Hello everyone,
I'm undertaking a study on household welfare using a composite measure of household access or ownership of basic amenities or assets. The individual variables that formed the index variable are dichotomous, taking the values 0 and 1.
I will appreciate your recommendations.
Rear @Ebuka Orjiakor, I think Logistic Regression (LR) is commonly used for binary dependent variable. Please also find the link below.
https://www.sciencedirect.com/topics/computer-science/binary-logistic-regression#:~:text=3.2%20Data%20analysis%20methods&text=Binary%20logistic%20regression%20(LR)%20is,or%20not%20readmitted%20(0).
What is key here are the properties of your composite/index variable. Is it the sum of multiple dichotomous variables? Is it a count variable? Can the composite/index variable be seen as continuous (metrical/interval scale)? Is it symmetrically distributed or other (e.g., zero-inflated)? Depending on the answer to these questions you can select the proper regression model.
As I understood the question, there are several, say k, "individual variables", which are Bernoulli variables (support is (0,1)). These are summed up or averaged into the "index variable" which is the dependent or response variable Y in the model. If these are summed, then Y has support (0, 1, ..., k) and could be modelled as a binomial variable with parameter p (0
Thank you everyone for your responses. Many thanks @Christian Geiser.
I apologize for not asking a comprehensive question.
The dependent variable is an index variable constructed from 10 dichotomous variables. The values of the index range from 0 to 1.
Thanks for your forth coming response.
Logistic Regression (LR) is commonly used for binary dependent variable. So go ahead with logistic regression.
- Badges
- Science topic
- Similar topics
- Indexes