In my fitted model i.e Determinants of Participation in CBHI , the Pseudo R-squared is 0.24 by using STATA package is it satisfactory ? or what shall I do to increase this value? I have used both continuous and categorical independent variables , does this matters for low or high Pseudo R-square?
This link may help you:
https://stats.idre.ucla.edu/other/mult-pkg/faq/general/faq-what-are-pseudo-r-squareds/
In my opinion, that R-square value is low, and indicates a bad fit. However, the use of R-square in logistic regression is problematic and requires others tests to better assess the model fit. I would recommend the use of the Hosmer-Lemeshow goodness of fit test when creating logic models. The test can be found in the following text:
Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied Logistic Regression, Third Edition. Hoboken: John Wiley & Sons, Inc.
When combined with assessment methods such ROC analysis you begin to see the quality of the model. The attached file demonstrates the test for several login models. If your model is not fitting well there several options. The first is to assess if all of your variables are significant. There may be a case for the use of "dummy" variables or even discretization of your continuous variables if fit issues remain. Finally you can examine variable interaction and test that interaction for significance (with care to avoid over fitting your model).
Most studies in the case of increasing the number of independent variables lead to an increase R-squared
Dear: One of the ways to increase the coefficient of determination is to choose the variables that affect the phenomenon under study, increase the number of independent variables, correct diagnosis of the problem, increase the size of the sample, or not choose the sample accurately.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing