which test for 4level categorical IV and continious IV?

[morepil]

I have a 4 category IV: What's your family type? Authoritarian, democratic, undefinable, inconsistent.
The other IV is idealization, continious variable ranging 0-15. Found by summing up answers to three 5-point-Likert questions asking how was your childhood, how was your family..
My DV is a score on the RWA (authoritarianism) scale. Continious type obviously.

Which test to use?

 

[ondansetron]

I have a 4 category IV.
What's your family type? Authoritarian, democratic, undefinable, inconsistent.
My DV is a score on scale. Continious type obviously.

Which test to use? My hypothesis is that participants who respond as authoritarian will have the highest DV scores.

Can you provide more information on the dependent variable? How is it generated? Do people answer a questionnaire 1-5 (low-high), for example?

 

[morepil]

Yes. The DV is the result of an addition of a 20 question 9 point likert scale. I'll add all the points from the 20 quetions and come up with a score. 0+2+5+2+8+2+5+.. etc: 65. For example.
Does this help? And why does it matter how the DV is generated?
Edit:thanks

 

[morepil]

Correction. I corrected the original post as there was no information on the continious IV. Please help!

 

[hlsmith]

What does the distribution of the DV look like? Post a histogram of its distribution please. Also what is your sample size?

 

[morepil]

Thanks.

I felt it was positively skewed but looks normal to me.. Here it is.
Population size: 272

 

[hlsmith]

You may be fine with multiple linear regression if the assumptions hold. My other question would be on what the distribution of the continuous IV? What I am looking for is how bounded it seems and any other irregularities.

 

[morepil]

Here's the graph of the continious IV. Does this help?
I'll try doing a multiple regression with assumption testing

 

[hlsmith]

Is variable bounded at ~13? Yes, that may be a good first step.

 

[morepil]

you mean the min and max? 3-13 for the second one; 20-149 for the first one I mentioned.
Thanks a lot.

 

[hlsmith]

Yeah, bounded variables have difficulty with symmetry and respectively confidence intervals. You have to watch out since 2 standard deviations one way may be fine but in the other direction it may not exist.

 

[morepil]

Thanks, ill take a look at these more in more depth.. Although the first step was not vey encouraging: Kolmogorov-Smirnow and Shapiro-Wilk are both at 0.0, so looks non-normal? Even at population size 272, I gotta check this?
And on SPSS, I'd input 4 dummy variables for my first IV(4 levels) and the continious IV as independent variables? and the DV as the DV..
I haven't done the rest of the assumpiton testings yet.