Questions regarding correlation

Question

I'm writing a thesis about cooperating competition between companies. I have a few problems with the last two research questions:

RQ7: Are there any differences between the perception of costs and benefits of the cooperating competition between the companies that have previously experienced the concept of cooperating competition and those who had not.

RQ8: Are there any differences between the perception of costs and benefits of the cooperating competition between different kinds of industries in which the companies conduct their business.

So, for RQ7 I have 3 major perceptions of costs which are scales ranging from 1(I highly don't agree) to 7(I highly agree), the same goes for perceptions of benefits and i have a nominal variable for previously experienced cooperating competition, which is just yes or no.

RQ8 is similar to RQ7 except it has a nominal variable with eleven possibilities for different kind of industries.

I used custom table, dragged perceptions in columns and previous competition in row and I got the means for every perception for yes and for no (previous competition) and for different kinds of industries.

Looking at the results I don't know if I answered my questions, is there maybe a better tool to for answering these questions, I'm thinking about correlation of some sorts, but don't know how to put it all together. please take a look and tell me what you think.

Answer 1

Welcome to the site.

It isn't clear what you did once you got the variables into a table. However, if you are looking for differences between two groups, I don't see how correlation is going to help. In RQ 7 you seem to have an ordinal dependent variable (going from 1 to 7). The simplest thing would be a t-test, but that is not technically correct since it is really for continuous variables. You could use a test of medians - medians don't depend so much on the data being continuous. Or you could use a test of ordinal differences such as the Jonckheere Terpstra test. Another possibility is ordinal logistic regression with the scale as the dependent variable and "group" as the independent variable. This would allow you to add any other variables you have to the model.

For RQ 8, you say you have a nominal variable. I am not sure I understand what that variable is, but, if you are right then two possibilities are chi-square test of independence and multinomial logistic regression.