*Bounty: 50*

*Bounty: 50*

*For lack of a better way of presenting my question I will use the example given*

A company wishes to quickly estimate the mean 2016 property taxes paid by its $1000$ stores

using a stratified random sample of $n = 100$ stores. Frequency data on 2015 property taxes

paid by these same $1000$ stores is available in the form of classification by $1000 increments

and appears in the following table.

(*The table is not necessary for the question so I won’t put it here*)

What are the assumptions necessary to have about the relationship between 2015 and 2016 property taxes to estimate the 2016 mean? I think they need to share the same variance but I want to be sure what the necessary assumptions are.