Regression analysis: redundant dependent variable

I'm not sure you want a regression analysis here. What is your independent variable? Or, in other words, what are you regressing the machine choice against? If the number of cups always equals five you can turn each response into a proportion for machine one (0.0, 0.2, 0.4, 0.6, 0.8, 1.0) and analyse the proportions. What is your hypothesis?

Thanks for your answer, here is additional info
H1: Students at the school have preferences towards the coffee from Caféen or Coffee Machines influenced by different factors .

The students have asked to evaluate how much they value the independent variables for the coffee machines and the cafee coffee (manual served coffee) on a 7 point scale. 1=strongly dissagree and = 7 strongly agree

First regression line
the dependent variable is Number of cups bought at cafeen=service+location+price+que+taste

seccond regression line
the dependent variable is numbers of cups bought at the machines=service+location+location+taste+que
then they have been asked: out of the last 5 cups bought, how many was bought at the machines and how many at the cafee.

the sum should equal five cups, meaning that the dependent variable becomes redundant?