Causal diagrams and Theory Maker: left-to-right? bottom-up? top-down?
There is an interesting discussion going on at LinkedIn about Theory Maker. One commenter asked why Theorymaker presents things left-to-right. This is what I wrote.
Well there are two things here:
The first is the direction of the arrows themselves. It sounds like you are saying that the arrows should go from, say, outcome to output because one thinks first about the outcome? Sure, … from a point of view of interpretation you can draw the arrows the other way round if you want. But I think the consensus is that the direction of the arrow means causation, so they should go from output to outcome. But if you want to think the other way round, you can draw the arrows the other way round in Theorymaker too.
Second, you could be talking about the presentation, and here it makes a difference where you come from … if you come from a left-to-right writing culture, left-to-right makes a lot of sense because it corresponds with the flow of time: outcomes come later than outputs. Personally I am not fond of bottom-to-top because I am not very happy with the interpretation that “higher” means “better”. In Theorymaker, you can mark any variable as valuable just by adding
_val to its name. Thus we separate value from position. But still, different people like different things so I added a button and now you can display the diagram anyway you want.
The same commenter also asked whether the parts of a causal diagram have necessarily to be ‘INPUTS’ , ‘ACTIVITIES’ and ‘OUTCOME’
This is what I wrote:
… well there is an example diagram which loads up when you load the app, and I pasted in a similar one at the start of this discussion, but Theorymaker is not tied to these designations at all. Personally I only distinguish inputs (completely under your control) and
valued variables, those variables which are important to us. In between, there could be any causal network at all, with any number of variables, layers, branches and loops. I totally agree that assumptions are important too, but most often an assumption is another variable which should be included in the causal diagram, not relegated to another column or an annex. I wrote a bit more about assumptions here. So I do think that causal networks such as those you can draw with Theorymaker are enough for project modelling - at least from a theoretical point of view. In practice, of course, there might be lots of additional features from cost to precise timing which you can’t reproduce in Theorymaker.