Master LLMs with our FREE course in collaboration with Activeloop & Intel Disruptor Initiative. Join now!


Statistics 101-Part 1-What Is Hypothesis Testing, Its Type, and Its Process?
Latest   Machine Learning

Statistics 101-Part 1-What Is Hypothesis Testing, Its Type, and Its Process?

Last Updated on July 25, 2023 by Editorial Team

Author(s): Kumar kaushal

Originally published on Towards AI.

Definition of hypothesis testing, the process of doing hypothesis testing, its types, and applications

Photo by Thomas T on Unsplash

This is a part of a series of articles related to Statistics 101. In Statistics, business analytics, and any other domain, we must have come across the term hypothesis testing. This article aims to provide a step-by-step guide for conducting hypothesis testing. We would understand the meaning of the term, its keywords, types, and applications. We would also elaborate on the meanings of important keywords such as p-value, significance value, and others.


  • Definition and importance
  • Hypothesis testing Process
  • Applications
  • References

Let’s begin!!!

Imagine a scenario where a manager of an electric vehicle company claims that the latest model runs 300 miles on a single charge. How would someone test or validate this assumption?

Hypothesis Testing would come to our rescue in such situations.

Definition and importance

If we refer to any dictionary for the meaning of a Hypothesis, it will point to the word “assumption”. For example, Merriam-Webster defines a Hypothesis as:

“a tentative assumption made in order to draw out and test its logical or empirical consequences”

Hence, hypothesis testing is the process of stating some assumption, testing it, and making some inference about the population basis the assumption made.

For the electric vehicle example stated at the beginning of this article, we could test the hypothesis on a sample and evaluate the results to make inferences regarding the average running of such vehicles, i.e., whether it is equal to 300 miles per charge or not. In the absence of Hypothesis Testing, it would be challenging to make inferences on the population parameters.

Hypothesis Testing Process

Step 1- Decide if the hypothesis is two-tailed or one-tailed

Step 2- Formulate the null hypothesis and the alternate hypothesis and set the significance level

Step 3- Basis the distribution to be considered, p-value to be calculated

Step 4- Compare the p-value with the significance level and select a null or alternate hypothesis to infer

Let’s go through each step in detail.

Step 1- Decide if the hypothesis is a two-tailed or one-tailed

Check the below statement:

Statement 1-Life expectancy in Asian countries is 60 years.

Here, this is a case of a two-tailed test where the hypothesized population parameter is compared for equality(=)or non-equality(!=) to a certain value.

Statement 2-Life expectancy in Asian countries is more than 60 years.

Statement 3-Average length of a video on YouTube a user watches is less than 7 minutes.

The above scenarios are cases of one-tailed hypothesis testing where the hypothesized population parameter is stated to be more or less than a certain value.

Step 2- Formulate hypothesis- null and alternate Hypothesis and set the significance level.

A null hypothesis(H0) is a hypothesis that is assumed as true until there is statistical evidence to “not accept it”. When the null hypothesis is not accepted, the alternate hypothesis (H1) is accepted.

For statement 1, the hypotheses statements would be:

Image by the author

And, for statement 3, it would be:

Image by the author

If we say that the Significance level is 0.05, it means the difference between the sample statistic(say mean- which has been calculated for the sample) and the hypothesized mean (60 for statement 1) has a 5% chance of occurring.

Step 3- Basis the distribution to be considered, p-value to be calculated

The below flow may be referred to for choosing the distribution to be used for the sample data.

Flow for 1 sample test (Image by the author)
Flow for two or more samples (Image by the author)

p-value (probability value), the probability of the null hypothesis is true. It is also called the calculated probability of committing a Type-1 error.

Step 4- Compare the p-value with the significance level and select the null or alternate Hypothesis to infer

Suppose the p-value is less than or equal to the significance level; hence, the probability of the null hypothesis being true is not significant. So, we reject the null hypothesis. If the p-value is greater than the significance level, we reject the alternate hypothesis.

A good reference to understand this is the Minitab blog.


Hypothesis Testing applies to scenarios where we are making or proposing any assumption and testing those. Be it manufacturing, clinical trials, marketing strategies, and any other domain, it is only left to the imagination of a business manager to apply Hypothesis Testing to these applications.


Definition of HYPOTHESIS

an assumption or concession made for the sake of argument; an interpretation of a practical situation or condition…

Statistical population – Wikipedia

In statistics, a population is a set of similar items or events which is of interest for some question or experiment. A…

Type I and type II errors – Wikipedia

In statistical hypothesis testing, a type I error is the mistaken rejection of an actually true null hypothesis (also…

Understanding Hypothesis Tests: Significance Levels (Alpha) and P values in Statistics

What do significance levels and P values mean in hypothesis tests? What is statistical significance anyway? In this…

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI

Feedback ↓