What Is a Null Hypothesis? (Definitions, Examples and FAQs)

By Indeed Editorial Team

Published September 2, 2021

The Indeed Editorial Team comprises a diverse and talented team of writers, researchers and subject matter experts equipped with Indeed's data and insights to deliver useful tips to help guide your career journey.

For research scientists and statisticians, it's important to understand what the null and alternative hypotheses are and how they function. The null and alternative hypotheses are important statistical concepts that help scientists develop research projects and test variables to answer research questions. Learning about the null hypothesis and how it's used alongside an alternative hypothesis can help you understand these terms and use them effectively in your research. In this article, we define what the null and alternative hypotheses are, explain how they work together and provide examples to help you understand how to apply them to your research.

What is the null hypothesis?

The null hypothesis is a concept in statistics used to represent the possibility of two or more variables having no statistically significant relationship or correlation with each other. When a scientist proposes a study or experiment, they test the possibility of there being a relationship between their chosen variables. If the researcher finds a statistically significant correlation between their variables, they reject the null hypothesis and conclude that there's a meaningful relationship between the constructs they tested. For example, a researcher might hypothesize that there's a positive correlation between temperature and ice cream consumption during the summer.

They design a study to investigate whether people eat more ice cream on hotter days, and they write a hypothesis describing their predicted outcome. During the study, they collect data about the air temperature and the amount of ice cream people eat each day. They process that data to determine whether the number of people eating ice cream on hotter days is significant enough to correlate statistically. If so, they reject the null hypothesis, which claims that there isn't a statistically significant correlation between ice cream consumption and air temperature, and they support their hypothesis that a correlation exists.

Related: Understanding What a Null Hypothesis Is

What is an alternative hypothesis?

The alternative hypothesis is the inverse of the null hypothesis, meaning the alternative hypothesis represents the possibility of a correlation between two or more variables in a study or experiment. Researchers attempt to prove the alternative hypothesis by disproving the null hypothesis. When the scientist discovers a statistically significant correlation between their variables, they support the alternative hypothesis and reject the null because they have collected enough data to show a statistically meaningful relationship between the variables.

For example, when testing the relationship between ice cream consumption and air temperature, the scientist proposes people eat more ice cream when the air temperature is hotter. The alternative hypothesis is that there's a positive correlation between air temperatures and ice cream consumption, and the null hypothesis is that there's no relationship between the variables. By proving this relationship, the scientist rejects the null and verifies the alternative hypothesis.

What is the null hypothesis vs. the alternative hypothesis?

The null and alternative hypotheses work together to create a model for determining the existence of a statistically significant relationship between variables. Here are some key comparisons between these two hypotheses:

Purpose

The primary purpose of these hypotheses is to provide a framework in which a researcher can either validate or refute the alternative hypothesis. Testing a hypothesis allows the scientist to answer important scientific questions and advance theories within their field of study. Asking and testing research questions promotes the development of new knowledge on a subject by building off the findings of previous studies. While the purpose of the alternative hypothesis is to verify a statistically significant correlation between variables, the purpose of the null hypothesis is to determine the absence of a correlation.

Related: Statistical Significance: Definition and Application in the Workplace

Principle

The principle of the null hypothesis is to create a statistical model in which the researcher can collect and process data to determine whether there is a plausible correlation between variables. After collecting data, the researcher chooses a statistical tool to process the numbers. The tool used to process the data depends on the type of data, quantity and how the researcher wishes to interpret the relationship between data sets. Based on these calculations, the researcher determines whether the data validates or refutes the null. These calculations can also provide information on the strength of the correlation between the data.

This principle works similarly for the alternative hypothesis. The alternative hypothesis also structures the statistical model by providing a claim to compare against the null. Using the results from testing the data, the researcher determines whether the correlation between the variables has statistical significance. If so, they validate the alternative hypothesis and reject the null.

Verification and rejection

These hypotheses work together in a system of oppositions, meaning the researcher can only verify one or the other. Researchers reject the null hypothesis when the results of their data processing method indicate that there's a statistically significant correlation between the variables. Statistical significance means that random chance didn't produce the data. To determine whether the results are statistically significant, the researcher processes the data and calculates whether the resulting correlation is strong enough to show a true, consistent relationship between the data sets. When these results are significant, the researcher validates their alternative hypothesis and refutes the null.

For example, a psychologist studying the relationship between social engagement and perceived happiness may determine that their data can have up to a 5% chance of randomness to be considered significant. They collect their data and process it using a statistical tool and find that there's a 4% chance that the correlation between the variables results from random chance. Since this value is less than 5%, the researcher confirms the results are statistically significant, meaning there is a true correlation between the variables and they can reject the null hypothesis.

Related: How To Calculate Statistical Significance (Plus What It Is and Why It's Important)

Differences between null and alternative hypotheses

Here are some key differences between these hypotheses:

  • Definition: By definition, the null hypothesis states that there isn't a meaningful relationship between the study's variables. The alternative hypothesis is a statement that claims there's a relationship between variables.

  • Claim: Statistically, the null hypothesis claims that any apparent correlation between variables results from chance. The alternative hypothesis states that there's a cause-and-effect relationship between variables that don't result from random chance.

  • Proof: Researchers look to disprove the null hypothesis through their studies by proving the alternative hypothesis.

  • Statistical significance: A statistically significant result proves the alternative hypothesis, while a nonsignificant result proves the null hypothesis

  • Importance: Both of these hypotheses are important. Verifying a null hypothesis supports existing theories and verifies continuity between studies while verifying an alternative hypothesis can lead to new theories or new ways of understanding established theories.

Related: Defining Hypothesis Testing (With Examples)

Examples of hypotheses in research

Here are two examples of research scenarios with considerations for the null and alternative hypothesis for each:

Example 1

A social psychologist designs a study to understand how perceived authority correlates with the degree of vocal accommodation between parties in a conversation. They randomly assign two participants to roles: one participant gets the authority role while the other is the nonauthority. The participants work together to solve a problem. As the pair works to solve the problem, the researcher records the conversation so they can analyze the degree to which the nonauthority participant changes their vocal cues to accommodate the authority participant. The researcher hypothesizes the nonauthority participant will change their vocal cues more than the authority participant.

In this example, the alternative hypothesis states that there is a correlation between perceived authority and vocal accommodation. The null hypothesis argues that there isn't a correlation between these variables. The researcher analyzes the data to determine if there is a correlation between the variables. If so, they assess the correlation to decide whether it's statistically significant. If the results correlate to a degree that shows statistical significance, the researcher rejects the null and verifies the alternative hypothesis.

Example 2

A medical researcher plans a clinical trial to determine the effectiveness of a new medication for treating chronic migraines. They design a study in which they give half the participants a placebo, or fake pill, and the other half receive the trial drug. Every week, the researcher meets with the participants, conducts a physical examination and asks them to self-report their migraine symptoms. The researcher hypothesizes the participants who received the trial drug will report fewer migraine symptoms than the group that received the placebo.

In this example, the alternative hypothesis states that there's a correlation between taking the trial drug and reduced migraine symptoms. The null hypothesis argues that there is no correlation between taking the trial drug and reduced migraine symptoms. The researcher analyzes the data gathered from both participant groups and compares the results. If the trial drug group reports a statistically significant amount of improvement in their symptoms compared to the placebo group, the researcher verifies the alternative and rejects the null hypothesis.

Explore more articles