# Null Hypothesis Explained: How And When To Use It

If a Null Hypothesis could talk, it would say something like this: “I have no relationship with the variables in this study.” In other words, the null hypothesis assumes that the measured phenomena [the dependent variable] and the independent variable are unrelated. If this sounds confusing, you’re not alone.

This article sets out to clear any confusion you might have on the null hypothesis, their meaning, purpose, and use in statistical research – with clear examples.

Before you continue, make sure you get familiar with what research is, research steps and concepts, and how the scientific method is used when formulating hypotheses.

## What is Null Hypothesis?

Definition: The null hypothesis is a statement of equality and acts as a starting point accepted as true in the absence of additional information. Think of it as a statistical assumption that claims there is no difference between specific features of a population or data-gathering methodology.

Take a look at the following null hypothesis examples:

1. There are no differences between the 15 and 20 years old score average on the ABC memory test.
2. Age does not affect how laptops are used to connect to the Internet.
3. Dogs do not choose their food based on color.
4. The type of candies preferred is independent of the buyer’s age, gender or income.

The four null hypotheses above share one thing: they are all statements of two or more items being equal or unrelated.

If we have to transcribe in an equation the example above, we would have:

H0 : 𝝁15 = 𝝁20

Where:

H0 = the generally accepted symbol for a null hypothesis.

𝝁15 = the Greek letter Mu symbol used to represent the theoretical average population of 15 years old.

𝝁20 = the Greek letter Mu symbol used to describe the theoretical average population of 20 years old.

Researchers usually work to disprove, nullify or reject a null hypothesis by testing it to find alternative hypotheses that better explain a phenomenon.

It is important to remember that it doesn’t mean it must be accepted just because a null hypothesis is not rejected. In other words, if there is insufficient data to show that the difference of means is not zero does not imply that the difference is zero. The devil is in the details.

## Why Null?

The concept of the null hypothesis was invented by the British statistician, geneticist, and academician Ronald Fisher.

The “null” term is often a cause of confusion among beginner researchers who often think null refers to “zero” or “amount to nothing.”

The term stands for a generally accepted fact. Researchers often work to disprove the idea in question through experimental observation, which brings up an important fact: a null hypothesis must be testable to meet the scientific standards.

For instance, if a null hypothesis states that “all candies are sweet” and we taste one that is sour, the hypothesis can be disproved.

Keep in mind that even if the confidence level is high [usually 95-99 percent], there is still a small probability of the null hypothesis being proven false.

It is true until proven wrong.

## What is the Basic Purpose of the Null Hypothesis?

Think of a null hypothesis as the starting point and the benchmark against which the outcome of a study will be measured.

A null hypothesis acts as a starting point because the statement is accepted as true in the absence of additional information.

## Null Hypothesis as Starting Point

For example, the first null hypothesis in our examples states that “there are no differences between the 15 and 20 years old score average on the ABC memory test.” Given no new knowledge about the 15 and 20 years old score average, there is no reason to think there is a difference between the two age groups.

You can speculate on differences that might impact the average score between the two age groups. However, without any evidence a priori [before the fact], there is no choice but to assume the average score is equal. In other words, until you prove the difference, you have to accept there is no difference.

Furthermore, suppose differences between groups exist. In that case, you must presume that the discrepancies are due to the most appealing explanation for disparities between any groups on any variable: chance!

That’s correct! Chance is always the most plausible explanation for discrepancies between two groups if no other information is brought forward.

But what exactly is chance? It is the random variability introduced due to the individuals participating in a study and numerous other unforeseeable factors.

If that sounds confusing, check the following example.

Let’s assume we are comparing the running speeds of two groups of athletes representing Argentina and the United States. How would you know if a group receives more training than others? Or which group practices more?

Maybe the way their speed is measured leaves room for chance, e.g., a cold day can affect the true running speed performance between athletes.

Our job as researchers is to eliminate the chance factor and all other factors contributing to group differences, such as factors identified as independent variables.

## Null Hypothesis as Benchmark

We also mentioned that a null hypothesis could act as a benchmark against the study outcomes and determine if chance or other factors cause any differences.

A null hypothesis defines a range within which any observed differences between groups may be ascribed to chance [the null hypothesis] or factors other than chance [e.g., manipulating the independent variable].

## How to use the Null Hypothesis?

It’s simpler than you might think. To use a null hypothesis in your study just follow these two steps:

Step 1: Start with a question, e.g., Does age affect how laptops connect to the Internet?

Step 2: Rephrase the question in a way that assumes no relationship between variables: Age does not affect how laptops are used to connect to the Internet.

There you go. Simple right? Here are some more examples [some needs your answer]:

NOTE: You can find the answers for Question 2 and Question 4 at the end of this post.

## When To Use the Null Hypothesis?

Usually, most experimental, quasi-experimental or correlational studies have an implied null hypothesis. Here, the goal is to show whether or not the test is supported, free from the researcher’s decisions or personal views. A null hypothesis serves as guidance for such types of research, and it states the exact opposite of what the researcher predicts as expected.

Example:

In contrast, descriptive and historical studies may not have/need a null hypothesis.

For example, suppose you research the life expectancy growth for women in the past 100 years [historical study] or how people feel about Amazon discounts [descriptive study]. In that case, you may not be concerned with a null hypothesis in your research.

## Null Hypothesis – Things To Remember:

• Are considered a generally accepted truth. However, that doesn’t mean that the null hypothesis can’t be falsified/rejected.
• They’re the opposite of the alternative hypothesis, which states a relationship between the variables.
• Assumes no variance between the characteristics of a population.
• It can be rejected within a certain level of confidence using the hypothesis testing method.

## Conclusion

It is essential to understand the role played by null hypotheses in statistical research. They may be used as starting point or benchmarking against a research outcome and help us conclude whether or not there is a correlation between two measured phenomena.

It can also help us determine whether or not two measurable events have a relationship and find if the outcomes are due to chance or are the consequence of controlling phenomena.

Q1: The shortest distance between two points is a straight line.
Q3: Mobile phones do not influence the sleep quality in humans.