Your math is fine.
Spec's on the units web site show variable input power 0.7 to 3.6A @ 12V and 1-2 a.hr/hr typical.
So, your 3.4*0.21= 0.7 sounds low. Measurement errors?
Here's a ballpark check:
Heat flow out of the box is area times temperature difference divided by resistance of insulation
So from the manufacturers site, the interior
surface area of the box is 12.2 ft2.
Your temperature difference is 84-38=47F
A guess at the insulation
- say 3 inches of R7 gives R21 (very optimistic, could be R10)
efficiency should be 40-60 percent. Let be generous and say 60%.
So heat flow = 12.2 x 47 / 21 = 27.3 btu/hr or 8 watts or .67 A @12V. (3.412 btu/hr is 1 W )
But at 60% eff. you will use 1.1 amps.
So, with these very optimistic calculations, I think you are light.
Maybe take more measurements? Is your system voltage 12V or 13.8V?