I don't know where the conclusion that there's less capability for cooling comes from, they are vented underneath. Certainly it would be expected as with any solar cell that it will be less efficient the hotter it gets. I don't know if anyone has any data on what temperature a typical racked panel gets to on a hot/sunny day, I'd be able to take surface thermocouple, contactless IR thermometer, or IR camera shot of my roof (as well as either of those two latter things work on a reflective surface), though getting good correlation to a panel would still be an issue unless it was right down the street, even for my town I'm in an unusually hot area, temps in my neighborhood are generally 5-15 degrees hotter than the 'official' San Jose temperature on any typical summer day. There does appear to be some degradation on hotter days, though of course there are many other possible variables such as haziness, inverter temperature, etc.
Last week here there were a few "clear" but increasing temperature days, if you look at my PVOutput data (link in signature):
5/23 (73.7 kWh) the peak temperature in my back yard was 93.9.
5/24 (71.0 kWh) the peak temperature in my back yard was 100.9.
5/25 (68.5 kWh) the peak temperature in my back yard was 108.5.
5/26 (67.4 kWh) the peak temperature in my back yard was 112.8.
5/27 (68.9 kWh) the peak temperature in my back yard was 110.3.
5/28 (69.7 kWh) the peak temperature in my back yard was 100.9.
[Yeah, last week was not fun, considering that my house has no AC and uninsulated walls.
Good news is the attic doesn't get as hot since the Solar Roof was installed (presumably due to that vent layer), so that has seemed to help the peak house temps a bit]
So 5/24 and 5/28 both had the same locally-measured high temperature (and same average wind according to my weather station also), and both curves on PVOutput are smooth (no clouds, etc), but the production is still ~2% different (and the sun should still be getting higher/longer each day, so 5/28 should have been the better day from that perspective). Clearly temperature isn't the only thing having an impact here, so I find it difficult to draw any real conclusion from the above data, the impact could have more been because the hotter days were hazier, for example.