Starting with your briefing in Chapter 1, a series of simplifying assumptions have been used for the case of NewTel. There are good reasons for making such assumptions. Two of the Guidelines G 3 (Chapter 2) and G 9 (Chapter 5) have already alluded to at the delicate balance between realism and usefulness:
Guideline 3 (G 3): Model parsimonyWhen developing and using causal diagrams, make sure that all relevant, and only the relevant, variables and causal links are represented.
Guideline 9 (G 9): Start with only the most important piece of structureWhen approaching a complex problem, start modelling with the most fundamental piece of causal structure and incrementally add new pieces of causal structure as you discover the need.
In other words, we have strived to keep the model as simple as possible in order not to dedicate effortsand time to unnecessary aspects. A model that has only the relevant variables and links also has the appropriate level of simplifications, and it satisfies two criteria related to the model’s purpose, which you can use to follow Guideline 3.
A model has an appropriate balance between simplification and realism when the following two criteria are satisfied:
- you cannot take away any variable or link without compromising to achieve of the model purpose;
- adding more variables or links does not improve the fulfilment of the model’s purpose.
If you fail to satisfy the first criterion, ...