Syntax

QRS.LOGISTIC.​REGRESSION(​X, y, [option name], [option value], …)

X[[number]]

Matrix containing the values of the independent variables with rows corresponding to observations and columns corresponding to variables.

y[number]

Column vector containing the values of the binary dependent variable.

option name/valueoptional

The name or value of an option. See the Options section for details.

Options

• LABELS: TRUE to include labels in the output.
• NAMES: Vector containing the name of the n^th independent variable in the n^th cell to include the names in the output.
• MASK: Vector containing TRUE in the n^th cell to indicate that the n^th independent variable may be selected.
• METHOD: String to use a variable selection method. None is used by default. The possible values are:
  • FORWARD for forward selection. Starting without any variable, the one with the smallest p-value less than PGREEN is selected recursively.
  • BACKWARD for backward selection. Starting with all variables, the one with the largest p-value greater than PRED is eliminated recursively.
  • STEPWISE for stepwise selection. Similar to forward selection, except backward selection is applied after each variable is selected.
• LRTEST: String of comma-separated flags to include likelihood ratio test results in the output. The possible flags are:
  • D for test statistic.
  • P for p-value.
  • RAG for red/amber/green rating.
  • ALL for all of the above.
• WALD: String of comma-separated flags to include Wald test results in the output. The possible flags are:
  • W for test statistic.
  • Z for z-score.
  • P for p-value.
  • SE for standard error.
  • RAG for red/amber/green rating.
  • ALL for all of the above.
• PGREEN: Number between 0% and 100%. A p-value less than this number corresponds to a green rating. Defaults to 5%.
• PRED: Number between the value of PGREEN and 100%. A p-value greater than this number corresponds to a red rating. Defaults to 10%.
• TRANSPOSE: TRUE to swap the rows and columns of the output.

Remarks

• Blank cells are read as zeros and dates are read as numbers due to Excel limitations.
• The length of y must be the same as the row length of X.
• The row length of X must be greater than the column length of X.
• The p-values for the variable selection methods are based on the likelihood ratio test, rather than the Wald test.
• The likelihood ratio test and Wald test are for the null hypothesis that the coefficient of a variable is zero.

Examples

• How to perform logistic regression in Excel using only a formula.