A principle we try to follow at Causal Map:
“leave the evaluation (the evaluative judgements) to the evaluator”:
We see the job of the software/algorithms as preparing everything data up to the point of comparison. For example, to help prepare information like this:
“you’d have to destroy 37 pieces of evidence in order to destroy any argument from intervention X to outcome Z, but in the case of intervention Y, there are only 8 pieces which you’d have to destroy, and the chains in the first case are shorter and involve more respondents. Let’s now discuss what this means ….”
… rather than this:
“the software says Intervention X had a bigger effect than Intervention Y”.