Academic references to outcomes theory: Duignan (2009d)*
The great thing about these is that just measuring that they have occurred is enough to prove that you caused them to happen.
The problem is that they'll tend to be at a lower level within your outcomes model (outcomes systems building-block one).
There will be a number of things that are controllable and can be measured about your program.
That is, they'll often tend to be more 'output-type' than 'outcome-type' indicators.
Where this is the case, the only thing you can do to prove impact on higher-level outcomes is to assess the possibility of doing impact evaluation (building-block five.)
How controllable indicators look when they're mapped onto an outcomes model
* This particular reference can be cited to refer to the material on this page.