Null and alternative hypotheses are assumptions researchers use during statistical analysis to understand relationships between two or more independent and dependent variables or phenomena. Analysts, researchers, and statisticians use statistical analysis software to perform complex null and hypothesis analysis testing.
Null and alternative hypotheses are mutually exclusive events and have their differences.
What is the difference between null and alternative hypotheses?
A null hypothesis (H0) is a statistical hypothesis that states there is no statistical significance or relationship among the variables in a dataset. An alternative hypothesis (H1 or Ha) in a statistical inference experiment directly contradicts the null hypothesis and states that there is a relationship between two variables.
While the null hypothesis presumes no change or status quo, an alternative hypothesis or the claim shows that a non-random cause influences the observations. That’s the key difference between null and alternative hypotheses.
The comparison table below shows how null and alternative hypotheses differ regarding their testing goals, observations, and acceptance criteria.
| |
Null hypothesis |
Alternative hypothesis |
| Definition |
A null hypothesis is the default hypothesis that suggests there is no relationship, difference, or observed effect between two variables. |
An alternative hypothesis states there is a relationship, difference, or observed effect between two variables. |
| Statistical notation |
H0 denotes a null hypothesis. |
H1 or Ha denotes an alternative hypothesis. |
| Symbols used |
A null hypothesis uses equal signs ( =, >=, <=). |
An alternative hypothesis uses not equal symbols ( !=, <, >). |
| Purpose |
A null hypothesis assumes that no relationships exist between variables. |
An alternative hypothesis suggests that a significant relationship exists between variables. |
| Types |
Simple, composite, exact, inexact |
Point, one-tailed directional, two-tailed directional, and non-directional |
| Testing goal |
Researchers aim to disprove or fail to reject a null hypothesis. |
Researchers aim to prove, accept, or support an alternative hypothesis. |
| Type of testing |
As it uses the population parameter, null hypothesis testing is indirect and implicit. |
Alternative hypothesis testing is direct and explicit because it indicates sample statistics. |
| P-value |
The p-value is smaller than the statistical significance level in a null hypothesis. Researchers favor the null hypothesis when the p-value exceeds the statistical significance level. |
The p-value is greater than the significance level in an alternative hypothesis. Researchers favor the alternative hypothesis when the p-value is lower than the statistical significance level. |
| Observations |
Observations in a null hypothesis are the outcome of chances. |
Observations in an alternative hypothesis are the outcome of real effects. |
| Acceptance criteria |
Researchers use a predefined significance level, also known as the alpha level, to find the rejection threshold. |
The statistical significance level and effect size determine the evidence strength to support an alternative hypothesis. |
Learn different statistical analysis methods to discover data patterns and trends.
