More data!
I repeated the test with the 300w external block heater. This is a bolt-on element style unit like this:
Results:
Note: I didn't display the "temp before starting the engine" in this graph because the readings were all within a couple of degrees of the "after 30 seconds of idling" values. That's probably because the location of the heater element is much closer to the coolant temp sensor.
In fact, more often than not, the temp dropped a few degrees after start-up (which tells me the warm spot in the water jacket created by the heater was close to the temp sensor).
It's not shown on the graph, but out of curiousity, I left the block heater plugged in for another hour after it reached 101 degrees F. When I checked it after the hour (no idling during that time), the indicated temperature had actually dropped a five degrees. That suggests the heat generated by idling the engine for 30 seconds every 15 minutes to get stable temp readings was skewing things by about 5 degrees at the end of the test.
Reposting the 800w chart for easy comparison (less scrolling!)