# What statistical tools should I use to analyze the data

#### Withnail73

##### New Member
I am working on my masters thesis in psychology. I have been using a scale to measure a disorder and relate it to 3 other variables. The scale works as follows:

I would like to do the multiple regression but it might be hard as the scale is not continuous.
Section A (psychological symptoms)
Likert scale 1-5 for each question
Questions A1 to A6 (3, 4, 5 points in at least 4 questions = 3 points,. If they do not answer from 3-5 in at least 4 questions, they get 1 point)
Scale goes from 1-5
Section B(physical symptoms)
Questions B1 to B8 (3 points in 3 sections= 3 points, no = 1 point)
Scale goes from 1-3
Section C( trauma)
Answer yes = 3, no =1 point
Section D( Significant distress)
Answer yes=3 points, no = 1 point
12 points = Disorder confirmed

I would like to do the multiple regression but it might be hard as the scale is not continuous. Not sure what I can do with this. I initially divided my subjects into two groups. Those who pass and those who do not pass but my advisor said the number who passed was too low to be meaningful. Out of 203 subjects, 10 passed. So,he suggested I use a scale instead of pass/fail and multiple regression but not sure how to do it with this type of scale.
What approach would be best with this type of scale?

#### Karabiner

##### TS Contributor
If you consider a measurement with k levels as non-interval,
then you can consider to transform it into k-1 dummy variables
which you enter into the analysis (one level receives no dummy,
as this would be redundant, and serves as reference level).
Or you use ANOVA or ANCOVA.

With kind regards

Karabiner

#### Withnail73

##### New Member
If you consider a measurement with k levels as non-interval,
then you can consider to transform it into k-1 dummy variables
which you enter into the analysis (one level receives no dummy,
as this would be redundant, and serves as reference level).
Or you use ANOVA or ANCOVA.

With kind regards

Karabiner
Thank you for your response!Just one more question. Does it matter that the scale above is my dependent variable? Basically, I want to show how other variables from other tests impact scores on the scale above. The scale above measures whether or not someone has a disorder and I want to show that other variables determine whether or not someone has the disorder. Take care.

#### Karabiner

##### TS Contributor
Ok, you did not say that it is the dependent variable. Unfortunately, what this scale measures and how
it is evaluated and what level of measurement can be assumed is not self-evident. In would look into
the manual and/or in previous studies which used this scale.

With kind regards

Karabiner