Which test for cont DV, 1 nominal IV (2 cat) + 1 nominal IV (4 cat) + 1 cont IV

#1
Hello,

I am a bit struggling with choosing which test I should apply to my data.

My dependent variable is continuous (participants could score between 0 and 6) and I have three factors:
1) a nominal variable with two categories (namely "native" and "nonnative")
2) a nominal variable with four categories (namely four different culture groups)
3) a continuous variable (namely a proficiency score). BUT I expect that the relationship between the DV and this variable is exponential.

So, I am a bit lost… Should I ignore my expectation that the relationship between my DV and IV3 is might be exponential, and just apply a linear regression?

Many thanks for your help!
 

Karabiner

TS Contributor
#2
You could transform the proficiency score in order to linearize its relationship with the DV. Taking the logarithm could be an option.

With kind regards

K.
 
#3
Thank you, Karabiner. And I guess I have to turn my second nominal variable (with your categories) into dummy variables, right? My question is then: In my regression, I would like to include another 2-cat nominal variable (i.e. nativeness, coded as 1- native, 2- nonnative) and my (transformed) continuous variable next to my dummy variables. But how does the program know that the dummy variables included have to be compared to the dummy variable that has been left out, and not to one of the other variables in my dataset that have been left out?
 
Last edited:
#5
Only integers. They had to answer six multiple-choice questions, for each question they could score 1 point if they chose the right answer (and 0 point if they answered it wrong)