I was questioning their proven effectiveness based on the "cleanliness" of the study. If that factor clouded the analysis for LED vs. incandescent, or red vs. amber, then I assumed that a similar limitation could apply to CHMSLs
That's a reasonable question. I don't think there's much danger that yesteryear's CHMSL study results have lost applicability. The range of difference in terms of size, shape, position and photometry is much larger for the right/left stop lights than it is for the CHMSL. There is some thought that a higher-mounted CHMSL mnight be better because it is less likely to be blocked by a large vehicle, but as far as I know, there hasn't been a study done.
I think there is a bigger "upshot" from your line of thinking, though: it looks to me like the CHMSL mandate was based on a much smaller, less realistic data set than the data set which NHTSA seems to think is insufficient to mandate amber rear turn signals. Safety regulations are as much about politics as they are about safety.
To me, the novelty of a third brake light was what was most attention grabbing.
Well, yes, to repeat myself, they were more effective when there was a novelty effect, but that was not the entirety of the effect (they remain effective crash-avoidance devices even though pretty much all vehicles have them now).
That brings to mind a more general question -- of all the various factors involved in signal lighting, what has the most influence? Photometry? Position? State change (i.e-flashing)? Something else?
Ah, that we know. At short range (dense traffic, from bumper-to-bumper to something like 30 or 40 feet), it's luminance. At long range (anything above about 30 to 40 feet) it's intensity. Luminance is how bright the lamp appears, expressed in candelas per square meter. Intensity is how much light the lamp is emitting, expressed in candelas. So if we have to name just one factor that is the prime factor in the safety performance of a vehicle signal light, it's intensity. Luminance would be next on the list. However, as always, there are caveats. Luminance distribution is also important. In the past, with simple optics that all had the lens as the light dispersing surface, luminance was fairly even across the whole lamp area. You might have a bright spot in the middle (direct view of the filament) and there might be some other minor variation owing to lamp geometry or design, but not a lot. Then came "jewel" optic reflectors and "window" optics-free lenses, and while some of them give a very even luminance, many of them do not.
It is common now to see lamps that look like a few very bright lines or spots on a dark field when viewed at close range. That might or might not be creating a safety problem. I certainly find it an annoyance in traffic sometimes. One offender is the Chrysler/Dodge minivans of the design that started in 2008. More recent ones have been coming from the Kia-Hyundai group. There's one in particular, I don't recall which (maybe the Sonata hybrid?) that has a small, very bright central spot surrounded by a large area of much lower intensity. The first few times I saw these, I thought they had been tampered with ("LED bulb" retrofit or similar), but no, it's factory equipment. The central bright spot is painful to look at if you're stuck behind one of these cars in traffic. Obviously it meets the legal standards, because the surrounding area is lit enough to count towards the EPLLA (effective lit area, basically) requirements, but there's no requirement for evenness of luminance and there's no limit on peak luminance. There's no control of luminance at all, in fact.
Anyhow, back on track: below the #1 factor (intensity) and the #2 factor (luminance), then we get to effects such as drivers reacting more quickly and accurately to stop lights on a vehicle with amber rear turn signals than on a vehicle with red rear turn signals, or on a vehicle with stop lamps that are body-color when off. Both of those studies were done by UMTRI.
I think I read somewhere that older drivers aren't attuned as much to the illumination of a brake light, but rather the act/rise of illumination itself; a state change
I'm not aware of any data suggesting anything like this, but it's common for manufacturers and vendors of dubious lighting equipment (devices that flash/blink/"pulse" the CHMSL, for example) to make such claims. If they bother backing up the claims with anything even vaguely scientific, it's always tiny little snips taken far out of context and not at all representative of whatever study they were originally a part of, or else these tiny snips are misrepresented as to what they mean.
One of my pet peeves is combination functions. I prefer each function to have its own distinct position. Off the top of my head, the tail fixtures on the early Chrysler minivans was probably the worst. A reverse light, and one red light for combined stop/tail/signal.
Actually, a few things here: combination stop/tail/turn lamps are widely common in North America where red rear turn signals are permitted. Most American-spec Chrysler minivans over the years have had such lamps, true, but they're among thousands of other year/make/models that have had this same type of lamp. Also, if the rear turn signal is red, I prefer it to be combined with the stop light. Separate red lamps (one for stop, one for turn) are confusing at best. At worst, the turn signal is effectively not visible when the stop lamp is lit. Most of this could be solved if there was a minimum separation distance required between stop lights and red turn signals, but there isn't. An amber (yellow) turn signal is better in every way (you might have read
this article on the subject already).
I thought it was an odd situation as well. The fixture is a NAL (nee AL) CHMSL.
No, NAL and AL have no connection to each other. They are entirely separate companies. NAL is North American Lighting, a wholly-owned subsidiary of Koito, which is a Japanese supplier of vehicle lighting and related equipment. AL is Automotive Lighting, the company formed when Bosch and Magneti-Marelli merged their car lighting business activities in 1999.