Hey,
My apologies for my ignorance on these issues. but i find conflicting answers on the web.
Data issue: How to translate all the survey questions into 1 variable
1) My dependent variable is constructed from 15 survey questions on a 5 point Likert scale. Do I take an average, or do I just take the sum of all questions?
2) My independent variable is in the same data form: 18 survey questions, but with three dimensions. Hypothesis are that each dimension is positively related, but that all of them together should work even better. Is this even possible? I'm not getting much support where i should be getting it, so i hope you guys can answer this.
3) my observations are from 4 companies. Does this change anything or add complexity?
4) From what i figured now, I should use the ordered logit/probit model. So I assume that summing all results is not the answer?
Thank you.
My apologies for my ignorance on these issues. but i find conflicting answers on the web.
Data issue: How to translate all the survey questions into 1 variable
1) My dependent variable is constructed from 15 survey questions on a 5 point Likert scale. Do I take an average, or do I just take the sum of all questions?
2) My independent variable is in the same data form: 18 survey questions, but with three dimensions. Hypothesis are that each dimension is positively related, but that all of them together should work even better. Is this even possible? I'm not getting much support where i should be getting it, so i hope you guys can answer this.
3) my observations are from 4 companies. Does this change anything or add complexity?
4) From what i figured now, I should use the ordered logit/probit model. So I assume that summing all results is not the answer?
Thank you.