If each of these parties involved is important, I think you should separate these response into three dicotomous/binary variable and apply logistic regression for each of them. However, if the important point is the number of satisfied parties, you may use the ordinal regression.

Hope this help.

Hello Ezra Putranda Setiawan , that analysis is also on the to do list, can I compare the results of these regressions?

Hello Jochem Groot Jebbink , i think you can compare the result if you use the ordinal regression. However, if you use three logistic regressions, a statistical procedure might available (see https://www.statalist.org/forums/forum/general-stata-discussion/general/1488661-how-to-compare-%CE%B2-coefficients-from-two-different-logistic-regression-models-using-permutation-test) to compare each coefficient in two logistic regression model. But if you only need to compare the significant variable (and not the coefficients), you can do it qualitatively.

Looks like binomial data: Y ~ Bin(k;n) with k = number of satisfied parties (0,1,2,...n) and n = total number of parties (that may or may not be satisfied). So I would use a binomial regression model.

https://towardsdatascience.com/the-binomial-regression-model-everything-you-need-to-know-5216f1a483d3

https://pdfs.semanticscholar.org/a124/bc91adacf5917e29807e71e363ac5dcecad4.pdf

http://www2.stat.duke.edu/~fl35/teaching/440-19F/decks/cs03_2_deck.html

Hi, both you and Jochen Wilhelm are correct and this is an ordinal logistic regression. Now what you do depends on the pupose of your model, explanation or prediction. To go from here see my post at this link for directions: https://www.researchgate.net/post/What_test_can_I_use_to_study_if_there_is_a_significant_impact_of_ageeducation_profession_marital_status_on_frequency_of_visit_to_a_mall Best wishes, David Booth