Is there a 100% chance if null hypothesis is true then parameter must be true

hypothesis

#1

I am currently studying about hypothesis generation while studying it I came across two type of hypothesis generation one is null hypothesis and other is alternative hypothesis.I have studied that we always assume that null hypothesis is true and we fill the evidence for disproving it if we don not find enough evidence to disprove it we will assume that null hypothesis is true,

My question is that if we prove that null hypothesis is true how accurate we are that parameter actually represents the population.


#2

Hi @sid100158

This depend on your choice of p value to accept or reject the Null hypothesis and the current p value and then the standardisation you did if you want to be in the original unit,

Alain


#3

hello @sid100158,

There can never be 100% probability because of randomness.
It is your p value which defines the prob of observing the particular outcome.
If it is less than the significance level you will reject the null else you will fail to reject the null.

In the above diagram the prob of observing the particular statistic under the assumption that the null hypothesis is true is 6.67%.Also we can never find the true parameter until and unless we have access to population data(which can be rare).We can only estimate the parameter by a sample statistic.

Hope this helps!!