You email the entire assignment package your tas

2. Don’t try to print the codebook. Instead, when you need original wording of survey questions, use search function (Ctrl+F) to locate variable names (variable names are unique) in the Acrobat Reader;

3. You need to copy-and-paste your Stata outputs to your report (right below each question) and remember to change the font in Word to Courier New (size 9) so that columns and numbers of tables can be vertically aligned;

PART II Requirements:

• A brief explanation of all procedures used;

PART II [50 Points]

B1. Inter-generational social mobility, or say reproduction of social classes, is an important topic for sociological studies. This topic can be studied by investigating earnings of fathers and their adult sons. One hundred father-son pairs were randomly selected from the labor market of Shanghai, China. The following table summarizes correlation, means, and standard deviations of earnings of fathers and sons (unit: RMB 1,000/Month):

1. Complete the procedure of an F-test for your regression model. You need to write down null [1 point] and alternative [1 point] hypotheses in English and then fill up the ANOVA table [9 point]. Go to textbook appendix to find F critical value based on alpha=0.05 and your degree of freedoms [1 point]. Compare your F statistic and F critical value and draw your conclusions [3 point]:

2. First, obtain the unstandardized regression coefficient (b) [2 point], the constant (a) [2 point], and the standardized regression coefficient (β*) [1 point]; then perform a t-test for regression slope b (set alpha=0.05 and critical values to be +-1.96). [3 point] Explain the effect of father’s earning on son’s earning. [2 point]

B3. Studies of gender inequality usually relate personal annual income to respondent’s gender. But gender inequality is an issue that is far more complicated than such a simple thought. To improve our understanding of gender inequality in the U.S., let’s consider “differentiated rewards to human capital between genders”. Use the GSS 2012 data and set realrinc (respondent’s annual income) as the dependent variable. Set educ (respondent’s educational attainment as a measure of human capital) and sex (respondent’s gender) as independent variables. Pay attention, educ is a continuous variable but sex is a dichotomous variable. Based on Stata outputs, complete following three regression models (you need to obtain slopes, constants, as well as adjusted R-squared to fill in each open white cell in the following table). Model 1 shows gender difference in income; Model 2 illustrates human capital effect on income; and Model 3 presents differentiated rewards to human capital between genders. If a slope or constant is statistically significant in a t-test when alpha=0.05, put a “*” right next to the number (see constant of Model 1 as an example). [1 point for each open white cell]

Among these three models, which one has the strongest explanation power (or say which one fits to the observed data best)? Why? [2 points] How do you conclude gender inequality in the U.S. using these models (remember to cite numbers from three models to support your conclusions)? [3 points]