Hi all,

I have some difficulties regarding interpretation of categorical variables when some of categories , not all categories are statistically significant in Binary logit model. To explain my question more clearly see the following categorical variable used in my study:

Outcome variable: Enrollment in CBHI scheme( enrolled=1, non-enrolled=0)

Some of categorical variables that I have used are as follows:

1. Age of household head(4 categories)

18-30(reference)

31-40 (odds ratio =0.3512 ; P =0.033**) ** shows signficant at 5%

41-50 (odds ratio= 0.7847; P= 0.604)

above 50 (odds ratio=0.4838; P=0.099)

2. Educational status of household head(5 categories)

illiterate( reference)

read and write( odds ratio=28.68; P=0.000***) *** shows significant at 1%

primary (odds ratio= 3.194408 ; P= 0.040**)

secondary (odds ratio= 5.164823 ; P=0.007***)

tertiary(odds ratio= 1.312708; P=0.677)

In addition to this , which way of interpretation is best( is it interpretating Odds ratio or Marginal effect?) in Logistic regression.

Any answer is appreciated

Regards!

Similar questions and discussions