I'm the one who did the test mentioned above. In case you wonder, let me explain how the test was done.
I put the compass (a Suunto sighting compass) on the edge of my wooden kitchen table (and away from any metal), so that I could look into the eyepiece without touching it, and I took a reading. Then I put the helmet on my head (the one with the Apex and Stenlight above it) and took another reading. I did not make deliberate efforts to get the Stenlight particularly close to the compass, nor did I try to keep it far: I was just trying to get a reading much like I would in a cave. Depending on the tilt on my head, I could easily get 2 degrees off the original reading. Just to make sure that the compass had not moved, I took the helmet off and took another reading. It had returned to withing 0.25 degrees of my original reading. (A small tap on the table got it to return to the original reading... the friction on the pivot was too much to get it to turn on its own.) I repeated the process described above 3 more times, getting the same results each time.
If you try this test, remember that the effect on the compass might not be seen, depending on the direction in which the compass is pointing. This is because any metal (or magnet) interfering will pull the north end of the compass towards or away from it. If this happens to be the same direction as magnetic north, then it won't make any difference! For my test, I happened to be looking west.
During the test, the Apex was turned off, but the Stenlight was turned on (to its lowest setting)... just in case the (small) magnetic field caused by the flow of electricity would make a difference. I really wanted to simulate a realistic situation.
Now, this test does not prove that the Stenlight is the culprit here, as the Apex was closer to the compass, but given that both lights were permanently mounded on the helmet (a Petzl Ecrin Roc if I recall), I didn't have much choice.
I understand that having a magnetic switch makes for a good separation between the inside of the lamp and the outside world, but this is not a good idea for surveying! Having a switch (on the "outside") that is eventually replaceable by the end-user should it ever die would have been a much better idea.
I must admit that before that test I was seriously considering buying a Stenlight, but I an now looking elsewhere. Too bad because the size of the Stenlight made it more interesting than its competitors.
Another user of the Stenlight had previously told me that he always took his helmet off when reading instruments (I presume that's done to shine light on the instrument). Might be ok in an horizontal cave, but I won't be doing this in the middle of a deep pit, or whenever there are people above me who could be knocking down rocks... did you ever try to get a good compass reading while receiving rocks on your head?
The search for the perfect light continues... but in the meantime I'd be interested in hearing others' opinion on this.