Standard practice would be to check the initial frequency count first, but Pallant goes direct to optional statistics from frequencies to find her cutting points and summary statistics, and produces a frequency count almost by default. Age itself may be a continuous variable, but the variable recorded here is age last birthday. Pallant’s calculated mean is actually inaccurate: for a more accurate mean you should really add 0.5, but the cutting points remain the same, even after this adjustment.
Frequencies…
Statistics…
In the Percentile Values panel, click on Cut points for ….
NB Pallant has missed a step out here as she doesn’t show how to get the required statistics: are they automatic? No, but they’re produced in the summary table: