You may be right that with the needle in the carb it might tend to be cooler because of the forward airflow (not preheated by the cylinder) but also in this (the old ) configuration vaporization is taking place right at the spray bar, where it should, even if caused by heat. Also all that vaporization at that point tends to cool the shaft, shaft housing and bearings and carb. The needle hanging off the back doesn't get this benefit.
I tend to run my engines a little over rich and as a result may have not had problems running in the air although I've certainly had bubbles. Also the few degrees difference in running temperatures caused by tight or loose parts fit could be the straw that causes a cascade of riseing temperature and boiling alcohol.
This could be why some people seem to be having problems with this engine going lean and quiting after a couple of minutes of flight. Smaller props and higher rpm will also help with this kind of problem. I'm currently running an ultrathrust muffler and when under proped just slightly it tends to go rich in the air.

It could probably be noted that most engines with remote needles don't seem to have this problem. But most engines use a bolt on needle assembly, I think, rather than one that is an integral part of the back plate casting. This means there is one more thermal junction impeding the heat getting to the needle housing. On the OS91 that housing also looks like a nice little pot for boiling alcohol. It's also suspended out there where vibration can be amplified for agitation.
Of course none of this is supported by good scientific testing supported by high dollar test equipment and well documented numerical results that have been scrutinized by other disinterested experts, and are therefore not conclusive.

It is conclusive enough for me to cut the suckers loose and mount them truly remote though.