I assume you know the reason for sensitivity analyses, namely to be able to specify the robustness of your results.  For a concise discussion of the basics, see:  http://en.wikipedia.org/wiki/Sensitivity_analysis    

The fact is that any of the variables in your model can be assessed in a sensitivity analysis.  Assuming that you are trying to reach a conclusion from the results of your model, the kind of question you should be asking yourself and addressing for those whom you want to benefit from what you have done is something like this: "What if our estimate of x (or y, or z) was wrong by a factor of 1.5 (or 2 or 3), how much would that have altered our conclusions?"  The more robust your results the less they will be affected if your estimates are incorrect by such factors.

Many thanks for this inside Stephen. I believe I can proceed to vary the useful-life estimates of some of my Capital cost components to see the effect on my outcome.

Once again thanks

You are welcome, Edmund.  I assume that if it is capital costs that you are varying, it should be relatively easy to compute upper and lower bounds for those figures - e.g., costs of materials, labor, if you are building something, or machinery, installation, training if it is capital equipment.

Hi Edmund,

The sensitivity analysis is very important in your cost analysis as you will likely have to make decisions in your analysis regarding what types of resources you include, how many of these resources you include and what value you give them.

An easy example (if you are costing over time) is discount rate…you have a choice to make in whether or not to discount and if you do discount, which discount rate to use.  Individuals and groups interested in your results may however want to see the results with and without discounting.

Another example is level of efficiency. You may collect date based on what is currently happening, but you may also want to show results given changes in implementation efficiency/effectiveness.  

Sensitivity analysis can help you divide your analysis into (1) a main analysis (based on your chosen assumptions) and (2) an analysis given different assumptions.

Hope this was helpful!

Good luck!

Tina

Hi Emdud, and many thanks to your question and our colleagues answers. 

I think you are well informed about fundamentals of Sensitivity Analysis. the I am not into repeat it again. I want to give some applied tips.

you can examine different approaches of depreciation ( Accountant or Economic ) , or as you said, different values of capital costs. I think it is also relevant for Building Investments, you can use depreciation costs or opportunity cost ( Rental costs). In my works for Iran. I have to notice some macroeconomic factors, because of economic fluctuations. for example, may be for next year our Inflation rate change, then for next year's expected unit costs I have to examine the impact of different inflation rates on equipments or consumables.

based on your study's aims, you can test different strategies of resource allocation, for example if ministry of health wants to outsource a primary health center with owning building and equipments or they want to assign all to private sector.

I think, sensitivity analysis gives this opportunity to examine the efficacy of your model in uncertain situations. all the variables up to you and your study's context.