Within outcomes theory all interventions are conceptualized as having an implicit underlying outcomes model (

Academic references to outcomes theory

Signup to the DoView School of Outcomes Tips Newsletter

“Boxes within an outcomes model can be: relevant; influenceable; measurable; controllable; attributable; and/or accountable"

“The best way to specify the type of box you are referring to (e.g. an outcome, an impact), is to specify it in terms of the technical features boxes in outcomes models can have"

The principle

Outcomes theory is focused on identifying, discussing and clarifying the conceptual issues faced by interventions and related outcomes work. In doing this, outcomes theory views all interventions as having an underlying implicit visual outcomes model. This is regardless of whether this implicit model has actually been surfaced and articulated within a real-world outcomes model built for the particular intervention. Imaging that all interventions have a notional visual outcomes model provides an efficient way of talking about outcomes theory concepts as they appy to interventions. 

The problem

Like any theory, outcomes theory seeks to provide the simplest possible conceptualization of outcomes issues that can do justice to the topic and provide practical ways of working with the issues faced in real-world outcomes work.

A central issue in thinking and talking about interventions and outcomes work is that one very quickly gets to the point that you need a way of talking about the various different levels of ‘outcomes’ that may end up resulting from an intervention (e.g. activities, outputs, intermediate outcomes, outcomes, results, goals, impacts etc.).

Unfortunately there is currently not enough consensus amongst different people working on outcomes issues regarding which words are the most appropriate to use. It  does not seem likely that any time soon it will be possible to enforce the use of a particular set of terms with standardized meanings.

This makes it difficult to have a concise and non-contested discussion about the different levels of ‘outcomes’ arising from an intervention because for instance there is not even an obvious word to use to describe the varous levels of ‘outcomes’ arising from an intervention. 

The problems is that the word outcomes is already firmly fixed in some people’s minds as being one partiuclar level arising from an intervention. Given that people think that an outcome is at a particualr level in an hierarch of effects from an interventions means that it is not a good word to use as a generic terms for all of the possible levels within a hierarch of effects.

The solution

Outcomes theory abandons at this stage the attempt to demand that that people us a particular set of terms for the various levels of ‘outcomes’ arising from an intervention, outcomes theory takes another approach. 

It proposes that we think in terms of there being an underlying implicit visual outcomes model that underpins any intervention. This then means that the effects of an intervention can be viewed simply as a set of boxes within this implicit outcomes model. This provides a powerful neutral generic term for different parts of the hierarch of effects. 

It allows one to talk efficiently about the different levels of effects: low, medium, high level boxes. 

In addition it makes it easier to make other points e.g. the distinction between an effect of an intervention and its measurement. Using the implicit outcomes model approach means one can simply say that there is a box and there is a measure of a box and these two are separate entities. 

The concept of an implicit outcomes model easily transfers to the real-world because people often use actual visual oucomes models for intervention work (called things like theories of change, program theories, results roadmaps, results chains, strategy maps, program logics, intervention logics etc.) 

problems around various types of boxes that can appear within outcomes models occur in part because different disciplines (managers, strategic planners, policy analysts, economists, accountants, performance managers, evaluators, social scientists etc.) and different sectors use different outcomes terminology. 

A lot of effort currently goes into attempting to find (and enforce) the ‘right’ terms for the different types of boxes that can appear within outcomes models.

However from a technical point of view, regardless of the names given to the types of boxes appearing within outcomes models, it is essential that each type of box is fully specified in terms of the features of that type of box.

The possible features of types of boxes within outcomes models include: relevance; influenceability; measurability; controllability; attributability; and, accountability.

The first table below shows these features and the values they can have. 

The second table shows how names for particular types of boxes within outcomes models can be unambiguously defined by specifying which values they have on which of the features. At the current time, when there are considerable differences of opinion about the best names to call different types of boxes in outcomes models, specifying boxes' features in this way is the fastest and most efficient way of cutting through the confusion in this area.


If a funder asks a provider to ‘give us a list of your outcomes’, the best way for the funder to communicate exactly what they want to the provider is to fully specify the features of the boxes they are calling outcomes. For example, they could fully specify an outcome as: relevant; high sequential position; influenceable; not-necessarily measurable; not-necessarily controllable; not-necessarily attributable; and, not-necessarily a direct accountability. 

If the funder wishes, it can vary this set of requirements. For instance some funders might want to require that the boxes they are calling outcomes also be measurable. This approach ensures that the provider is completely clear about what is required and avoids any confusion which can occur if they are just asked for their outcomes without further explanation.

It is appropriate for providers, whenever they are asked by a funder to provide a list of their outcomes, results or impacts to ask that the provider fully specifies what they mean by using Duignan’s Features of Boxes That Appear in Outcomes Model List and reference this page.

Duignan, P. (2009d). Using Outcomes Theory to Solve Important Conceptual and Practical Problems in Evaluation, Monitoring and Performance Management Systems. American Evaluation Association Conference 2009, Orlando, Florida, 11-14 November 2009.