On the face of it, your statement should be correct.
Extensive testing with a variety of different reactors and resistance heaters, including bare resistance heaters, shows that my statement is correct. I am not making this stuff up. I am not speculating. I am telling you what the data shows from three air flow calorimeters (two from Mizuno and one from Saito), and from many water flow calorimeters.
I have some concerns about the exterior heating method, since it is directly heating the air as much as, if not more than, the reactor or dummy.
Why would this matter? It is not clear to me what "directly heating" means in this context, but whatever it means, how can it transfer more energy over time? I can see why it might heat up the air more quickly, if it emerges from the reactor sooner, or if there is less thermal mass. But over the total course of the reaction the heat balance will be exactly the same no matter with any heater. The heater does not measurably affect the inlet or outlet thermocouples. They are not directly exposed to the calibration heater or the active reactor.
Anything that increases the outlet temperature will appear as increased power, all else being the same.
How can any configuration of the heater increase the outlet temperature? This would not "appear" as increased power; it would be increased power. It would be a perpetual motion machine. The terminal outlet temperature must be the same at a given power level for any heater, with any geometry. If it is not the same, I suppose the outlet thermocouple is directly exposed to the heat source, or there is some other problem. The whole point of calibrating with different kinds of heaters is to ensure there is no problem of this nature. I have never seen a problem of this nature, but anyway, if there were one, calibrations with different heaters should reveal it. If you don't think so, what test do you propose to show there is "direct heating"? How can we detect this?
(Some people here -- not you -- have a bad habit of suggesting there might be thus-and-such a problem, but when I ask: "Okay, how would you detect this problem? What test will reveal it?" they do not respond. That makes their assertion not falsifiable. Meaning it is not true or false. If you cannot describe a test to demonstrate a problem, that problem does not exist. It is imaginary.)