# Choosing the correct Statistical test

#### taspinar

##### New Member
I am writing my first year project about working conditions in Denmark and have done some statistical analysis. But I am in doubt whether I have used the right methods. The reason behind my despair is my absent teacher which resulted in a huge lack of guidance. (Enough self-pity) My problem is that I have made and reflective index of 6 ordinal variables and have an ordinal dependent variable. So far I have done an ANOVA-test and a post hoc Tukey-test, which I think is wrong. My biggest problem is that my DV is not normally distributed, which is an assumption behind the ANOVA. I therefore feel quite lost and I have been looking at this website, http://www.ats.ucla.edu/stat/mult_pkg/whatstat/, for some answers but I can’t figure out what “levels” means.
My DV is ordinal with 4 categories ranging from not at all to always. (Satisfaction)
And my IV is an index constructed by 6 ordinal variables with 5 categories each, and therefore 25 in the index, ranging from not at all to always. (Freedom) The index is not a problem I used a text-book guide to construct it.
I am trying to uncover a correlation between job satisfaction and freedom in the work place.
I would very much appreciate if you had any clue about what tests I could do with the data I have. I am using STATA.
Best regards,
taspinar :wave:

Last edited:

#### Karabiner

##### TS Contributor
My DV is ordinal with 4 categories ranging from not at all to always. (Satisfaction)
An ordinal variable like this you cannot use as DV in an ANOVA.
I am trying to uncover a correlation between job satisfaction and freedom in the work place.
You could use Spearman's rank correlation rho.

With kind regards

K.

#### Karabiner

##### TS Contributor
I have 1069 observations and actually i dont know how to do it?
The statistical software will do that for you.

With kind regards

K.