I have a linear regression with a dependent variable (revenue) explained by a dummy coded (0-1) variable (charging ticketing fee no:0 yes:1) and a time variable (date format). The dummy variable (ticketing fee) is used to indicate the date from which a ticketing fee was charged. Since the ticketing fee has been charged since a certain date, the dummy variable changes only once, before a certain date: 0 and after a certain date:1. (This is a very simplified version of the real model without taking into consideration of trend, cycle, AR-component, etc.)
Since the influence of taking the ticketing fee might change with the time, I would like to determine the interaction between the dummy variable and the time variable. Would that be possible by building the product of the two variables and include it into the linear model? I could not find a solution in SPSS Reg. Models for interactive terms.
Example:
input program.
loop time =1 to 10000 by 1.
end case.
end loop.
end file.
end input program.
EXECUTE.
DO IF time < 5000.
COMPUTE ticket = 0.
ELSE.
COMPUTE ticket = 1.
END IF.
EXECUTE.
COMPUTE revenue = RV.NORMAL(1000,10).
COMPUTE ticket_by_time = ticket * time.
EXECUTE.
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT revenue
/METHOD=ENTER time ticket ticket_by_time.
Dr. Frank Gaeth
Locked
