or manuals.
The ABCD matrix offers a standardized way to include this information:
|Behavior Change Principles||Conditions for effectiveness||Applications||Sub-determinants||Determinants||Sub-behaviors||Target behavior|
|Persuasive communication||Messages must be relevant and not deviate too much from existing beliefs; can be stimulated with surprise and repetition; contains arguments.||An infographic shows how the effects of ecstasy change as the dose increases.||If I use a high dose of ecstasy, I will feel less connected to others.||Attitude||Decide to follow the dosing recommendations||Following ecstasy dosing recommendations|
|Persuasive communication||Messages must be relevant and not deviate too much from existing beliefs; can be stimulated with surprise and repetition; contains arguments.||An infographic shows how the effects of ecstasy change as the dose increases.||If I use a high dose of ecstasy, I will feel more isolated.||Attitude||Decide to follow the dosing recommendations||Following ecstasy dosing recommendations|
|Persuasive communication||Messages must be relevant and not deviate too much from existing beliefs; can be stimulated with surprise and repetition; contains arguments.||An infographic shows how the effects of ecstasy change as the dose increases.||If I use a high dose of ecstasy, I will remember less||Attitude||Decide to follow the dosing recommendations||Following ecstasy dosing recommendations|
|Information about others’ approval||Others do indeed approve of the target behavior.||Show the Party Panel result that illustrates that most people want to dose relatively low (compared to the strength of available ecstasy pills).||Most people approve of avoiding a high dose of ecstasy.||Perceived norm||Decide to follow the dosing recommendations||Following ecstasy dosing recommendations|
|Modeling (vicarious learning)||The message recipient must identify with the model; the model has to be a coping model, struggling with the behavior, not a mastery model; the model must be positively reinforced.||A comic with examples of how to discuss the dose you plan to take.||I can explain why I want to follow the dosing recommendations.||Perceived behavioral control||In advance, with the groups of friends, discuss everybody’s planned dose.||Following ecstasy dosing recommendations|
This matrix can be processed into an ABCD, producing this result:
Even though in this book, the font will likely be too small to easily read, you can already see that it is much easier to get an idea of the assumptions underlying this hypothetical mini-intervention. Apparently, the intervention targets two sub-behaviors and three determinants, using three behavior change principles in three applications. It is easy to see whether important determinants are omitted, or whether conditions for effectiveness were not taken into account.
In addition, the visualisation of the logic model makes it easier to communicate with, for example, members of an intervention planning group, such as stakeholders or target population members. It also facilitated supervision of the executive program producers, such as advertising agencies. Because they often lack knowledge about behavior change, instead having specialized in creative processes, an ABCD is a convenient tangible tool to make sure none of the targeted sub-determinants or conditions for effectiveness gets lost in translation.
You can create an ABCD using whichever spreadsheet editor you like. Because the ABCD is generated from the ABCD matrix, any application that lets you edit a table with seven columns can be used. There are at present three ways to produce an ABCD from an ABCD matrix. First, you can use jamovi; second, you can use R; and third, you can use an online app. Using jamovi is the most userfriendly, because it offers both the ability to edit the spreadsheet with changes immediately reflected in the diagram. However, as yet, it is not yet possible to save the generated diagram. We will also describe how to use R.
One of the datasets supplied with the
behaviorchange jamovi module (see Section 22.214.171.124) is an empty ABCD template. You can simply open it and adjust it right in jamovi’s spreadsheet editor. To order the diagram, open the Behavior Change menu that contains the
behaviorchange module analyses, and select “Acyclic Behavior Change Diagram (ABCD)” (see Figure ??).
In the dialog that appears, enter the seven columns of the ABCD matrix (see Figure ??).
As soon as all seven columns are specified, jamovi will generate the diagram (see Figure ??).
With every change, jamovi will automatically regenerate the diagram. This means that the ABCD matrix can now be edited to immediately see the changes reflected in the diagram. On the one hand, this provides a convenient method for working with the ABCD. However, at the time of writing this chapter, jamovi does not yet support wrapping the text in the spreadsheet editor, which makes editing longer texts inconvenient.
In addition, when collaborating with multiple people, it can be convenient to use an online spreadsheet that can be edited simultaneously by everybody. Multiple such services exist: the free/libre and open source Collabora is one; Google docs is another; and Microsoft Office 365 yet another one. The contents of such a spreadsheet can then simply be copy-pasted into jamovi.
It is still convenient, however, to first load the template, so that all variable/column names are already configured properly. Then simply select the cell in the first row and the first column and copy-paste the ABCD matrix, so that the cells from the template are overwritten.
To export an ABCD, right-click it.
To create an ABCD in R, the
abcd() function in the
behaviorchange package can be used. If the ABCD matrix has already been loaded, it is very simple. For example, imagine that the ABCD matrix has been loaded into a data.frame called
dat. In that case, the ABCD can be generated using this command:
To load a file into R, many methods exist. An easy one is to use the
getDat() function from the
ufs package. Because the
behaviorchange package also uses
ufs functions, the
ufs package is automatically installed if you install
behaviorchange, so you should have
ufs installed already. The
getDat() function opens a dialog that allows you to select a file, which
getDat() then opens, storing the data in a data.frame called
To save an ABCD to a file, an
outputFile can be specified. The ABCD is then stored to that file.3 For example, the following command saves the ABCD to a file called
abcd.png in the user’s home directory (e.g. the “Documents” directory on a Windows PC):
Other extensions can also be specified. For example, if the ABCD is exported to a
.svg file (a Scalable Vector Graphic), it can then be edited using for example Inkscape, and excellent Free/Libre and Open Source Software package.
Under the hood, the function
DiagrammeR::export_graph()is used. The
file_typeis extracted from the