I have worked with categorical variables and i reviewed many source related to categorical variables. However, there is no explanation how to select reference baseline. Please provide me some guideline
I suppose you are talking about choosing reference category of your categorical variable when doing regression.
Choose a category appropriate for your interpretation and/or your scientific aims. There is no rule for that.
Yes Mehmet is right. There is no rule for baseline selection of categorical variable.
For example if you have 4 categories : 1, 2, 3, 4.
and category 3 is interpreting your aim then you could make 3 as your baseline reference point and compare categories 3:1, 3:2 , 3:4.
1. if preventive factor for negative outcomes the preventive factor become reference.
2. the reference group should be well defined (avoid like others)
3. reference should have sufficient cases to provide precise estimate of sub group mean
Actually, there are no rules but it is better to select or choose the category with a large number of observations to narrow the confidence interval and minimize the standard error.
I get difficult in interpreting results from logistic regression especially when there is no effect of exposure on outcome variable, can someone please help?
You can choose arbitrarily. It depends on the aim of the study. If you are studying some behavior of citizens in a country then it will be reasonable to choose the national level as the benchmark to which other ethnic groups will be compared.
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation