I’m not very programming-oriented so I can’t really engage with the technical side of things, but I am curious whether/to what extent you think that a quantitatively oriented tool like Pedant would be complemented by a conceptually-oriented framework like the [TUILS framework] (https://forum.effectivealtruism.org/posts/SyWCmqtPsi4PsxnW8/the-tuils-framework-for-improving-pro-con-analysis-1) ? The latter is not so (inherently) focused on specific calculations or parameter harmonization, but it does try to highlight bigger picture conceptual steps/assumptions (e.g. counterfactual effects, the normative value of a given change).
However, I don’t fully understand your project. Is it possible to refine a Cost Effectiveness Analysis from this? I’d probably need to see a worked example of your methodology before being convinced it could work.
The purpose of the TUILS framework is to break down advantages and disadvantages into smaller but still collectively exhaustive pieces/questions (e.g., are the supposed benefits counterfactual, how likely will does the desired state materialize).
I’m not sure if you have a very particular definition for “Cost Effectiveness Analysis,” but if you just mean calculating costs and benefits then yes: the whole point of the framework is to guide your reasoning through the major components of an advantage or disadvantage. There is a spectrum of formality in applying the TUILS framework, but the informal version basically treats the relationship between the factors as roughly multiplicative (e.g., anything times zero is zero, if you only solve half the problem you may only get half the originally claimed benefit—assuming there is a linear relationship).
I haven’t fully sketched out a truly formal/rigorous version since I’m not a mathematician and I don’t see that as the main value of the framework (which tends to more about assumption checking, intuition pumping, concept grouping/harmonizing, and a few other things). However, if you’re actually interested in the more-formal application, I could write out some of my rough thoughts thus far. (It’s basically “imagine an n-dimensional graph, with one of the dimensions representing utility and everything else representing input variables/measure…”)
In terms of example applications, I gave some simplified example applications in the post (see the subsection “ Examples of the TUILS framework being applied”), but I haven’t done any written, deep, formal analyses yet since I haven’t seen that as a valuable use of time since organic interest/attention didn’t seem very high even in the simpler version. That being said, I’ve also used a less refined/generalized version of the framework many times orally in competitive policy debate (where it’s referred to as the stock issues)—in fact, probably more than 90% of policy debate rounds I watched/participated in make reference to it. (All of this just to say that a version of it is pretty widely used in the policy debate community)
I’m not very programming-oriented so I can’t really engage with the technical side of things, but I am curious whether/to what extent you think that a quantitatively oriented tool like Pedant would be complemented by a conceptually-oriented framework like the [TUILS framework] (https://forum.effectivealtruism.org/posts/SyWCmqtPsi4PsxnW8/the-tuils-framework-for-improving-pro-con-analysis-1) ? The latter is not so (inherently) focused on specific calculations or parameter harmonization, but it does try to highlight bigger picture conceptual steps/assumptions (e.g. counterfactual effects, the normative value of a given change).
Maybe, your work there is definitely interesting.
However, I don’t fully understand your project. Is it possible to refine a Cost Effectiveness Analysis from this? I’d probably need to see a worked example of your methodology before being convinced it could work.
The purpose of the TUILS framework is to break down advantages and disadvantages into smaller but still collectively exhaustive pieces/questions (e.g., are the supposed benefits counterfactual, how likely will does the desired state materialize).
I’m not sure if you have a very particular definition for “Cost Effectiveness Analysis,” but if you just mean calculating costs and benefits then yes: the whole point of the framework is to guide your reasoning through the major components of an advantage or disadvantage. There is a spectrum of formality in applying the TUILS framework, but the informal version basically treats the relationship between the factors as roughly multiplicative (e.g., anything times zero is zero, if you only solve half the problem you may only get half the originally claimed benefit—assuming there is a linear relationship). I haven’t fully sketched out a truly formal/rigorous version since I’m not a mathematician and I don’t see that as the main value of the framework (which tends to more about assumption checking, intuition pumping, concept grouping/harmonizing, and a few other things). However, if you’re actually interested in the more-formal application, I could write out some of my rough thoughts thus far. (It’s basically “imagine an n-dimensional graph, with one of the dimensions representing utility and everything else representing input variables/measure…”)
In terms of example applications, I gave some simplified example applications in the post (see the subsection “ Examples of the TUILS framework being applied”), but I haven’t done any written, deep, formal analyses yet since I haven’t seen that as a valuable use of time since organic interest/attention didn’t seem very high even in the simpler version. That being said, I’ve also used a less refined/generalized version of the framework many times orally in competitive policy debate (where it’s referred to as the stock issues)—in fact, probably more than 90% of policy debate rounds I watched/participated in make reference to it. (All of this just to say that a version of it is pretty widely used in the policy debate community)