Standby power consumption, budget PSU (CX430)

I got a CX430 for a project last week and I'm getting some strange readings from my "kill-a-watt" meter. In standby, either installed in the system or unplugged, the PSU draws 25 W. Stranger still, when you flip the switch on the back powering it off, the PSU still consumes about 15 W (fluctuating between 13 and 16 actually).

The kill-a-watt meter seems quite accurate on other PSUs, bulbs, monitors and what have you. I have an older HX520 PSU that consumes 9 W installed and in standby.

Any ideas?

I should add that I asked the same question over at Overclock.net. They seemed to have a long CX-series thread going. And they were quick to dismiss this as a kill-a-watt issue, claiming that the active PFC design on this PSU could be making it impossible to measure the consumption using consumer level equipment.

What is your take on this?

I have read many stories of people getting weird readings on APFC PSUs while under heavy load, making it seem as though the PSU is operating at over 100% efficiency and such. But I have read nothing about how this would affect standby readings or how common this is. Or how it could show a power draw while it is physically powered off. Is there some type of circuitry that precedes the on-off switch?

Tested a new VX550 and the results were equally or more erratic. 15 W switched off. 0 - 3 W powered on. The conclusion has to be that at least this kill-a-watt meter is useless for measuring modern APFC PSUs. Case closed