The answer depends on the type of variables you have to use as criterion variables. If you work with a bivariate framework, the pearson correlation coefficient could be a good choice with continuous and normally distributed variables; or a spearman rank correlation if you use non-normal or categorical variables. In a multivariate framework I suggest you to use Structural Equation Modeling (SEM) to test different theoretical relationships between MBI factors and other variables. If you prefer regression analyses, the model depends on the response variable. For example, if the response variable is discrete (count variable), a poisson regression model must be applied. But if the response variable is continuous and normal distributed, you should use a linear regression model. Other example could be a logistic regression model, which must be applied with dychotomous response variables (0-1).
I hope this information could help you.
All the best.
That is very helpful; in non-anonymous surveys, with respect to burnout, variables may not be normally distributed given responses sometimes may be skewed. If one wanted to keep the burnout scale as a continuous variable; would a linear regression model still be applicable?
The problem is not the burnout variable, that I assume that it is continuous. The question is which type of variables you use as criterion variables. For example, if you use job satisfaction or organizational climate, and these two variables are continuous, then a linear regression model is OK. But if the criterion variables are not continuous, you must change the regression model (logistic, poisson, multinomial, etc.).
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress