Types of Regression Models - Assignment Example

?Econometrics Regression Model Question A The other hypothesis to be investigated by this regression was weather the level of education of women had an impact on their level of availability or participation in the work place. This was to be monitored using the coefficient educ. The availability and increase of other sources of income for the family was also considered to find out weather the same affected their level of commitment at work. This was done using the variable nwifeinc. The estimated coefficients had the signs as expected and were of statistical significance, this was demonstrated clearly buy the patterns displayed by the respective graphs of the variable coefficients. It therefore demonstrated beyond any reasonable doubt that the availability of married females in the work place was directly proportional to the size of their families. This could be explained by the competition for time brought about by the existence of families, Such that the woman had to make a choice between taking care of the family and coming to work. Question B The estimated logit model had a lot of resemblance to the LMP in terms of statistical significance and estimated coefficients. The few deviation (variations) were identified in cases where the females involved had some sought of secondary education. Weather this was as a result of better time management skills learnt in school or not is a subject of debate which can be investigated further but what is certainly true is that the level of education had an impact on the availability of married females at the work place. Question C An increase in a married female’s education had a substantial impact on the estimated logit model. The deviation from the logit model caused by an increase in female’s education was up to a maximum of three and a minimum of one. This was very evident from the resultant patterns plotted by the graphs. This confirmed the previous hypothesis that the level of education of a married female had a significant impact on their availability in the work place. The amount of money earned by other family members in the family was found to have a marginal impact on the logit model but the impact was not of statistical significance. It therefore appeared as though the amount of time the women put in to their work did not depend on the availability or increase of other sources of income brought into the family by other members of the family. Regression Model 2, Question D The Tobit model was to investigate weather the various variables had an impact on the amount time married women spent at work. The main variable to be investigated here was the size of the families that these females were responsible over. The model and the associated variable focuses on the availability of married females in the labor force, it was expected that their level of availability would be dependant on the size of their families indicated by the number of children they had. Such that, the more the children, the bigger the family thus less participation at the work place. On the logit model, the inlf coefficient was therefore expected to be inversely proportional to the kid_s and kid_m coefficients. The expected signs were realized and they were of statistical significance. The coefficients demonstrated the effect of large families on the amount of hours married females spent at work. It was noted that women with relatively smaller families spent more time at work than those with larger families. Question E There was statistically significant difference in the estimated coefficients between the Tobit and the OLS estimations which were very much expected. The differences were evident I both the signs they plotted and the magnitude of those signs. This led to the confirmation of the previously stated hypotheses namely, that is increased education and reduced families enhanced the availability of married females at work. The other hypothesis was that increase in alternative incomes by other family members had a negative impact on the availability of mothers at work. The later hypothesis was nullified since the deviation plotted by the coefficient (nwifeinc) was not statistically different from the Logit model. Thereby asserting that the presence or increase in sources of income by other members of the family had no bearing on the amount of hours married women put into their work. Regression Model 3, Question F The regression model three was meant to evaluate the effect of the explanatory variables namely experience, hours worked and education on the amount of wage offered to married female workers. It was expected that the three variables put together would have a significant positive impact on the amount of wages married females earned from their work. The hypothesis was that the amount of wages earned by married women is directly proportional to their level of education, experience and the number of hours they worked. There were significant similarities between the MLE of model 3 and the signs demonstrated by the education and experience coefficients. Question G This indicated a confirmation of the hypotheses put forward for those two variables. But there was statistically significant difference between the MLE of model 3 and the signs plotted by the coefficient indicating the number of hours worked. It therefore follows that while it’s obvious that increased education and experience increased the wage earned, the same could not be said of hours worked. Working long hours was not a definite guarantee that a married female would earn more, more often that not they would earn the same they would have earned had they worked for only the prescribed hours. Question H A likelihood ratio test was done to find out if a sum of the coefficients on exper, exper2 and age were equal to zero in model three. The Null hypothesis was that they are not equal to zero thereby meaning that there is significance difference between exper and exper2. The alternative hypothesis was that a sum of the three equals to zero thereby meaning that a sum of exper and exper2 is inversely proportional to the age. The test confirmed the later hypothesis. Read More