Diagnostic software with live data (so, Alfadiag, or FIATECUScan I presume) will show the Lambda voltage switching about once per second low-high-low-high etc. if the Lambda sensor is working and the mixture is correct.
Problem is, the Lambda sensor is really the last reading after many factors that affect the fuel/air ratio. In other words, if there is any fault in the injection system (faulty fuel pressure regulator, blocked fuel filter, air intake leak, etc.) then that can cause a lean (or rich) condition and the output of the Lambda sensor will be 'jammed' low or high (not switching) but that does NOT necessarily mean the sensor is faulty - it is just reading the condition that results from everything else. It is even possible for an exhaust leak to affect the Lambda reading.
Therefore, if a Lambda sensor is reading correctly, everything is probably OK, but otherwise you should probably investigate for other faults before replacing Lambda sensors. On the CF3 engine (2002) there are several Lambda sensors - would they all fail at once? Perhaps if one is stuck and the other two are switching, it's safe to assume one has failed.
Unless Smaky knows something I don't about these sensors (common failure?), which is very possible.
How's the engine temperature?
So many of these cars have the thermostat stuck open (so the temperature gauge needle hardly moves) - that will affect economy.
Paying out a little money to have someone check everything out with electronic diagnostics and decide on the required actions should pay for itself in fuel savings pretty quickly (unlike the magnets, which won't pay for themselves
Meanwhile I managed 1000km on one tank, in open-road driving. That's 620 miles. Smaky's 510 miles is more realistic for most conditions, though. And driving in town makes a huge dent in the economy.