this work uses Stata software

You must create a table of descriptive statistics, and you must explore meaningful relationships between variables in the data. Think of this in terms of hypotheses between the dependent variable and the independent variables. Demonstrating a simple relationship between these variables motivates the research into the overall t opic as well as justifying the use of multivariate regression. You are required to demonstrate such simple relationships use graphs , but you should also use simple statistical methods - not regression - to verify the existence of these relationships. Since t he data contains many variables, you should only explore relationships for the most important variables, or only a few of the variables that are included in the regressions. Your table of descriptive statistics should not include variables that are not use d in the regression analysis. Think carefully about what you are doing and try to come up with hypotheses regarding the relationships between the variables. Do not fall into the trap of letting the data dictate how you think about the relevance of the var iables. For the regressions that must be completed for this exercise: - you are required to present all the regressions in one table; - your table of regression results must include the adjusted r-squared for each regression - your table of regression results must include the standard errors for the coefficients - once you add new variables to the regression you must state whether the coefficients on the other variables have changed; that is, make a statement about the omission of important variables in the simp ler regressions. Please note that you are required to consider the statistical significance of variables when interpreting coefficients. However , don’t forget that we are far from being able to talk about causality in our models. You are required to su bmit the do file: - The do file must include all the necessary Stata code to complete the tasks laid out above. - The do file must also be sufficiently annotated so that there is no ambiguity about the tasks being completed. - For all variables that were gener ated, you must carefully label them. - I strongly recommend that you also label the values within the variables . - Your do file must run without any errors and must create a log file. You are also required to submit a word document. However, instead of answering the questions or providing a one -sentence description of a figure, you are now required to write up your answers as a brief essay. This brief essay must now include only two sec tions: a data description section where you present the descriptive statistics table, one figure, and one non -regression statistical test ; and, a data analysis section where you present the regression results table and discuss your findings . Note that the data description section should include a brief statement about the data itself. 2 The data for this question comes from the Annual Social and Economic (ASEC) supplement of the Current Population Survey, 2019 . The variables in this data are identical to the ASEC data you have seen in-class and on previous assignments . 1. Clean the data , as per our previous use of this data. Generate variables necessary to answer the questions below . 2. Estimate the following model . This is the “original” model refer enced in parts (c), (d), and (f): log (ℎ )= 0+ 1 + 2 + 3 + 4 + where hrwage is hourly wage, eduyears is years of education, exp is years of experience , married indicates whether the individual is married or not , and black indicates if the individual is black or not . a. Interpret the coefficients . Note that this is specifically asking you to interpret the coefficients, even on the dummy variables. b. What is the predicted h ourly wage for someone who is black? c. Add experience squared to the original model. Does the inclusion of this variable improve the model? Interpret the relationship between experience and hourly wage. What is the predicted turning point in the relationship between experience and log(hourly wage)? d. Now extend the original model to allow the return to education to depend on being black . Does the inclusion of this variable improve the model? e. Now extend the original model to allow wages to differ across four groups of people: married and black, married and nonblack, single and black, and single and nonblack. What is the estimated wage differential between married blacks and married nonblacks? f. Is the use of the variable black sufficient for capturing the relationship bet ween race and wages? g. We know that gender is an important variable in explaining wages, however it is excluded from our regression. How does the exclusion of this variable affect the interpretation of our results? h. Now extend the original model to allow the return to education to depend on being female . Does the inclusion of this variable (s) improve the model? i. Based on your regression to part ( h), what is the predicted wage for someone who is male , has 21 years of education, 1 2 years of ex perience, is not married, and is not black ? How many people in the sample match th ese specific characteristics? j. Based on your regression to part ( h), what is the wage differential (i.e., change in wage, or difference in wage) for someone who is female, has 16 years of education, zero years of experience, is not married, an d is black ? How many people in the sample match th ese specific characteristics? k. Now rescale the hourly wage variable so that it become s weekly wage . To do so create a new vari able t hat converts the hour ly wage vari able to weekly wages using the mean of usual hours worked per week . Use the log of weekly wages as the depen dent variable in the model you spe cifi ed in part (h) and r un the regression. What impact did rescaling the dependent variable have on your results ? Do you think tha t this was a useful change to your model?