Reference no: EM132271383
Homework 1 -
Tests the following hypotheses with the first replicate of SCF data and interpret the results. Unless explicitly requested to do otherwise, please set the significance level (alpha) of the tests to 0.05.
1. Two independent samples t-test. (Rely on NETWORTH)
H0: The subjective expected living horizon (X7381) of millionaires is the same as non-millionaires. (Hint: define a household as "millionaires" if their net worth is over $1 million.)
H1: The subjective expected living horizon of millionaires is not the same as non-millionaires.
2. Chi-square test. (Test of independence. Two categorical variables.) The SCF provides information on subjectively-evaluated health status (X6030) and expectation how the economy will perform in the next year (X7489). Use the chi square test procedure to test whether or not the optimism regarding performance of national economy is independent of health status. (Observed - Expected)^2 / Expected.
3. Conduct the ANOVA analysis of the expected living horizon (X7381) by expectation how the economy will perform in the next year. Do you reject the hypothesis that the average subjective life expectancy differs by the subjective economic outlook at alpha=0.01 level? (One must be dependent (continuous...mean)) Independent (group).
If the test indicates that subjective life expectancy differs by subjective economic outlook, identify which specific paired groups are significantly different from each other at the alpha = 0.01 level. Use the Tukey's test in your ANOVA.
Write a short report with your findings. Your report should look like a section of a publishable paper, i.e., any tables or figures included in your report should be formatted as if you were submitting a paper for publication (it is not ok to just randomly copy & paste output from SAS). Please include a table with descriptive statistics for all variables used.
Homework 2 -
Identify four continuous variables in the SCF. Designate one of these variables to be the dependent variable and the remaining three to be the independent variables.
1. Is there a conceptual/logical reason why the dependent variable should be correlated with the other three variables? Explain briefly.
Dependent variable - Net worth, Independent variables - Income, Age (X14), Education
2. Examine the correlations between all variables.
3. Run the simple least squares regression model of the dependent variable on the independent variable with which it is correlated the most. Examine and interpret the coefficient estimates and model fit statistic, R2.
4. Run the multivariate regression model by including the remaining independent variables. Interpret the results from the multivariate model and examine how the model fit statistic changed.
Write a short report with your findings. Your report should look like a section of a publishable paper, i.e., any tables or figures included in your report should be formatted as if you were submitting a paper for publication (it is not ok to just randomly copy & paste output from SAS). Please include a table with descriptive statistics for all variables used.
Homework 3 -
Return to your regression from the last homework.
1) Identify at least two categorical variables in the 2016 SCF data that are likely related to the dependent variable. One of the categorical variables can be a dummy with only two levels, but at least one of the categorical variables should have three or more levels (for example, education coded using separate dummies that indicate high school, some college, college, etc.). Explain what relationship you expect to see between the dependent variable and the categorical variables.
2) Modify your regression model from previous homework by including the categorical variables. Interpret all coefficient estimates (including significance tests of the coefficient estimates) and the overall model fit statistic R2. You can choose to have the binary variables either indicator- or effects-coded. Also, examine the plot of residuals vs. predicted values and identify/discuss any problematic patterns. Calculate the VIFs and examine if excessive correlation among independent variables might be a problem to statistical inference.
Write a short report with your findings. Your report should look like a section of a publishable paper, i.e., any tables or figures included in your report should be formatted as if you were submitting a paper for publication (it is not ok to just randomly copy & paste output from SAS). Please include a table with descriptive statistics for all variables used.
Homework 4 -
Identify a binary variable in the SCF and estimate the logistic regression model to measure the relationship between the likelihood of the outcome measured by your binary variable and a few independent variables (please include both continuous and categorical independent variables). Is there a conceptual reason why the dependent variable should be correlated with independent variables? Interpret the results of your logistic regression (both the magnitude of effects and their statistical significance). Finally, measure the predictive accuracy of your model by calculating and reporting the Pseudo R2 and the percentage of correctly predicted outcomes.
Write a short report with your findings. Your report should look like a section of a publishable paper, i.e., any tables or figures included in your report should be formatted as if you were submitting a paper for publication (it is not ok to just randomly copy & paste output from SAS). Your report should also include a table with descriptive statistics for all variables used in the analysis (please report both weighted and un-weighted means).
Homework 5 -
The SCF asks respondents about their expectations about the future. In particular, it asks if individuals think that the economy next year will perform better, worse, or about the same as today (variable X7489). Estimate the multinomial logistic regression and the ordinal logistic regression with the intention of testing whether or not the total family income affects the expectation for future. Include in your model the basic demographic control variables such as age, gender, education, marital status, race, and household size. Evaluate the predictive accuracy of both models by calculating the percentage of correct predictions and provide full interpretation of results of the estimated model that performs better in predicting the outcome.
Write a short report with your findings. Your report should look like a section of a publishable paper, i.e., any tables or figures included in your report should be formatted as if you were submitting a paper for publication (it is not ok to just randomly copy & paste output from SAS). Your report should also include a table with descriptive statistics for all variables used in the analysis (please report both weighted and un-weighted means).
Attachment:- Data File.rar