Reducing the sensitivity of a PIR sensor?

You can attenuate by using thin sheets of polyethylene. Polyethylene is the same material that the PIR lens is made of. It will pass, but attenuate the wavelengths you are interested in. Find something around .015" and begin stacking until you hit your desired range.

Here is a source: http://www.mcmaster.com/#polyethylene-plastic-sheets/=hc4uvj

You can either do this physically using an IR opaque material as Jason suggests, or electronically by modifying the sensor.

Most (all?) of these small PIR detectors seem to be based around a BISS0001 chip. These have two amplifier stages for taking the signal from the PIR FET. If you take a look at the example circuit on page 4, most of the detectors I have seen follow this almost exactly, down to the component labels being the same.

The resistors don't always seem to have the same values on the boards. But by simple observation, you can see that for opamp OP1, it is operating in non-inverting mode. The gain is R7/R8, 2M/47K or approximately 20x.

OP2 is in inverting mode and the gain is R5/R6, 1M/10k or approximately 100x.

I would adjust the values of R7 and R8 to reduce the gain and make the device less sensitive.

Alternatively, you could adjust R1 (connected to the source of the PIR detector) or try reducing Vref (which sets several of the voltage references inside the chip).

All of these require removing and adjusting components though, so maybe the physical way is easier.