What is the best statistical test on SPSS to predict the effect of more than two independent categorical variables on a (one) dependent categorical variable?
*Socio-demographic characteristics of the subjects on their practice.
Dear Alkalmi
For your data you can do regression analysis for predicting the effect of dependent variable over Independent variables in SPSS. For testing you can use F-test where the hypothesis will be all the independent variables are equal over At lest one independent variable is not equal.
Thank You.
Dear in this situation you should use Loglinear analysis. For detailed understanding of this technique you can consult book of Agresti on categorical data analysis.
Would like to present my thanks and appreciation for all of you. May ask one question, what if the proposed test failed to find any sig. correlation (i.e, p>0.05) .even there is interrelation between the independent variables ?
Dr. Sami can you justify using of MLR instead of LR? (all the items binary measures)
Ramadan,
Look at the work on Hierarchical Multiple Regression. Also the recent work of Prof Andrew Hayes in Conditional Process Analysis which now appears in several journal articles and his PROCESS macro which is Hayes' add-in for SPSS and SAS. The references below illuminate:
http://www.processmacro.org/index.html
https://scholar.google.com/scholar?cites=2148876485119216290&as_sdt=5,36&sciodt=0,36&hl=en
http://www.afhayes.com/
http://www.ats.ucla.edu/stat/spss/seminars/process_macro/process_training.htm
Silburn
You may use logistic regression (binary or ordinal or multinomial depending on the nature of the dependent variable). For logistic regression the independent variables may be categorical or scale variables, but remember to identify the categorical independent variables as such ( as factors). CHAID analysis could also work provided you have enough cases.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS