Estimation of a Logistic Regression Model to Determine the Effects of the Factors Associated with the Likelihood of Skilled Workers in the Garment Sector of Bangladesh (Published)
This study attempts to estimate the logistic regression model to find the factors associated with the likelihood of skilled workers in the garment sector of Bangladesh. The study is based on primary data. From the maximum likelihood estimates of the logistic model, it is found that the variables, years of education, years of experience, wage rate, grade, and working years at the present company have significant positive effects on the logit of the outcome variable Y, but the variable age of the companies has a significant negative effect. From the estimated odd ratios, it can be said that the male group is 3.0695 times more likely than the female group to be a skilled worker, which is significant at a 10% level; a group of rural workers is 10.3727 times more likely than the urban group in favour of skilled workers, which is significant at a 15% level; a group of workers have a training program degree is 38.5552 times more likely in favour of skilled workers, than a group of workers who have not, which is statistically significant at any significance level. The estimated results of squared correlation between Y and , pseudo-R2, Mcfadden’s , Maddala’s , and Count-R2 support that the model fits the data very well. Based on the findings, different policies are formulated for the sustainable development of Bangladesh’s garment sector.
Keywords: Fit very well., Logistic regression model, Maximum Likelihood method, Primary data, pseudo-R2