I would like to conduct a regression analysis to figure out if an attitude affects the number of accepted objects. The attitude/independent variable is a metric 7-point likert scale (1-7). I need to create a dependent variable to measure the accepted objects. At the moment I have 15 dichotomy variables which measure if a participant accepts an object or not. For example item 1: do you accept object 1? (yes; no), ... item 15: do you accept object 15? (yes; no).
Now I would like to if it's possible to calculate the sum of all 15 dichotomy variables as a metric variable?
In my opinion, it's possible because the variable has a neutral point and the distance between 1 to 2 points is the same to 5 to 6 points. That means the criteria for a metric variable is satisfied.
Is this solution sustainable?
I think you should apply the Principal Component Analysis in order to get a composite index which you can use as a dependent or exogenous variables in your models.
You can look at our attitudes towards mental illness-scales
- Badges
- Science topic