[sigi45]

40? so nothing relevant? :(

[mlechha]

Number of samples depends on the effect size. e.g. You only need n=1 to claim a dog can talk.

[elchief]

You realize that statistics is capable of using very small sample sizes right?

[sigi45]

And you do realize that 40 is just a tremendious small sample size in comparison to a complex system like a human beeing?

I mean yeah you need one dog who speaks but apparently, when you read other comments, there are enough papers saying the opposite to this one.