A hypothesis by definition is a proposition or a number of propositions that reflect a prediction. In simpler words, a hypothesis is a statement that assumes a relationship between an independent variable and a dependent variable. In research, hypotheses provide the basis for data collection and data analysis. The data analysis is essentially applications of research tools and techniques to prove or disprove the hypothesis. However, before testing, it is important to frame the hypotheses properly. This sometimes poses a challenge.
Need for framing a hypothesis from research questions
Not all types of research require a hypothesis. The need for framing a hypothesis stems from the research questions and the research methodology. It depends on the research approach, research type, and research method. Framing a hypothesis is essential in cases:
When the research type is descriptive
Descriptive research is the one that aims to collect information and analyze it statistically to draw conclusions. For example, let a primary study want to investigate the effects of organizational factors on job satisfaction of employees. Moreover, several organizational factors identified in the literature review are:
- relation with employers,
- job engagement,
healthyenvironment at work and, scopeof promotion.
The hypothesis, in this case, will test the effect of these factors on job satisfaction of employees.
When the research approach is deductive
A deductive approach of research emphasizes causality and aims to test a theory with the help of a hypothesis. Suppose one wants to examine the Economic theory of export-led growth. This theory states that a country can achieve an accelerated rate of growth by relying on the expansion of exports. There are several factors such as foreign direct investment (FDI), trade openness, exchange rate, bilateral and multilateral trade agreements among others that the literature identifies as related to export expansion. In this case, the hypothesis will be framed to determine the effects of these factors on the growth rate.
When the research method is quantitative
A quantitative research method is one that is based on the measurement of variables. It useful to assess the effect of GDP and the current account deficit of a country on its fiscal deficit. GDP, current account deficit and fiscal deficit all are measured and published in monetary units. The hypothesis, in this case, will be framed to measure the effect of GDP and CAD on the fiscal deficit.
Types of hypothesis
A hypothesis can be stated in a number of ways. Consider the example research on student performance (grades) in relation to
- The hypothesis of difference. There is a significant difference between the average performance of students who receive counseling and those who do not.
- The hypothesis of association. There are equal numbers of students in the classroom who receive counseling and who do not receive counseling.
- Null hypothesis. There is no relationship between counseling and the grades received by students in the classroom.
- Alternative hypothesis. There is a significant relationship between counseling and the grades received by students in the classroom.
Furthermore, if the research predicts that there is a significant relationship between counselling and student performance. The alternative hypothesis reflects this prediction. The null hypothesis is framed in such a way that it can be refuted to confirm the alternative hypothesis.
Format of a hypothesis
Consider the example of organizational factors and job satisfaction mentioned above.
- Make a flow chart before framing the hypotheses. This is called a conceptual framework and it helps in framing the hypotheses in a systematic way. In the conceptual framework, list the independent variables or the factors on the left-hand side. The dependent variable ‘job satisfaction’ should be on the right-hand side. Use an arrow in between. The directionality of the arrow should be from the independent variables to the dependent variable.
- Following the conceptual framework, the independent variables should come on the left-hand side of the hypothesis. The dependent variable should be on the right-hand side of the hypothesis.
- Include words like ‘impact’, ‘influence’, ‘effect’, ‘relationship’ or ‘association’ within the hypothesis. This is to indicate what tests can be used in testing it.
- Use notation H0 to denote the null hypothesis.
- Use notation HA to denote the alternative hypothesis.
Following the above rules, the null and the alternative hypotheses in the case of the above example are:
H0: Organizational factors given by remuneration, rewards, relation with employers, training, job engagement, workload, healthy environment at work and scope of promotion have no effect on the job satisfaction of employees.
HA: Organizational factors given by remuneration, rewards, relation with employers, training, job engagement, workload, healthy environment at work and scope of promotion have a significant effect on the job satisfaction of employees.
Steps for constructing a hypothesis
- The first step before constructing a hypothesis is a thorough review of existing literature on the topic of research.
- After the literature review, identify gaps in the literature. Then narrow down the research problem to fulfill the gap.
- The research problem needs to be stated in terms of research objectives or research questions.
- Following the research question, identify the dependent and the independent variables.
- Frame statements or hypotheses that reflect a prediction and are testable.
- The results of hypothesis testing directly help to answer the research questions and draw conclusions for the study.
While framing hypotheses, these are the important points one needs to remember.
- The hypothesis should be precise and clear.
- It should be stated in simple terms.
- The hypothesis should propose a relationship between two variables or a set of variables namely dependent and independent variables.
- The scope of the hypothesis should be specific and narrow.
- The hypothesis should conform to the research questions.
- It should be consistent with the findings of the previous researches or facts that are known and established.
- The hypothesis should be testable with primary or secondary data.
- The results of hypothesis testing should address the study’s aim and objectives adequately.