You should very much know that the null hypothesis for a proposed behavioural modification technique is that it isn't a benefit and doesn't work and may be harmful. It's on the proposer to show otherwise.
What's less harmful when it comes to human health and safety in non-essential activities, to assume no risk of harm or injury or to assume risk of harm or injury? I assert that the null hypothesis is that it does not have a benefit and is harmful. This is not an assertion of fact, simply a position of likelihood and a judgement that doing a physical and emotional punishment to a person is generally harmful and that doing so must be balanced against ensuring a just outcome. This is not an unreasonable stance. You, on the other hand, if I understand properly, are saying that we should be able to do whatever we like to a punish person unless we show it directly causes harm. Down that road lies such human rights abuses that holding such views should be considered unconscionable.
> What's less harmful when it comes to human health and safety in non-essential activities, to assume no risk of harm or injury or to assume risk of harm or injury?
I agree with this statement. Assume a risk.
> I assert that the null hypothesis is that it does not have a benefit and is harmful.
Why did the word 'risk' disappear here? I don't think that's a good null hypothesis at all.
> You, on the other hand, if I understand properly, are saying that we should be able to do whatever we like to a punish person unless we show it directly causes harm.
And you need to make sure you're looking for both positive and negative effects.
"It's harmful" is a huge bias for a null hypothesis.
> Then what are you saying?
That you shouldn't assume it is harmful.
I didn't say anything about what parents should be "able to do". I didn't even give an opinion on the candy thing. I just think your justification is a big overreach.
But also a null hypothesis implies you're currently doing the testing, so that's a different scenario too...