0
$\begingroup$

In this article which explains types of t-tests
for “Independent Two-Sample t-test” example:

We can confirm that the t-statistic is again less than the t-critical value so we fail to reject the null hypothesis. Hence, we can conclude that there is no difference between the mean screen size of both samples.

Null hypothesis was that the average screen size of the sample does not differ from 10 cm.
Alternate hypothesis was that the average screen size differs. And also the data is only about sizes of screens.

Why can't null hypothesis be that the average screen size differs. and alternate hypothesis, that the average screen size doesn't differ.

Not just this example, in every article I have read, every author only told what was chosen for null and alternate hypothesis. But, no one talked about why they chose what they chose. I always think that, why can’t the statement chosen for alternate hypothesis be chosen for null hypothesis and vice versa. And the data won’t have much information too. It’s always up to the article writer to what to choose for null and alternate hypothesis.

So, my question is, are there some set of rules, that have to be followed to choose a statement/condition for hypothesis. Or what actually causes a particular statement to be chosen for hypothesis. Or am I missing something.

$\endgroup$
1
$\begingroup$

You always choose the statement that you want to disprove as Null hypothesis. You can either reject it in favour of alternative hypothesis or not reject it. We don't say we accept null hypothesis. Because we don't have enough evidence for the null hypothesis to be true. It might be true, might be not. But with significant t/p value we can reject null hypothesis in favour of alternative hypothesis

Like in this example, if you don't get any significant value after the experiment you do not reject that the average screen size of the sample does not differ from 10 cm.But you don't have any proof to accept it either.

Always choose the statement you want to disprove as null hypothesis.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks. Also, What if I now perform a t-test with same data but with null hypothesis that screen size differs and alternate hypothesis that the screen size doesn't differ. So, with the p and t values of hypothesis test(will be same as the former test). Will it now be safe to conclude that screen size differs? $\endgroup$ – Naveen Kumar Jul 2 at 4:51
  • $\begingroup$ Usually null hypothesis is taken some exact value. Like scree. Size is 10, weight is 60 etc. Not some range of value which is difficult to work with. $\endgroup$ – SrJ Jul 2 at 8:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.