comparing feedback of customers in two branches of an organization

#1
Hello , I am facing a problem.
A voting kiosk ask persons in two branches of an organization about the organization's services.each person has an answer to question as a grade from 0 to 100 . Suppose 200 person in the first branch refer to kiosk and vote and the average of their opinions is 10. In the second branch there are 6 person vote, and their average opinions are 80. Because the number of samples in these two branches is very different, comparing these two branches can not be valid. What is my solution ? How can I compare service in these two branches?Is there a way, for example, to make the results realistic by multiplying a coefficient proportional to the number of samples obtained in the average?
thank you all.
 

hlsmith

Omega Contributor
#2
The numbers are what they are. I don't think you should try to reweight them. You can directly compare them and their measures of dispersion will come into play and provide precision around parameters. Without seeing your data, I would recommend beta regression. A cruder approach given your small a sample size could be the Wilcoxon rank sum.

Is it not possible to acquire more data points? You also have to keep in mind that people self-select to participate and what influence that could have on the smaller of the samples.