Sorry? A >£15,000 Spectrascan PR650 spectroradiometer with an integrating cosine receiver dome is NOT a test at all?
I'm afraid I got interrupted whilst I was typing that earlier reply. I have done several things.
One was to take spectral and illumination readings using the PR650, which showed there was a sufficiently high level of light around the 450nm mark to bleach rods completely. This compares to the ~590nm of the low-pressure sodium lights that were previously fitted, though I didn't cart over £15k worth of equipment 20 miles on the train to make readings of street lamps at the time the LPS lamps were still operating - it didn't seem worth it. At 590nm, the absorption of photons by human rhodopsin is practically nil.
I brought the expensive gear out after I made a first set of measurements where I got someone to walk into the park outside the flat holding the end of a groundsman's measuring tape until I could no longer make out their outline against the grass whilst viewing them from the path on the far side of the road. I then did the same in a different park about a mile away where the streetlights were still the low pressure sodium type. Under the LED light, it was two metres, under LPS lighting, it was forty metres. Not a controlled experiment, I know, but what the heck.