If you've got actual lasers and know they're spectrally pure, don't waste time with color filters.
With a nice, tight beam I'd be particularly interested in the difference between illuminating the cathode versus the anode. You're likely to get some reflection either way so I doubt either will give you a zero reading, but you should see a significant difference. RCA clear tops sound like a good choice. However, the tube glass might attenuate or even block UV depending on exactly what sort of glass it is, so be ready for it.
Can you vary the laser peak power, or just total power (i.e., pulse width)?
I'd probably start with a 1 kHz square wave, and monitor the control signal and the output simultaneously using a storage oscilloscope. Should answer the basic question of cause very quickly.
If the effect is definitely there, the question would be what the efficiency of the system is -- how many volts out versus how much power in. Should be totally linear with respect to pulse width, more-or-less linear with peak intensity. If you have a strong enough laser you might even hit a saturation point where the cathode is fully illuminated and can't spit out any more electrons, but I'm guessing this would take a LOT of light.
Using a Foreplay I'm guessing the volume control will have no effect on the light-induced output signal, but you should probably check this too.
