I have 2 IV:
- Scale data of skills achievement by our University alumni (They rate from 1 to 7 their acquired skills at University)
- Gender (Already converted it to a dummy variable)
And I have 1 DV:
- Job finding difficulty that has 5 options: (Very Easy, Easy, Neither Easy nor Difficult, Difficult, Very Difficult).
I have 2 questions that I need your assistance with:
1. Can I use a dummy variable with 0, 1, 2, 3, 4, 5 values for my DV?
2. My first IV is not normally distributed, can I still run this regression with it?
I am trying to see the contribution of gender or any specific skills on job finding difficulty.
Please answer as clear and as short as possible.
Thank you!
1. It sounds like your dependent variable is ordinal in nature. So you could use ordinal regression...
1. I agree with Sal Mangiafico
2. yes the IVs are assumed fixed
Best, David Booth
Sal Mangiafico Do I have to change my dependent variable into a dummy variable? What would be the values of the dummy variable if so?
David Eugene Booth Meaning that I can go ahead with them, right?
Thank you so much for your responses!
You're correct in your comment. However you need to us ordinal logistic regression since your DV is ordinal . Google for further into. Best wishes, David Booth
You may want to look at this free course - Module 9 is on multinomial regression for ordered categories:
http://www.bristol.ac.uk/cmm/learning/online-course/
if you are not familiar with it, it may be worth starting with the binary logit model and then the unordered nominal regression - there are graded modules on each type of model.
Dzhamshed Salomov , if you are looking at ordinal regression, then you won't need to convert your DV to dummy variables. The software you are using should handle this aspect, but you may need to make some adjustment to your variable. For example, in the R package ordinal ---- at the least the last time I checked ---- the DV needs to be coded as an ordered factor variable.
You can use ordinal logistic regression without converting it to 0,1.... however you can change it into dummy variables and apply the logistic regression analysis. moreover, no need of checking the normality assumptions for your IV.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression