Introduction

We have discussed types of hypotheses and the difference between significance and confidence level: https://researchmethodsclass.com/download/understanding-hypothesis-testing-in-research-significance-level-pdf/

This blog will discuss the steps followed while testing hypotheses and the decision(s) to make after testing hypotheses.

Steps involved in hypothesis testing

Hypothesis testing follows a systematic process. The steps below use an education example to show how the process works.

Step 1: Formulate the research question

Begin with a clear investigative research question.

Example:

What is the influence of digital learning on the academic performance of students in higher education institutions?

The research question identifies the issue the study seeks to investigate.

Step 2: State the null and alternative hypotheses

Suppose preparedness is measured using a score, and the researcher compares. Ideally, state the alternative non-directional hypothesis in Chapter One but test the null in Chapter Four

The hypotheses are:

H₀: There is no significant relationship between digital learning materials and the academic performance of students in higher education institutions

H₁: There is a significant relationship between digital learning materials and the academic performance of students in higher education institutions

Step 3: Select the significance level

The researcher chooses a significance level before conducting the statistical test. In social science, a significance level of 0.05 is normally used. However, one should be sensitive when using SPSS. Sometimes, based on the sample size and the statistical tool selected, SPSS may automatically select the level of significance for testing the hypothesis rather than 0.05. The purpose of the significance level is to establish the decision threshold for evaluating the evidence.

For this example:

α=0.05

Step 4: Select the appropriate statistical test

The choice of test depends on the scale of measurement and assumptions of the statistical test.

For example, an independent-samples t-test may be appropriate when comparing the mean preparedness scores of two independent groups, provided its assumptions are reasonably satisfied.

The chi-square test is a test of the relationship between two variables measured at the nominal level

Simple and Multiple Regression is a predictive statistical tool that predicts the value of Y given X for variables measured at interval or ratio scale.

Step 5: Collect sample data

The researcher selects a sample of students and measures their academic performance.

For illustration, suppose the following results are obtained:

Group	Number of students	Mean performance score
Male	50	78
female	50	70

The males have a mean score that is 8 points higher. However, the difference in sample means alone does not establish whether the population means differ. We need to consider the variability of the scores and the uncertainty associated with sampling.

Step 6: Calculate the test statistic and p-value

The selected statistical test uses the sample data to calculate a test statistic. For the above sample, we shall use an independent-samples t-test to test the mean difference in academic performance between males and females.

The statistical test analysis also produces a p-value. This is called the calculated p-value. This p-value is the probability, assuming the null hypothesis and the statistical model are correct, of obtaining a result at least as extreme as the one observed.

Step 7: Compare the p-value with the significance level

This is a critical point because it determines the decision that the researcher makes.

Suppose the analysis produces a p-value of 0.02.

The researcher compares it with the significance level of 0.05.

• If p≤0.05, reject the null hypothesis.
• If p>0.05, fail to reject the null hypothesis.

In our illustration:

0.02<0.05. In this illustration, the calculated p-value is less than the set p-value. The decision is therefore to reject the null.

Step 8: Interpret the findings

Rejecting the null (or failing to reject the null) does not mean that your research is good or bad. It simply means that the researcher does not have sufficient evidence to reject or fail to reject the hypothesis.

In this step, the researcher should explain what the statistical decision means in relation to the hypothesis.

For example:

Though the males have a higher mean performance than females, the analysis provides statistically significant evidence of a difference in mean academic performance between male and female students. Note that males having a higher mean score does not make them more academically endowed than females in the population. It just means, in the sample collected, the males have a higher mean score than the females. Finally, this does not also lead to causality; i.e., we cannot conclude that there is a factor X that caused the “good” performance in the males. 

However, testing the hypothesis has shown that there is no statistical difference between the mean score of males and females.  

Step 9: Report the results

Good research should present the findings, interpret them, and discuss the findings. Presentation is the visual display of the findings either in Tables, Numbers or Graphs. Interpretation is the process of assigning meanings to the findings. Interpretation ties the findings to the research problem. Discussion of findings is tying the interpretation to the literature review. This is where you identify concurrence or contrasts in literature. A good research report should include:

1. The statistical test used.
2. The sample size and relevant descriptive statistics.
3. The test statistic and degrees of freedom, where applicable.
4. The p-value.
5. The effect size and confidence interval, where appropriate.
6. A clear interpretation linked to the research question.

NOTE:

1. Always start with descriptive statistics followed by inferential statistics.
2. Statistical significance should not be confused with practical or educational importance. An effect can be statistically significant but too small to make a meaningful difference in practice.