Logistic regression is a variation of ordinary regression which is used when the dependent (response) variable is a dichotomous variable.  A dichotomous variable takes only two values, which typically represents the occurrence or nonoccurrence of some outcome event and are usually coded as 0 or 1 (success).

The example dataset below was taken from the well-known Boston housing dataset.  The information in this dataset was gathered by the US Census Bureau from census tracts within the Boston area.  Each of the features (or variables) describes a characteristic impacting the selling price of a house.

Logistic Regression Example Dataset

To run a logistic regression:

  1.  On the XLMiner Analysis ToolPak pane, click Logistic Regression
  2. Enter C1:C40 for Input Y Range.  This is the output variable.   
  3. Enter A1:B40 for Input X Range.  These are the predictor variables. 
  4. Keep "Labels" selected since the first row contains labels describing the contents of each column. 
  5. If "Constant is Zero" is selected, there will be no constant term in the equation.  Leave this option unchecked for this example.
  6. Select Confidence Level 95%.
  7. Enter F1 for the Output Range.
  8. Click OK.

Logistic Regression Pane 

The results are below.

Logistic Regression Results

Using these results, the regression model can be written as:  Median Value of Owner Occupied Home = 0.98 – 29.68 * CRIM – 1.484 * ZN.  For a more detailed explanation of these results, see any standard statistics reference text.