shifting the burden
Posted: Mon Oct 18, 2010 3:43 pm
Hi
Of course both qualitative models described are only suppositions, as there may be something important missing in these qualitative models. One can too object that due to the extreme generality of the problem, these kind of models are necessarily very approximative or are eventually plain wrong depending on the case they are supposed to describe.
To be correct, the structure of these models should be validated by transforming them into quantitative models, running them and comparing the results with different real cases of burden shifting, which should be practically very difficult to realize.
In this case, where the structure of the model is difficult to validate, another solution is to build qualitative and quantitative models made by diffierent people with no contact between them, run these models, compare the results and try to understand what the differences are coming from. This method of building a model with different groups, whould greatly increase the credibility of models built in an environment where there is a lack of rererence modes or if the past reference modes cannot be projected in the future with enough credibility.
Regards.
JJ
[Edited on 18-10-2010 by LAUJJL]
[Edited on 19-10-2010 by LAUJJL]
Of course both qualitative models described are only suppositions, as there may be something important missing in these qualitative models. One can too object that due to the extreme generality of the problem, these kind of models are necessarily very approximative or are eventually plain wrong depending on the case they are supposed to describe.
To be correct, the structure of these models should be validated by transforming them into quantitative models, running them and comparing the results with different real cases of burden shifting, which should be practically very difficult to realize.
In this case, where the structure of the model is difficult to validate, another solution is to build qualitative and quantitative models made by diffierent people with no contact between them, run these models, compare the results and try to understand what the differences are coming from. This method of building a model with different groups, whould greatly increase the credibility of models built in an environment where there is a lack of rererence modes or if the past reference modes cannot be projected in the future with enough credibility.
Regards.
JJ
[Edited on 18-10-2010 by LAUJJL]
[Edited on 19-10-2010 by LAUJJL]