Null hypothesis is effectively stating the default situation. It is as you said there is nothing going on.
When some one puts forward a claim the null hypothesis is that this claim is not true and needs to be proven.
When we are testing between a new claim and the null hypothesis, a null hypothesis is H0 and the claim is the alternative hypothesis H1.
we consider the null hypothesis to be true unless the data shows a strong indication that it is not.
In hypothesis testing, we test the probability of finding a specific value for a sample statistic given that H0 is true.
If the null hypothesis is true, the difference between a sample statistics and the population parameter is only because of sampling error just a variation in the sample from the population. This comes when the probability of finding the sample statistic as extreme as the data results is big.
However, if this probability is very small, we generally reject the null hypothesis.
Imagine we have found a p value of 0.46 called p0 and another p value of 0.001 called p1, then p0 is considered a strong evidence in favor of null hypothesis and against the claim or the alternative hypothesis while p1 is a weak evidence in favor of null hypothesis and a strong evidence in favor of the claim.
What does a p value of 0.15 mean?
A p value of 0.15 means that there’s a probability of 15% of obtaining a similar result or more extreme as the data gives us given that the null hypothesis is true.
In future posts we will discuss how to calculate p values and make inferences.
