0
$\begingroup$

In a study, we have the disease as dependent variable ( 0 or 1 patient or healthy); some independent variables are real and some of them are confounding (some of varibles are quantitative and some of them are 0/1 like smoking); i am using spss for analysis; which of these should be used? GLM--->univariate or GLM----->multivariate or on of regression subdivisions? What about custom models? It seems meaning of these words in spss is different from them in statistics books? Why in multivariate it shoulde be two dependent variables? I read some researchs using multivariate analysis while they had only one dependent variable.

$\endgroup$
0
$\begingroup$

If your dependent variable is dichotomous, you should use the binary logistic regression.

Regarding the independent variables to use, you can use both categorical and quantitative variables. However, regarding the independent variables, you should check for outliers and avoid using collinear variables (variables strongly related with each other). In SPSS, don't forget to mark as categorical the corresponding variables.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you so much $\endgroup$ – user241033 Mar 14 '19 at 22:30
  • $\begingroup$ @Slow You are welcome, By the way, if you believe that my answer solves your problem, you are welcome to accept it. Of course that you can also wait for other helpful answers before accepting one. $\endgroup$ – Ertxiem - reinstate Monica Mar 14 '19 at 22:34
  • $\begingroup$ Thank you @Ertxiem; it was helpfull, i found the green color means "accepted". $\endgroup$ – user241033 Mar 15 '19 at 20:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy