MGS 495 Business Forecasting                                                      2/23/2004

v     Subjects

¡P        Perform Linear Regression using SPSS for Windows

¡P        Calculate skewness and kurtosis for residuals

v     Procedures

¡P        Input

¡P        Process

¡P        Output

v     Data Input

¡P        What Format to use? Depends on your own convenience and objective

¡P        How to Input Data?

¡P        Windows Excel

STEP 1: Input Format

STEP 2: Open Excel file from SPSS

Go to File> Open> Data.

Things to do:

(1) Select data type as ¡§All Files [*.*]¡¨

(2) Check ¡§Read variable names from the first row of data¡¨

¡P        SPSS Data Editor ( as shown above)

v     Data Process

¡P        Regression Analysis ( here we only demonstrate Model 1)

¡P        Perform multiple regression analysis with (Yt) as the dependent variable and (X1t), (X2t), and (X3t) as the independent variable (predictors).

¡P        Go to Analyze>Regression>Linear ¡K..

• Under Statistics, click Descriptive. Under Options, make sure you have clicked Exclude cases listwise.

• Practice Model 2 and 3 on your own.

v     Output ( The outputs are for 3 models we have discussed in class)

• Model 1 (Dependent Variable: Yt. Predicators: X1t, X2t, and X3t).
• Model 2 (Dependent Variable: Yt. Predicators: X1t, and X3t).
• Model 3 (Dependent Variable: Yt. Predicators: X2t only.

v     How to calculate skewness and kurtosis of  for model 1, 2, and 3 (Assuming that you have derived results for all models)?

• STEP 1: Calculate   and (Definition: ) using the estimates of regression coefficient derived from SPSS for each model. (do the calculation by using EXCEL, and then open the file via SPSS)

Model 1:

Model 2:

Model 3:

• STEP 2: Calculate skewness and kurtosis of  for model 1, 2, and 3.

Go to Analyze> Descriptive Statistics> Descriptive¡K.

Pick ethm1, ethm2, and ethm3 as Variables to calculate skewness and kurtosis.

Under Options, make sure skewness and kurtosis are checked.

• STEP 3: Output