Hey, I have binary dependent variables and around 12 independent variables..
Now for 4 categorical variables (independent), I have more than two levels, for which I have created dummy variables.
Now I don't know what all variables I should enter in SPSS to run binary logistic regression, everything in one covariate or different ?
(all I have known is that usually for dummy variable we don't feed the reference category (for which no variable is made) to assess the odds)
so is it the same way here ?
Thanks..
Why are you not using the CATEGORICAL sub-command to handle categorical explanatory variables? Here is an example from the Command Syntax Reference manual:
LOGISTIC REGRESSION
VARIABLES = PASS WITH GPA, GRE, MAT, CLASS, TEACHER /CATEGORICAL = CLASS,TEACHER.
https://www.ibm.com/support/knowledgecenter/en/SSLVMB_25.0.0/statistics_reference_project_ddita/spss/regression/syn_logistic_regression_categorical.html
Given that you have 12 explanatory variables, you'll also need to have a large number of events to avoid over-fitting your model. See Mike Babyak's nice article (link below) for details. HTH.
https://www.cs.vu.nl/~eliens/sg/local/theory/overfitting.pdf
Article What You See May Not Be What You Get: A Brief, Nontechnical ...
Hey Bruce, thank you.
I'll go through the links, I'll get back to you in case of queries.
I will recommend this textbook (A Gentle Introduction to Stata by Alan C. Acock).
A Gentle Introduction to Stata by Acock is a very good book for people wanting to learn Stata. But Riddhi is using SPSS. ;-)
Hey Bruce ! Thanks for sharing the information.
I understand that this categorical function can be utilized. But my categorical data is ordered and my reference group is at second level.
So I guess I should denote reference group as “0” only, and assign the variable as “nominal”?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS