undefined | Better HN

0 pointsschiffern1y ago0 comments

  >If... your headache flares up precisely when the wifi router's lights blinks, but not when it is actually on... then that's pretty conclusive...

...of nothing, because that's also how the body will react to known stressors.

A recovering alcoholic can feel sick at the mere smell of alcohol (simultaneous stimuli), but that's not "pretty conclusive" evidence that drinking alcohol has no deleterious effects.

That "experiment" is only evidence of one thing: that you can trick people using Psych 101 combined with bad experimental design. If you wanted to actually prove or disprove the effect, a properly designed experiment would omit the trickery (or at least run a matrix experiment with a trickery/no-trickery variable).

0 comments

3 comments · 2 top-level

jerf1y ago· 1 in thread

"...of nothing, because that's also how the body will react to known stressors."

You say that like it's some sort of brilliant revelation, but that they are reacting to some other stressor and not actually to Wifi is the point.

Since the question under study is not "do people react to perceptions" but "does this person actually react to Wifi signals, specifically Wifi signals and not anything else", it is quite to the point. If Wifi gives you a headache immediately, which is to say, a strong, relatively quick signal, but you can't detect it reliably in a contained environment, then you're wrong about the cause of your headaches, which, as I said, is not even remotely a shocking result. Headaches do not come with labels attached telling you what caused them, unfortunately.

schiffernOP1y ago

  >You say that like it's some sort of brilliant revelation, but that they are reacting to some other stressor and not actually to Wifi is the point.

The "and not actually" clause used above is unscientific and not causally connected. That's precisely my point.

Whether or not they're reacting to other stressors doesn't tell you whether or not they're "actually" reacting to WiFi. It only confuses the matter.

A proper experimental design would disable the lights and just have the RF signal be on/off, then you can cleanly measure the effect. With the lights, your measurement variable is invariably distorted by overlaying a known (placebo) signal.

(saying the placebo effect is non-zero and measurable isn't a "brilliant revelation" either, so you'll probably poo-poo me again based on a claim I never made, but c'est la vie :-P )

The presence of the lights is completely unnecessary. We already know people will react to placebos. It doesn't tell us anything scientific, it only serves to confuse the participants and reduce the number of useful trials measuring the actual variable. Adding lights has negative scientific value.

Real science is about eliminating confounding variables (see Young's rat maze experiments[1]), not throwing in extra ones solely so you can later spin a better "gocha" narrative in the public press. -_-

Perhaps the experimental design in the original paper was much better than this game-of-telephone retelling suggests, but lacking a citation (or even an authorship hint) it's hard to really know.

[1] https://gwern.net/maze

nerdponx1y ago

It would be conclusive if the smell of alcohol consistently made people sick but actually drinking (magically deodorized) alcohol had no apparent effect on the same people.

j / k navigate · click thread line to collapse