I want to compare the awarness of patients who are on anti-coagulation toward anti-coagulation treatment/prophylaxis based on many demographic variables,such as education, gender, comorbidities, personal and family history of venous thromboemolism...
when I need to use multinomial logistic regression ?
is there specific guidance for such analysis using SPSS?
thank in advance
Abdulsalam Halboup
For your problem, I am suggesting you perform "Ordinal Regression".
Because, the outcome of your dependent variable will be ordinally categorized e.g. minimal awareness, moderate awareness, etc.
You could perform "Multinomial LR" if you had nominally categorized dependent variable i.e. causes of minimal or no awareness, causes of venous thromboembolism, etc.
Thanks
Multinomial models are for situations in which several alternative outcomes can occur, of which one is a natural baseline outcome. For example, in pregnancy the woman may have normal blood pressure (baseline) or may either develop gestational hypertension or pre-eclampsia (but not both – that's important).
It allows us to compare the predictors of each outcome versus the baseline outcome.
Abdulsalam Halboup is right. It's not what you need, based on what you've told us.
A suitable regression model is selected for our analysis based on the type of the dependent variable and therefore its distribution. As we know data on-hand may be numerical (discrete/continuous) or categorical (nominal/ordinal). In the multinomial logistic regression model the dependent variable is nominal categorical variable with more than two categories.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS