Relation between hypothesis testing and null hypothesis

datamodeling

#1

I am currently studying a different type of hypothesis which helps in finding the importance of the variable.I have studied the hypothesis testing and null hypothesis.

Hypothesis testing is the use of statistics to determine the probability that a given hypothesis is true.

The null hypothesis is a hypothesis which the researcher tries to disprove, reject or nullify.

What I want to know is there a relation between them if yes then what is the relation and if no then how they are different from each other.


#2

Hi @hinduja1234,

I think you’re getting a but mixed up here. Actually ‘hypothesis testing (HT)’ is used to test the ‘null hypothesis’. So when you said “HT is the use of statistics to determine the probability that a given hypothesis is true”, null hypothesis is the ‘given hypothesis’ which you are testing.

Let me give you an example. For instance, you’re trying to check whether ‘Men are Taller than women in a particular organisation’ (P.S: don’t judge me from the example :D)

So we will collect the data for say 100 students in the organisation and lets define a null hypothesis, say ‘Men are taller than women’.

The hypothesis testing will try to find evidence to disprove the null hypothesis. Generally every HT technique would give a ‘pvalue’ and thumbrule is that a ‘pvalue’ less than 0.05 means that null hypothesis is invalid. ‘pvalue’ is actually the probability of the observed data given the null hypothesis is true. If this probability is less than 0.05 (or some other threshold), then we reject the null hypothesis.

The last part about ‘pvalue’ is a bit tricky. Feel free to discuss further.

Cheers!


#3

Hi @hinduja1234

I have s slightly different explanation than @Aarshay. First when you mentioned hypothesis testing you mean you when to test a positive, for example that water turned into ice at 0 degree. Well now it starts you take water in Delhi and freeze it 0 degree it works , Paris similar, you go by the sea it does not work salts you should go at -3 minimum. Now the issue if you are positive you should test all possibilities. It is easier to go by negative as you will need one case and we are in the null hypothesis there is no differences in water freezing temperature ? Well this easy to prove and you need only one difference (not all) you go by the sea -3 to freeze, that is it you can reject you null hypothesis. There is a difference.
This is an explanation without statistics, if go by the p-value the famous one you got:

  • p-value it is defined as “probability of observing the observed difference or greater difference when the null hypothesis is true
    Which is certainly not a intuitive but sounds scientific !!!
    Alain