Jesweez
1d ago
The null hypothesis is indeed one of a set of competing hypotheses, but it’s typically the one that assumes no difference between populations.
If I want to show that two distributions are statistically different then I start with the assumption that they are not and then set out to disprove that.
Similarly, if I believe the populations are not actually significantly different, I believe it’d still be common to set up a null hypothesis that they are not different and then either confirm or reject the alternative hypothesis.
Jesweez
1d ago
H0 (null): no difference between populations
H1 (alternative): radiation resistance of the new population > radiation resistance of the reference bird population
This is the typical formulation. Null typically assumes no effect or no difference between the populations being considered.
Jesweez
4d ago
If you seriously think people in your day to day want you dead I think this is more indicative of a medical issue than a political philosophy debate.
This is the most extreme kind of political thinking that I’ve seen
You should not treat your political opponents as a homogenous group made of their most distasteful members
This (quite common) cognitive mistake becomes particularly egregious when the conclusion is: they all want to kill me
Jesweez
6d ago
Do not insult the flag of Cusco
Jesweez
6d ago
I think the problem tends to arise when people use these symbols and then also want to deport undesired races from their homeland.
That’s kind of like asking how does algebra have a place here, we’re just trying to solve for a variable in this equation.
A hypothesis test is a method to provide evidence for or against two competing theories using data and the way that they’re commonly constructed is to assume a null hypothesis as being the one where the data are not from significantly different distributions.
A standard hypothesis test is not the only method and its use in science is sometimes over stated but it’s by far the most common approach to address such questions, and that’s just how it’s structured.
It’s kind of Occam in the end. It’s simpler to assume that there’s no difference between how fast this group of monkeys climb trees vs that one. If I wanted to posit that the second group climbs faster, I can collect data and argue that it backs up my assumption, but the null case is null because it makes less assumptions.
More options
Context Copy link