Regression: ASCII data

Run|Regression and Tools|Regression provide a process for building a regression model by successively adding fields (variables).

1. Add one Dependent field

Enter the record/column location and width for the Dependent field (for example 1/7:2).

2. Add one or more Independent fields

Enter the record/column location and width for one or more Independent fields (for example 1/8:2).

3. Choose the Display options:

Show means

Creates a Mean Summary section of the Regression report. This section reports the Variable Name, Variable Label, Variable Number (a sequence number for the selected Independent variables), Mean, Standard Deviation and Correlation value.

Show correlation matrix

Displays the Correlation Matrix section of the Regression report.

Show model summary

Displays the Model Summary section of the Regression report for each step of the regression unless the Show only the final step option is selected. This section reports the Step Number, Variable Number (a sequence number for the selected Independent variables), Variable Name, Variable Label, Regression Coefficient, Standard Error of Regression Coefficient, Computed T-Value, P value, R-Squared value, Adjusted R-Squared value and % Contribution.

Show only the final step

Displays the Model Summary section of the Regression for only the last step in the regression. The Model Summary section reports the Step Number, Variable Number (a sequence number for the selected Independent variables), Variable Name, Variable Label, Regression Coefficient, Standard Error of Regression Coefficient, Computed T-Value, P value, R-Squared value, Adjusted R-Squared value and % Contribution.

Show variable labels

Displays the Variable Label for all of the variables selected in all sections (if displaying) of the Regression report.

4. Choose the option for calculating % contribution:

Use R-squared

Uses the R-squared value for calculating the % Contribution in the Model Summary step(s) of the regression.

Use adjusted R-squared

Uses the Adjusted R-squared value for calculating the % Contribution in the Model Summary step(s) of the regression.

5. Choose the option for missing values:

Skip record

Uses the Skip record option when there are missing values for any of the selected variables. The record(s) with missing values is excluded from processing. This is also referred to as Listwise Deletion.

Mean fill

Uses the Mean fill value when there are missing values for any of the selected variables. The Mean fill is calculated and used to replace all missing values for the selected variables.

6. Choose one of the Stepwise options (mutually exclusive):

Use T-score to limit steps

Uses the T-score to limit steps of the regression.

Use F-value to limit steps

Uses the F-value entered here to limit steps of the regression. The WinCross default for F-value is 3.96.

Stop when coefficient is negative

Halts the regression process when a regression coefficient is negative.

Run all steps

The Run all steps option processes all steps of the Regression. This is the equivalent of a Linear Regression.

7. Run Filter (optional)

Enter a Filter Title (optional) for your Regression processing.
Enter Filter Logic to be applied to your Regression processing. This logic limits the records included in the regression processing to those that match the Filter Logic.
The Run Filter is active check box can be used to remove the Run Filter without removing the Filter Title and Filter Logic that was previously entered. This allows you to reuse the Filter Title and Filter Logic if desired.

8. Run or Cancel the Regression selections:

Run

Choose Run to process the Regression.
Saves the regression specification as dataname.wro to the folder where the open data file currently resides.
If dataname.wro already exists, it will automatically be replaced.

Cancel

Choose Cancel to cancel the Regression and close the Regression dialog.

Restore Defaults

Resets all Regression settings to the WinCross defaults.