PROC TABULATE by Example, Second Edition

Number of Observations Seems Too Small

Your code runs fine with no error messages, the table has the rows and columns that you wanted, and the values are formatted correctly, but the table is still wrong. What do you do when your table shows a number of observations that is much smaller than what you expected?

Most of the time, when you produce a table, you know how many observations to expect. In fact, one of the best ways to double-check your work is to include the number of observations in your table, at least for the first draft. After you have reviewed the table and the number of observations looks okay, you can always choose to delete the N statistic from the final report. In the following example, the N statistic is used to check the ...