Business Statistics 202
Regression and Correlation Analysis Project
1. This project is worth 15%.
2. The majority of the calculation and data analysis will be done using MegaStat, SPSS or Minitab or other software.
3. The penalty for a late project is 3 marks per day. Your mark for the project will be zero after 5 working days. You are, however, still required to hand in a satisfactory project. This component is compulsory. You will receive incomplete grade for the course if you fail to comply. This Incomplete grade will automatically change to Failure after three months by the Registrar’s office if you have not fulfilled all the course requirements.
4. Each project must be unique and original…it must be your own effort. Plagiarized work is not acceptable. Please refer to the College directive regarding plagiarism.
5. The project will not be returned to you, so please keep a copy for yourself before handing it in.
Regression and Correlation Project Particulars
You are required to conduct a complete multiple regression and correlation analysis using the Least Squares model.
You should decide on the dependent variable (Y) and three related independent variables (X1, X2 and X3). For example, you may wish to predict $ Computer Sales (Y) based on Advertising Budgets (X1), Number of Sales Outlets (X2). Number of Employees(X3)You should then collect information on several brands of Computer Sales (Y), their respective Advertising Budgets (X1), Number of Sales Outlets (X2) Number of Employees(X3),
1. Obtain data from WDI (World Development Indicators) or other source. Save the data.
2. You will need a minimum of 20 observations, preferably 25 observations if possible for each of the variables.
3. You are required to provide the MegaStat output using Excel, SPSS or Minitab (i.e. included with your project).
5. The report must be word-processed.
Page 2 of 3
The report must reflect the minimum standard of report writing techniques, meaning that you will be responsible for grammar, including proper spelling and sentence construction.
Grades such as A and A+ are awarded to students who demonstrate superior research skills.
The project should be presented with the following sections clearly indicated. Please provide all of the following information in the exact order outlined below, with each number below forming a section in your report.
Do not make your reports longer than necessary; a well written report will provide all of the required information accurately and clearly with as little words as possible.
1. Table of contents
The table of contents should list, in chronological order, the important topics covered in the project along with page numbers. This will make it easy for the reader to “walk through” the report by referring to topics that are of interest to him or her.
This section should explain to the reader the intent of the report. You should define your dependent and independent variables. If possible, provide background information on each of the 4 variables. This will give the reader some insight into your research. Explain why you decided to study these variables.
3. Correlation Matrix
Based on the output describe the relationship between each of the variables and the dependent variable. Also, be on a look out for multicolinearity between the independent variables.
Stepwise Regression:Eliminate the worst independent variable and explain how you decided to eliminate this variable, making reference to your MegaStat output. Print out this output to help explain and support your decision
Page 3 of 3
4. Multiple Regression Analysis
With the remaining 2 independent variables perform the following analysis:
o Obtain the least squares regression equation from your Excel output.
o Interpret the partial regression coefficients in terms of the dependent and
o Determine and explain the confidence intervals of each slope.
o Interpret the meaning of the standard error of the model. Also, interpret
the standard error of each slope
o Indicate how the significance F value is calculated in the ANOVA table.
o Select the highest value from each of your independent variables and
make a point prediction, confidence interval and prediction interval.
Interpret each prediction. Be careful to explain the difference between
o Perform the global test to determine the overall validity of your regression
model. You must use proper test procedures.
o Test for significant contribution of each of the independent variables.
5. Assumptions and testing
List the multiple regression assumptions.
Do you think the independent variables that you have chosen are valid variables? What factors could be considered to improve the multiple regression models.
You must identify your references – any sources that you have used for this report (periodicals, etc.)
Appendices should include the following:
o Excel output from MegaStat, or SPSS, Minitab or other software
o Excel output of correlation matrix with 4 variables.
o Summary, ANOVA and Predictions.
o Supporting documents if any. Relevant sentences and/or paragraphs must be highlighted, as I will not be reading the entire article.