"physics said it shouldn't work but tests showed a small effect with widely varying results"
A valid point, and why I phrased it as I did. If you claim that Wifi gives you a headache in, say, 5 minutes, then that's a strong signal that can be tested for. If it fails, then there's a lot of statistical power there. If you claimed that being in a Wifi field for a few days gave you a headache, then my answer is... actually that you're very unlikely to be correct because carefully testing such a problem requires individual trials that take weeks apiece, which you have almost certainly not done. I know from experience; I've done some of my own personal health optimization that takes similar time frames. Some of my results have required years to determine because of the insensitivity of the testing and how long a single trial takes to get significance, and I still wouldn't be my life I've got all the details correct.