IBM SPSS Modeler Cookbook

Using an @NULL multiple Derive to explore missing data

With great regularity the mere presence or absence of data in the input variable tells you a great deal. Dates are a classic example. Suppose LastDateRented_HorrorCategory is NULL. Does that mean that the value is unknown? Perhaps we should replace it with the average date of the horror movie renters? Please don't! Obviously, if the data is complete, the failure to find Jane Renter in the horror movie rental transactions much more likely means that she did not rent a horror movie. This is such a classic scenario you will want a series of simple tricks to deal with this type of missing data efficiently so that when the situation calls for it you can easily create NULL flag variables for dozens ...