Receiver spurs can come from many sources, the most notorious of which is mixers. They can also come from sidebands on local oscillators, power supply conducted signals, circuits picking up radiated signals, magnetic coupling, even mechanical vibration.
But the one everybody worries about are mixer spurs, because they are present by design and if they are there, no amount of shielding or filtering will get rid of them. They are caused by the mixing process itself being less than ideal - a perfect mixer would take in the RF and LO frequencies and output only an IF at RF-LO and RF+LO frequencies. But of course nothing is ideal and you get lots of other garbage out. Distortion in the mixer, particularly on the LO signal give you (M times RF frequency)+/-(N times LO frequency) products essentially to infinite frequency, although most people only worry about spurs with M and N both less than 10th order. Manufacturers will sometimes give you a table of spur levels for a given set of frequencies and power levels, but often, you're on your own.
Here's an example. The objective is to make sure that for the LO frequency you want to use to mix your DESIRED RF down to your IF (say at 70 MHz), nothing else also mixes down to your 70 MHz IF. So With a 928 MHz receiver the LO could be at RF+IF or 998 MHz. Let's look at the 3x3 mixer spur on see where it lies.
We know the LO and IF frequency, so where would the spur be?
(3 x unknown RF)+/-(3 x 998 MHz LO) = 70 MHz IF
so the 3x3 spur can be at either 1021.3333 or 974.666
And if you measured this, you would find the receiver would receive a signal at these frequencies, although not as well as the desired frequency, because the mixer has a lot more conversion loss at higher order conversions.
So why doesn't this make the receiver unusable with all these hundreds of spurious response frequencies? Preselection.
Those dielectric and SAW filters before the mixer prevent stuff outside your desired frequency from ever getting to the mixer. Stuff still leaks through the filter, but unless you cram a big signal into the receiver front end, the mixer never sees it and causes little problem.
This receiver is relatively simple in this regard because it only receives a narrow, easily filtered band of frequencies. The problem gets much tougher as the frequency band to be received is wider, say extended to 1000 MHz in this case so the 974.666 MHz spur was in band and cannot be filtered. Then this story gets longer.
I don't think your receiver has any mixer spurs less than at least 10th order. See the great engineering you never knew was in there...
Check out http://www.hittite.com/spurchart/SpurCalculator.htm
This is an automated version of a classic mixer spur chart. It takes a little getting used to, but it's the standard tool for avoiding mixer spurs.