I'm a big advocate of the digital spedometers. I'm always leery about taking a top-speed claim at face value unless it has been verified either with a car (as Wayne has done) or with a digital spedometer.
I also find that the spedometers don't vary by a constant percentage.
Ex: One of my Peugeots stock spedometer is accurate up to about 40km/h (0% error). When it says 55km/h I'm really going about 50km/h (10% error). When it says 70km/h I'm really going about 60km/h (15% error).
If you don't want to add a digital spedometer, you might want to consider <i>temporarily</i> putting one on until you learn what speeds your moped is really going. As an alternative you can verify with a car at different speeds. You can also do the one mile time trials as Don suggested. Once you know how your spedometer behaves you can reliably translate the speed into "real" units.
As a point of interest: Does anyone out there have a stock spedometer that actually gives you a <b>conservative</b> (as opposed to inflated) reading? Broken spedometers that always say "0" don't count, of course!