Question: Perform a multiple regression analysis using business data of your choice from the Internet, the library, or your company and write up the results as a report to upper-level management, either as a background summary report or as a proposal for action. You should have a significant F test and at least one significant t-test (so that you will be able to make some strong conclusions in your project). Your report should include five to seven pages plus an appendix and should be based on the following format:
a. Introduction: Describe the background and questions of interest and the data set clearly as if to an intelligent person who knows nothing about the details of the situation.
b. Analysis and Methods: Analyze the data, presenting displays and results, explaining as you go along. Consider including some of each of the following:
(1) Explore the data using histograms or box plots for each variable and using scatterplots for each pair of variables.
(2) Use a transformation (such as the logarithm) only if this would clearly help the analysis by dealing with a big problem in the diagnostic plot.
(3) Compute the correlation of each pair of variables and interpret these values.
(4) Report the multiple linear regressions to predict one variable (chosen appropriately) from the others by explaining the regression equation and interpreting each regression coefficient. Comment on the quality of the regression analysis in terms of both prediction accuracy (standard error of estimate) and how well the relationship is explained (coefficient of determination). Report statistical significance using p-values both overall (F test) and for each regression coefficient (t-tests). In particular, are your results reasonable?
c. Conclusion and summary: What has this analysis told you about the situation? How have your questions been answered? What have you learned?
d. Appendix: List the data, with their source indicated. (This does not count toward the page limit.)