I became interested in agency as a programme outcome because many familiar measures describe what a programme delivers without describing what changes in the participant’s capacity to act afterwards. A person can receive information, complete a service, report satisfaction and even achieve an immediate outcome while remaining dependent on the same external conditions the next time a similar situation arises.
Agency raises a different question: has the intervention changed the person’s practical ability to influence what happens next?
I use practical deliberately. Agency is easy to invoke at a level where it becomes difficult to distinguish from confidence, autonomy or choice. For programme design, I think it needs to be tied to more specific changes. Someone may become better able to understand a situation, identify plausible options, obtain relevant information, navigate an institution, mobilise support, make a decision, initiate an activity or pursue something that was previously outside their effective reach.
Those changes can arise through very different mechanisms. A person may acquire knowledge. A programme may connect them to a relationship they can use later. A barrier may be removed. They may develop a skill. Their understanding of the institutional environment may improve enough that a previously opaque pathway becomes navigable. In each case, the outcome is not merely something the programme has done for them; something has changed in the set of actions they can realistically undertake.
This does not imply that agency should replace other programme outcomes. There are many contexts in which the relevant outcome is appropriately direct: receiving food, securing housing, resolving a legal matter, treating an illness. It would be perverse to treat immediate material outcomes as secondary merely because agency sounds more developmental.
The useful distinction is that some interventions have both a direct outcome and a capability effect. A service may resolve the immediate problem while also leaving the person better able to navigate similar problems in future. A community programme may provide a worthwhile social experience while also making it easier for somebody to find collaborators or contribute to later work. A support intervention may help someone take the next step while improving their ability to identify and assess options independently.
Once agency becomes an explicit intended outcome, programme design has to become more precise about these mechanisms. It is insufficient to state that participants will be “empowered”. We need to specify what they should become more capable of doing and why the intervention is expected to produce that change.
Measurement then becomes less abstract as well. Depending on the programme, we might examine whether participants can identify realistic next steps, reach resources without intensive mediation, navigate relevant systems, initiate activity, exercise greater discretion over decisions, or mobilise relationships and knowledge that were previously unavailable to them.
There is an important methodological difficulty here. Agency is highly context-dependent. A person can have substantial capability in one domain while facing severe constraints in another. Self-reported confidence may improve without corresponding changes in practical possibility, while somebody may become objectively better able to act without describing themselves as more empowered. No single indicator is likely to capture the construct adequately.
For Real Good, this argues for treating agency as a programme-specific capability outcome rather than as a universal score. The relevant question should be grounded in the action environment of the intervention: what should this person become better able to understand, choose, access, initiate or sustain?
I find that formulation useful because it disciplines both design and evaluation. It forces us to move from an attractive value — empowerment — towards a claim that can be examined: what has changed in the person’s effective capacity to act?