Q1:
1. The data for this (and all other) assignment is located on blackboard. You should save a copy to a flat ASCII (space delimited) file and a formatted ASCII (comma delimited) file. This gives you three copies of the data.
2. You are to use the IMPORT command to read EACH of the three datasets into SAS. Confirm that the data are correct.
3. You are to write a DATA step that correctly reads the two ASCII files into a SAS dataset.
Q2:
1. You are to create dummy variables for each month and each day of the week. Make sure that these are functioning properly. Also create a log variable for the two temperature variables.
2. Create a deviation from mean for the temperature variables.
3. Create squared terms for each of the temperature variables in deviation from mean form.
4. You are to get the summary statistics from the dataset and save the output file as a .lst file. Send me both the .sas and the .lst files.
Q3:
1. Do a regression of sales on all of the relevant variables using the linear versions of the temperatures. Be sure to remove any variables that create multicolinearity.
2. Do the same regression as in 1) above but using the log versions of the temperatures.
3. Do the same regressions as in 1) and 2) above but in each case do a separate regression for each month of the year.
Q4
You are to take the linear regression from Q3 and perform the following tests.
1. Test each coefficient against the null hypothesis that it is 0.
2. Examine the coefficient for rain and test the null hypothesis that it is -1500.
3. Examine the coefficient for Temperature and test the null hypothesis that it is 38.
4. Examine the coefficient for Min Temperature and test the null hypothesis that it is 20.
5. Examine the coefficient for Holiday and test the null hypothesis that it is less than 2000.
You are to turn in the SAS program and a short WORD document that describes the results completely. Also, include a complete description of the regression from Q3.
Q5
1. Carry out a White test test for heteroskedasticity.
2. Carry out a Durbin-Watson test for serial correlation.
3. Test your data for multi-colinearity in both fashions that you know how to do this test.
You are to turn in the SAS program and a short WORD document that describes the results of your tests.
Q6
1. Test the null hypothesis that day of week is not important for predicting sales.
2. Test the null hypothesis that month of year is not important for predicting sales.
3. Test the null hypothesis that the squared temperature variables are not important for predicting sales.
You are to turn in the SAS program and a short WORD document that describes the results of your tests.
Attachment:- data.xls