Epi Info™ User Guide
Command Reference
Analysis Commands: REGRESS
Description
This command performs multivariate linear regression with support for automatic dummy variables and multiple interactions. If a variable has more than two values, it is automatically turned into a series of ‘yes/no’ variables called ‘dummy variables’, one for each extra value.
Syntax
REGRESS <dependent variable> = <independent variable(s)> [NOINTERCEPT] [OUTTABLE=<tablename>] [WEIGHTVAR=<weight variable>] [PVALUE=<PValue>]
- <dependent variable> represents the dependent variable.
- <independent variable(s)> is an independent variable that can be a numeric variable, a non-numeric variable, or a variable surrounded by parenthesis. Any text or yes-no variable, or a variable surrounded with parenthesis, is automatically recoded into dummy variables. A dummy variable is created for all but one of the levels of a variable. This dummy variable measures the contribution of its level relative to the excluded level. Interactions are specified by * between variables.
- <weight variable> represents a variable describing each data row’s contribution to the regression
- <tablename> is the table to store the residuals. If no table name is present, no residuals are produced.
- <PValue> represents the size of the confidence intervals; may be specified as percent or decimal; if greater than .5, 1-PValue is used. The default is 95%.
Comments
REGRESS uses the least-squares method for determining coefficients.
Example
READ {.\Projects\Sample\Sample.prj}:BabyBloodPressure REGRESS SystolicBlood = AgeInDays Birthweight