> And how would you even double-blind test this anyway?

Get lights that either start fast and slow down or blink at speeds that randomly change then use serial numbers to track which lights are which, but send them out at random to test subjects?

Which treatment you're in is relatively easily detectable, so unfortunately that wouldn't work.

I like the approach though, it would be a reasonable control condition (assuming that the 60bpm is a core part of the effect).

you would blind the participants to the effect you are looking to measure. If they don’t know what to expect or what conditions there are, they can still get a placebo effect from random patterns.

> you would blind the participants to the effect you are looking to measure. If they don’t know what to expect or what conditions there are, they can still get a placebo effect from random patterns.

Maybe. Generally you'd do a manipulation check at the end to ensure that people couldn't identify what condition they were in.

Additionally, that's single blind rather than double blind.

The researchers can be blinded by not experiencing the intervention and handing out identical looking devices. You’re right that you would need a manipulation check at the end but this experiment is totally design-able properly.

I still think that the manipulation check would reveal failures of blinding, but yeah it is doable (but we needed a while to come up with a reasonable design).

And more generally, my philosophy (as someone with a PhD in the placebo effect) is that harnessing any expectancy/placebo effects in one's personal life is totally reasonable, and if it works for people and there's low risk of harm then more power to them.

We shouldn't gatekeep potentially useful personal treatments based on the gold standard. It's kinda the medical equivalent of YAGNI, I suppose.

Trivially, given a small amount of money.

He sees a correlation, correlations are amongst the easiest thing to test.

Causation is the difficult one