A null hypothesis is a hypothesis that says there is no association or statistical significance between the two variables in the statement. When talking about a test hypothesis we must take a closer look to the null hypothesis and the alternative hypothesis as they are the base of proving if your experiment is true or false.
As an example, a company believes that sending a coupon for each testimonial received from their clients will increase the quantity of testimonials to their website. They split the number of clients receiving coupons and not receiving coupons (sample) for 2 weeks (the experiment), aiming to prove their hypothesis true or false.
The null hypothesis would be “clients that receive coupons for their testimonial does not help increase the quantity of testimonials”. In a null hypothesis we try to reject the statement with sample data, as it is presumed to be true until statistical evidence nullifies it for an alternative hypothesis.
When stating a null hypothesis, you would have to present the alternative hypothesis. It is simply the opposite of the null hypothesis. Continuing with the above example, the alternative hypothesis would be something such as “clients that receive a coupon for their testimonial will help increase the quantity of testimonial to their website”.
The null hypothesis (or H0) assumes that any kind of difference or significance you see in a set of data is due to chance and we use the alternative hypothesis to prove the opposite.
There is a simple 4-step process to conduct a statistical hypothesis:
When accepting or rejecting the null hypothesis, you have to take into consideration the statistical significance of the results and for that we look at the p-value. A p-value that is less than or equal to 0.05 can indicate that there is strong evidence against the null hypothesis and therefore you can be sure you have disproved the null hypothesis and can accept the alternative hypothesis.
So why should someone test a null hypothesis just to find it false? Simply because it’s much easier to reject a hypothesis than to accept one. This is because there can be several alternative hypotheses depending on certain factors that could ultimately determine the alternate hypothesis as incorrect.
No matter what direction you take for testing hypothesis, the importance is to keep on testing with factual data and not hunch-based data.
What is the Multi-Armed Bandit problem? How to choose between actions with unknown payouts.READ MORE
What is Multivariate Testing? Simple, full factorial and fractional factorial MVT.READ MORE
Retargeting a consumer while engaged on your site can deliver significant improvements to your conversion rates.READ MORE
There's nothing like getting real customer feedback - instantlyREAD MORE