Michael Gregersen wrote: > > Is it possible to make SPSS show the variable/value labels in say a > crosstab, when you dont have any cases matching the values. In the > output you just have 'blanks' or zeros. My problem is, that I only get > an output of the variables and values with corresponding data.

[snip] -- It is not possible to make CROSSTABS show rows or columns for values which do not occur in the data. To show the labels you want, you need Fry's Horrible Hack No. 23 (FHH23).

Procedure for FHH23:

1. If your data do not include a weighing variable, make one.

COMPUTE WGT = 1. WEIGHT BY WGT.

This step will not change the results of any procedure.

2. Add some cases to your data set which have the following properties:

a. They all have valid values for all the variable involved in the CROSSTABS you want to fix, so they are not discarded for missing values.

b. They "complete" each variable, in the sense that with these new cases included, all the values you want to appear in the table are in the data set.

c. Their values for the weighting variable are positive, but very small, say 1e-20. Cases with weiths this small will have no effect on summary counts or statistics, except where the number of rows or columns is used in calculating a statistic.

3. Run the CROSSTABS.

Thw simplest way to add the cases is to use the data editor with your original file. But if you plan to run the same report many times with changing data, then it will be worth your while to prepare a separate file with only the add-in cases. Then you can use the ADD FILES command to add the cases to each new data set before running your reports.

FHH23 works for graphs as well as tables. You can use it to add an axis label for a category that did not show up in the data. -------------------- Jonathan Fry SPSS Inc. jon@spss.com