On the Use of Logistic Regression Model and its Comparison with Log-binomial Regression Model in the Analysis of Poverty Data of Nepal

Background: Previous literatures have indicated that log-binomial regression model is an alternative for the logistic frequent occurrence of event outcome. The comparison performance these two models has been found with reference to clinical/epidemiological data. Nonetheless, application and its poverty data not described. 
 Objective: To compare in terms variable selection, effect size, precision goodness fit, diagnostics, stability model, issue failure convergence.
 Materials Methods: Cross sectional 5988 households Nepal Living Standard Survey 2010/11 used analysis. compared each covariate, fit using Hosmer - Lemeshow (H-L) test, diagnostics bootstrapping method, Results: Logistic overestimates yields wider 95% confidence interval than log binomial covariate. greater elevation risk covariates varies from 13% 173%. satisfies (p = 0.534), tests, model. However, grossly violates 0.0004) but criteria.
 Conclusion: Log-binomial all criteria development except gross violation including On basis entire performance, better fitted fitting set Nepal.