What do we call the "middle" variables in a complex research model?

How do I model normal distribution with upper max and lower min in Excel?

• Hello everyone, can you please help me with something that is driving me nuts? :-) I am trying to run a Monte Carlo simulation on potential sales of a mobile application that serves people who do yoga on a regular basis. The key variables are: 1.  Total number of people with a smartphone and who regularly do yoga. 2.  The % of those people I think I can expose the app to in the first 12 months (through app store marketing, etc.) 3.  The % download rate following marketing exposure. 4.  The cost of the app ($1.99) which is fixed. Multiplying each of these variables with each other will give me an expected 12 month revenue number.  Having estimated the 90% CI upper / lower bounds, and mean for each of the first three variables I've already built a Monte Carlo simulation model using the NORM.INV function.  Specifically I'm using the following equation: =NORM.INV(RAND(),MEAN,(UPPER BOUND-LOWER BOUND)/3.29) So here's the issue:  In reality, each of the variables I am modelling has a known upper and lower limit.  For example, I have market research data which says there are at least 2 million regular yoga participants with a smartphone in the US but no more than 5 million.  However, the normal distribution generated by my current model routinely delivers values below 2 million and above 5 million. So here's the question:  Is there any way I can model a normal distribution BUT with fixed upper and lower limits? Any help you guys can give on this will be hugely welcome and very gratefully received. Many thanks, Mike p.s.  I've tried using the BETA.INV function but, to be honest, I'm already at the limit of my statistical modelling capabilities so if you could bear that in mind with any replies that would be appreciate.

• Answer:

  I don't think you can. If you want a normal distribution, you can't specify upper and lower limits. Or alternatively, if you want to specify upper and lower limits, you can't have a normal distribution. One approach you can use is to just discard those that are outside your limits.