You are continuing to confuse the concepts of power and current. These are two very different concepts.
For a given p-n junction, the heat transfer is proportional to the current through the junction. That is the mechanism by which the Peltier Effect works. However, most Peltier devices are made up of several junctions. Also, different Peltier devices may be made from different materials, so you cannot compare one to the other in the manner you suggest. If you measure the applied current vs. the change in temperature for the same device, then the relationship will be linear (or at least close).
When you look at the spec sheet for your device, does it say the following:
= 38 W
I would be very surprised if it did.
Chemists are physicists who don't do math.
