Hypothesis Testing – A Complete Guide with Examples

Characteristics of a Good Hypothesis

A hypothesis should be:

• Clear and precise

• Testable and measurable

• Related to specific variables

• Stated in simple terms

• Consistent with known facts

• Limited in scope and time

Steps in Hypothesis Testing

There are 6 main steps in hypothesis testing:

1. State the null and alternative hypotheses.

2. Collect data.

3. Select an appropriate statistical test.

4. Select the level of significance.

5. Decide whether to reject or fail to reject the null hypothesis.

6. Present the outcomes of your study.

What are the Null Hypothesis and the Alternative Hypothesis?

A null hypothesis is a hypothesis when there is no significant relationship between the dependent and the independent variables. 

In simple words, it’s a hypothesis that has been put forth but hasn’t been proved as yet. A researcher aims to disprove the theory. 

The abbreviation “Ho” is used to denote a null hypothesis.

An alternative hypothesis (H₁ or Ha) is completely opposite, as it shows that there is a relationship or difference, which the researcher has yet to prove. 

Example

In an automobile trial, you feel that the new vehicle’s mileage is better than the previous model of the vehicle. You can write it as;

 (Null Hypothesis)

• H₀: There is no difference between the mileage of the old and new car models.
  

 (Alternative Hypothesis)

• Ha: The new vehicle has better mileage than the old model.

If H₀ is rejected, this can be due to the following errors:

• Type I error:  Rejecting H₀when it’s actually right.

Type II error: Occurs when failing to reject H₀ when it is actually false.

How to Conduct Hypothesis Testing?

Here is a step-by-step guide on how to conduct hypothesis testing.

Step 1: State the Null and Alternative Hypotheses

Once you develop a research hypothesis, it’s important to state it as a Null hypothesis (Ho) and an Alternative hypothesis (Ha) to test it statistically.

Example
For example, you want to test whether the rate of weight gain is higher in women as compared to men. 

Then you write it as;

• H₀: Women, on average, do not gain weight faster than men.

• Ha: Women, on average, gain weight faster than men.

Step 2: Collect Data

Hypothesis testing follows the statistical method, and statistics are all about data. It’s challenging to gather complete information about a specific population you want to study. 

You need to gather the data obtained through a large number of samples from a specific population. 

Example

• To check the difference in obesity rate between men and women, collect equal samples from each gender group. 
• Consider the factors like diet, lifestyle, and profession. 
• Determine the scope of study, whether it is applicable to the local or global population. 

Step 3: Select an Appropriate Statistical Test

There are many types of statistical tests, but we discuss the two most common types below, 

• One-sided Test

Values of rejection of the null hypothesis are located in one tail of the distribution.

Example

If you want to test whether all the mangoes in a basket are ripe.
H₀: All mangoes in a basket are ripe.
If tests prove that all mangoes are ripe, H₀ is true.

• Two-tailed test

Values of rejection lie outside the null hypothesis on both tails of the distribution.

Example
H₀: All mangoes in a basket are ripe.
But if testing proves that some are unripe, H₀ is rejected.

Step 4: Select the Level of Significance

When you reject a null hypothesis, even if it’s true during a statistical hypothesis, it is considered significant. The significance level (α) is the probability of making a Type I error. It is usually set at 0.05 (5%).

• If p-value < 0.05, reject H₀.

• If p-value ≥ 0.05, fail to reject H₀.

Note: If the significance level is minimum, then it prevents the researchers from making false claims.

Example
You apply a one-sided test to test whether women gain weight quickly compared to men. You get to know about the average weight between men and women and the factors promoting weight gain.

Step 5: Decide Whether to Reject or Fail to Reject the Null Hypothesis

Use the p-value to decide whether to reject or fail to reject the null hypothesis after running the test.

Example

• If p < 0.05, reject H₀ (Women do not gain weight faster than men). Accept Ha (Women gain weight faster than men).
• If p ≥ 0.05, fail to reject H₀.

Step 6: Present the Outcomes of Your Study

The final step is to present the outcomes of your study. You need to ensure whether you have met the objectives of your research or not. In the research results are presented in the following sections;

• Conclusion section: Summarize data, test results, and p-values.

• Discussion section: Explain whether the findings support Ha or not.

Example
So, for example, our study found that the value of p < 0.05. H₀ will be rejected as (“women do not gain weight faster”) and conclude Ha is accurate.

Note: In academic writing, instead of writing  “accepted” or “rejected” hypotheses. You should say:

• “We reject H₀ in favor of Ha.”

• “We do not reject H₀.

Frequently Asked Questions