Actor-Based Change Frameworks

What are they?

The Actor-Based Change Framework (ABC) is a participatory approach that was initially developed by Andrew Koleros and published in 2018.[1] It is an invaluable aid to developing a meaningful, evidence- and evidence-informed Useful Theory of Change (what is this?).[2] While the initial paper reflects on a fairly academic model of it, many of us have been applying and developing the idea beyond it (and hopefully into a practioner’s guide some day!). As soon as this is ready the references in this will be updated. In the meantime, I have been using it extensively for a range of purposes including: programme design, MEL system design, Theory of Change development, evaluation, and even diplomacy. Here are my reflections on it that I will continue to update as I play with the framework.

ABC weaves together systems-thinking, complexity science, and behaviour change theory to help you map out the informal systems that give rise to the problem that your programme is aiming to address. While often a Cold Hose over programme plans of perceptions of change, ABC ensures that you have actively understood the terrain within which your programme will operate and therefore accurately plot your hypothesis for change. It then becomes a powerful monitoring and evaluation tool that aids swift programme adaptation by holding a symbiotic relationship with your theory of change.

---

[1] Koleros, A. Et al, The Actor-Based Change Framework: A Pragmatic Approach to Developing Programme Theory for Interventions in Complex Systems, American Journal of Evaluation, 2018

[2] Mayne, J., Useful Theories of Change, Canadian Journal of Program Evaluation, 2015

How does one do it?

ABC is a powerful tool, but not a silver bullet. Nevertheless I have found it invaluable to my work, and my programme teams have as well.

ABC is all about behaviour and influence, not about formal hierarchies or existing formalised structures. It capitalises on contextual knowledge to construct informal systems that exist around – and give rise to – problems. It then supports you to solve part of that problem with the resources you have to hand.

ABC is best understood through a workshop so do reach out if you want to talk about that option, but in lieu of that, please find guidance and support below. Please make sure you’ve read my guidance on Theories of Change first.

Six Steps

Until more formal practitioner guidance comes out, please read the paper. However, you may find that this 6-step model I find makes ABC easier. These steps break ABC down into parts that make it feel more manageable. Please note that this all speaks to the Useful Theory of Change model. Then, once it feels coherent, below you will find templates for some pre-work exercises that I often use in trainings as well as the kinds of skeletons and checklists that I use as a trainer. However, I will emphasise that you do need an experienced facilitator for this, and for groups larger than 8 you will need two.

Step 1: Identify your problem.

This is a problem focused approach. It doesn’t always have to be, but when considering change it is sometimes easier to think of it through the lens of a problem. What is that current state you want to change? Identifying and analysing this helps us understand what the current state of play is.

My favourite way to do that is to define the problem and draw out the causes and consequences.

Sometimes causes are themselves consequences. Be aware of these situations and find the slight differences that occur in cyclical causal situations. If harmful gender norms are both causes and consequences, what does that mean? Maybe the cause is that women are seen as having fewer rights, and the consequence is that women receive even lesser status in their community. They’re different ends of the same field of harmful gender norms. Getting this right is crucial as the first steps to an accurate ABC-F hinges on this as this is how we set our system boundaries. You can represent a problem definition however you want: as a causal map, a logic tree, or in this PCC format that many of us use:

Step 2: Identify your actors associated with it.

Then, once defined, we think about the actors associated with it. an actor can be a particularly influential individual – such as a minister – or a group  – such as community faith leaders.

Defining these actor groups is really important. Often we fall into traps of defining beautiful homogenous groups such as ‘girls’ or ‘LGBTQI community’… But is this truly realistic? It is not. Resist the temptation of the easy. There is no such homogenous group. Instead, try to think through the lens of behaviour and influence in relation to the problem you have analysed. Use your evidence of how the populations you are considering really act. Perhaps ‘girls’ as a group does not exist, but ‘urban girls between 16 and 20 who use contraception’ might. That kind of a behavioural and geographical group might exist. With a skilled facilitator and the right research and/or evidence you can better pull out groups that mirror the scale, granularity, and nuance that is appropriate.

Step 3: build your AB systems map with influence lines.

The next step is to map out the actors we have identified in the right groups, and map how they are connected. This gives you an A-B SM! Crucially, this sets your system boundaries. By pre-defining the problem and actors, you have clear limits on the system you will be working with.

NOW. This can be really tricky as it’s easy to fall into mapping hierarchies or existing systems. That’s not what we are after. We want who influences who.

For example, in the Kyrgyz central government apparatur (GA), an ABC participant may want to draw an influence line from the head of the GA to the deputy head, and from the head to the senior managers. This is the formal reporting structure, sure, but not where the influence lies. Really the deputy head influences all of those stakeholders as he is protected from political whims and stays in position more safely. That’s where the power lies. THIS is what our systems map is after.

These relationships between our actors need not be restricted to influence, mind you. These can be defined in a manner most suitable to the context. If you have a lot of time and evidence, you can characterise these connections through multiple lenses: you can map when actors are connected by dynamics such as power, finances, information flow, religion and overlay to find the patterns. If resource poor, however, try to focus on the question of influence and define it for your context.

Ultimately, we map this to acknowledge the fact that behaviour is inherently systemic. That is to say, no one ever behaves in isolation. Teachers utilise corporal punishment in school party due to a lack of knowledge of other methods (‘capability’ for my COM-B aficionados out there) and a culture of acceptance (‘opportunity’), but also because they are responding to a behaviour from students that holds sway over them: a behaviour that triggers theirs, and is compounded by other COM actors. This is why we want to understand interrelationships between actors, as it helps us understand the systemic behaviour, and means we are more likely to understand how we can change it systemically as well (please see my Chemonics blog on how I used ABC to support doing exactly this as part of rapid research in Syria, work I since took to the UKES and EES conferences).

So… All of this mapping has a purpose. Hands up if you’ve been half-way through a programme and realised that you’re working with the wrong people? It’s a really common issue, and ABC helps you try to avoid that ahead of time if being used for programme design purposes.

Once we have mapped the connections, then we have the informal system. Our flock of birds. All we need is to understand how each is moving. What are their key behaviours contributing to, driving, or relating to the problem? Identifying that gives us an actor-based behavioural systems map.

I find it easiest to consider current state behaviours in three buckets: helpful, unhelpful, and neutral (recognising neutral behaviours can often, in their own way, be unhelpful). All of these should be identified in relation to that original problem – we all have many behaviours! We are just seeing the relevant one.

For example, if a problem statement is something relevant to the criminal justice system not fully functioning, an unhelpful behaviour of a police actor group could be ‘engaging in task based policing’ (as intelligence led policing may be more ideal). However a helpful behaviour from some criminal justice system actors might be ‘effectively prosecuting a small number of cases’. In this situation, this may be an actor group whose behaviour you want to scale or maximise. A neutral behaviour could be, for example, in an actor group within the public such as ‘ignoring the news about cases’. This could matter little and therefore be neutral (but it could also be problematic depending on your programme!)

If short on time, no need to do this across the whole system. Identify your entry points (step 4) and map only their behaviours.

Step 4: find a meaningful in. If you can’t, reconsider your intervention or find a friend.

To change a system, you need a good lever to pull. In systems-speak these can be called nodes or hubs, but I’ll call them ‘ins’. This is a useful step for ABC for programme design, but I have also found it valuable to assess intended entry points as part of evaluation.

A good ‘in’ is an actor group that has good causal potential through the informal system you have just identified. That is to say, if your ‘in’ starts behaving differently, will they bring others with them? If not… time to bring out the cold hose (ABC is a great cold hose) and reconsider your ‘in’ or your intervention altogether. Alternatively, see if you can buddy up with another implementer in the system and combine your ‘in’s to achieve a more catalytic change.

There are more considerations than just causal potential, however. Consider your reach: can the programme reach the actors with good causal potential? If not, can you work with someone who does and build reach? If not… Cold Hose time.

Relatedly, is your programme trying to operate in a crowded do not or implementation space? If so, it might be wise to consider this in your entry point analysis. There is no point trying to pull a lever that seven other people and their cat are trying to pull.

Step 5: identify the problem behaviours, and do a change frame for the problem behaviour of your ‘in’, and identify which you can affect. If it’s out of scope, reconsider your intervention or find a friend.

Now, I have talked a lot about systems maps and potential for change, but less on the hows of change. Once you’ve established a decent ‘in’ in terms of affecting this system, ABC has a straight forward approach to unpacking the behaviour change elements of this in such a way that it translates into a theory of change.

You will have identified the current state behaviour of that actor group that contributes to the problem. You can then proceed to use a handy COM-B model to unpack how this might turn into a ‘future state’ behaviour that is more desirable. You do not have to use COM-B, but it is useful.

For example, moving from a police force with the behaviour of ‘ineffective and poorly coordinated service provision’ to ‘effective, targeted, and coordinated service provision’

Once you have your COM-B model… time for the cold hose again. As a team, you have to assess which changes are in- or out-of-scope. If the behaviour change is not possible for your intervention to bring about, you simply will not see that catalytic change. Either reconsider the intervention, change the intervention, or buddy up with another implementer to bring about the change.

For example: you may be able to provide capacity building to the police force but cannot fund technology that would make coordination easier. If, however, another actor could fund the technology… then collectively you can bring about the behaviour change. I have done this when using ABC to help two programmes coordinate and have seen powerful results.

Do this step for every single ‘in’ you have in the system, and you will be able to build a series of mini ‘theories of action’ for each actor group: what the behaviour change will be, the COM change to bring it about, and the reach & reaction is automatically filled in. The next steps are just to consider what your activities will be and what goods and services they will produce to bring about these COM and behaviour changes.

Step 6: align this work to your ToC, and add all relevant assumptions.

These theories of action will start to stack up: you have a string of activities, goods and services, reaches and reactions, COM changes and behaviour changes. It starts to become clear how this mirrors your Useful ToC. The next stages are to connect the behaviour changes with Direct Benefits and a Sustainable Positive Change… well.

We started with a problem, right? The inverse of that will be your sustainable positive change. All that is left is to identify the beneficial intermediary step between your behaviour changes and a mitigated systemic problem and…voila! You are done.

If you want more, some of my tools and templates are below. As always feel free to reach out if you want to talk.

Intro to ABC: slides with presenter’s notes.