Alternative Hypothesis-Definition, Types and Examples

A hypothesis is a working assertion based on scant data. As a result, there arises a requirement for additional testing. A statistical hypothesis test is a technique for determining if the available data are sufficient to support a given hypothesis. Hypothesis testing can be used to make probabilistic claims about the characteristics of the population. There are two types of hypotheses:

1. Null hypothesis

2. Alternative hypothesis

Alternative hypothesis

The alternative hypothesis is a hypothesis under which a statistically significant association exists between two variables. It is denoted by Ha or H1. It is also called the research hypothesis. The alternative hypothesis is typically a claim that a researcher believes to be true and that rejects the null hypothesis. An alternative hypothesis is complementary to the null hypothesis and thus, only one of them can be true at a time. Suppose, your null hypothesis is ‘Harry will get more than 20 marks in the test’ then the alternative hypothesis will be ‘Harry will get less than or equal to 20 marks in the test’.

Types of alternative hypothesis

There are three types of alternative hypotheses:

1. Left-tailed: In this type of hypothesis, the sample proportion (ᴨ) is less than a specific value (ᴨ0).

Ha : ᴨ < ᴨ0

2. Right-tailed: In this type of hypothesis the sample proportion (ᴨ) is greater than a specific value (ᴨ0).

Ha : ᴨ > ᴨ0

3. Two-tailed: In this type of hypothesis the sample proportion (ᴨ) is not equal to a specific value (ᴨ0).

Ha: ᴨ ≠ ᴨ0

Null hypothesis

A null hypothesis is a hypothesis in which there is no statistically significant relationship between two variables. It is denoted by the symbol H0. Researchers try to reject the null hypothesis and always write the null hypothesis in terms of ‘no effect’, ‘no difference’, or ‘no relation’. One should never prove or accept the null hypothesis.

Differences and similarities between null and alternative hypothesis

Differences between null and alternate hypothesis:

Similarities between null and alternative hypotheses:

• They are both solutions to research problems.

• Statistical tests are used to examine them.

Examples

• We want to determine if the mean salary of employees in an office is different from $75,000. The null and alternative hypotheses are:

H0: ᴨ = $75,000

Ha: ᴨ ≠ $75,000

• Let us take a decision in a court as a statistical hypothesis test. The null hypothesis is that the defendant is assumed to be innocent while the alternative hypothesis states that the defendant is guilty. The defendant is innocent until proven guilty likewise it is presumed that the null hypothesis is true in a hypothesis test until proven otherwise. There must be sufficient evidence to prove the alternative hypothesis true that is to prove the defendant guilty.

Only legitimate evidence may be used in a court as the basis for a trial. To determine the statistical significance of the null hypothesis in hypothesis testing, an acceptable test statistic should be used. If the null hypothesis is proven wrong at a particular level of significance, evidence would back up the alternative hypothesis.