When conducting building performance evaluation for a development of dwellings, is there a formula for calculating the minimum sample size? For example, rather than conducting full evaluation on all 200 dwellings in a development, which would be both impractical and costly, is there a minimum that would give an acceptable statistical confidence to the results?
Does the answer depend on the parameters being investigated and properties of the population?
I appreciate this question could be broadened out to other fields - building performance is my field.