I summed up five linkert variables measuring satisfaction with life scifea+,scifeb, scifec+, scifed+, scifee...I want to run a bivariate regression with the sum of these fiver linkert variables as the dependent variabel, and a continous variable like time spent caring for older adults as independent variable, does the sum of the linkert variables maintain being a categorical variable so i run a logistic linear regression or does it become a continous variable so i could run a sumple linear regression?
The sum of variables in Likert scale is categorical. However, it can be approximated by a continuous variable.
The quality of a continuous approximation depends on the number of levels and the variability. For example, if the 5 variables are assessed on 11-point scale 0, 1, 2, ...10 with some variability, continuity may be a good approximation. If all variables are = 4 or 5, the sum is equal to 20, 21, ...,25, for which the approximation is not very good.
the usual practice is to use factor analysis first, save factors variables and use the in the logic model
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running