Hypothesis: definition; characteristics; types; difference & importance

Introduction

A good research question should guide you to formulate the right hypothesis of your study. But before we go in to details, what is hypothesis and what does it entail? Remember we are defining hypothesis NOT research Hypothesis!

Definition

There are diverse definitions that refer to hypothesis. Some of them are as follows;

Definition 1: Hypothesis is official uncertain statement that the researcher frames to portray the expected correlational perspective between two or more variables under investigation or being studied.

Definition 2: Hypothesis is a rational forecast of a specific event happening before the researcher carryout any empirical confirmation or evidence to support it.

Definition 3: Hypothesis in scientific expression. It is a tentative theory or testable statement about the association between two or more variables.

 Therefore, a hypothesis will guide the audience on how or why an event happens or occurs in the natural phenomenon. It achieves this objective by giving a clear explanation using facts on the ground or using some reasonable/logical assumptions although not yet tested for approval.

NB: That, a research question should lead the researcher to hypothetical claim of logical expected outcome of the research study.

Therefore, a hypothesis is a scientific suggestion of the expected research finding of the association between or amongst the study variables and this means the hypothesis should be specific, testable and falsifiable. Therefore, these three key aspects drives us to curiously establish the key characteristic indicators of a good research hypothesis.

Sources of hypothesis

How does hypothesis come about? Where does this proposition, supposition, suggestion, premises or theory come from? The following are some of the origins of a hypothesis. These are;

1). Premonition or intuition/personal insight;

This is an original idea or a virgin idea that one comes up with. This implies that it is an idea that has never been in existence before. Although virgin ideas may bring in unique contribution(s) to the body of existing knowledge, it may suffer two major challenges, one; the association that is discovered between the two or more variables as per the hypothesis, may fail to be in existence in other studies. Two; the concept or relationship may lack other similar theories to back it up. But they can be a Centre of interest for further research by scholars. In fact, it can translate into an explanatory theory.

2). Research Findings

From other research findings as per past literature review, one can develop a new hypothesis. Once the hypothesis is proven through empirical tests, then this confirms the previous studies. Hence no re-inventing the wheel by the researcher although it will always appear as if it is a replication of past studies which were conducted in distinctly different conditions. In fact this is the reason as to why most of the institutions of higher learning like Universities incorporate a chapter section of SUMMARY, CONCLUSIONS AND RECOMMENDATIONS mostly in chapter four or five of the project and/or thesis work so as to portray that the current hypothesis is in tandem with the existing research findings and so nothing like re-inventing the wheel.

3). Existing Theory or set of Theories

A Theory is a supposition or a system of ideas intended to explain something, especially one based on general principles independent of the thing to be explained.

OR

A THEORY is the natural, observable, logical, realistic and practical working or relation of two or more characteristics or behavior of a particular subject matter whether of a natural person or otherwise.

Therefore, a theory is a logical relationship between or amongst variables and it is also referred to as conceptual framework. Out of the existing conceptual framework/theory, the researcher out of logical deductions may establish a hypothesis.

4). Social cultural setting

Cultural values and beliefs initiate hypothesis development by social researchers/scientists. They carry out careful observations and generates a number of testable suppositions or premises in the format of hypothesis. So, most of the times, hypotheses may carry the same message but in different contextual setting.

5). Analogy

An analogous situation is a case of similarity, equivalence or likeness which the researcher can replicate to form a hypothesis. To achieve this goal, the researcher needs to test the analogy relationship with a similar characteristic in a different setting or environment. It should be noted that the success of this endeavor is pegged on the researcher’s appreciation of the theory underpinning the analogy and its relevance to the new hypothesis. 

6). Personal Experience

Past experience of the researcher may help in designing the hypothesis. From inference due to continuous interactions and exposure by the researcher can lead to a way of forming a hypothesis. The researcher will establish research questions and look for tentative answers (hypothesis) which will solve the problem at hand in the future. So, the researchers unique life history personal exposure influences his/her perception and conceptualization and hypothesizing of issues.

Characteristics of a good hypothesis

A good hypothesis has the following characteristics

1. Must have conceptual clarification
2. Referent in empirical viewpoint
3. Objectivity
4. Specificity
5. Relevant
6. Testability
7. Consistency
8. Availability of technique
9. Purposiveness
10. Verifiability
11. Productivity of effects
12. Economical

All these viewpoints of high quality for research purposes are as explained as follows;

2.1 Conceptual Clarity

A good quality hypothesis should maintain clarity in terms of concept definition. The researcher has to ensure that the concept and the study variables used in the study carry the intendent implication to avoid any ambiguity or confusion. This is achieved if the process of operationalization of the study variables is firm and distinct. Operationalization means a way of putting a measurement indicator on a variable to make it observable and measurable. For instance, if we want to measure ones intelligence, we can consider class performance based on marks scored in certain subject.

You see, you cannot touch ones intelligence for all heads look alike but what is inside (i.e. intelligence level) is invisible hence cannot be touched or seen. Therefore, with an examination test we can rank those who have highly scored marks like 80% to 90 or 100% as having high level of intelligence. Otherwise, it will mean medium or low intelligence level. In short the researcher must define how the variable will be manipulated and measured in the study. By clearly detailing the specifics of how the variables were measured and manipulated, other researchers can better understand the results and repeat the study if need be.

2.2 Empirical reference Origin

The hypothesis should be anchored on some empirical literature from past studies. A hypothesis cannot be a stand-alone as far as its origin is concerned. In other words am saying that a researcher cannot wake up one morning and decide to formulate a hypothesis from nowhere. A hypothesis should be from other researched work or findings. The researcher should consider drawing hypothesis from previously published research work based on the theory.

2.3 Objectivity

The hypothesis should observe objectivity in various aspects such as data collection to avoid researcher’s biasedness when choosing the study variables.

2.4 Specificity

 A good hypothesis ought to be specific and not general to the contrary so as to be able to straightaway predict the expected correlation between the study variables. In other words, a hypothesis is a specific, testable prediction about what you expect to happen in a study, so buying a leaf from previously published studies which have been based on the theory to develop a hypothesis is much in order.

To achieve the objective of specificity of the hypothesis, one should ask himself or herself the following questions;

1. Is the dialectal aspect clear and fixated?
2. Is there any correlation between your hypothesis and your research topic? The hypothesis and the topic of your study should be harmonious or in congruence.
3. How can I test my hypothesis if I claim it is testable?
4. In my new exploration what explanations in my study do I wish to portray to my audience?
5. Is my study variables, both independent and dependent well-articulated in my study?
6. Is it possible to manipulate my study variables without violating ethical standards?

The aforementioned questions ensure that your hypothesis is based on a sure underpinning. They can also be useful in guiding you on revising your hypothesis to eliminate any weaknesses therein.

2.5 Relevance

The hypothesis should be aligned to both the research questions and the research objectives to make sense. On the same breath, it should match the theory to be tested for approval or disapproval.

2.6 Testability

There must be a method to test the validity of the hypothesis to ensure that the researcher does not make moral judgment on the study variables. This ensures that the variables are measurable and verifiable. The quality of testability is key for it portrays in a clear manner how the independent variable can be manipulated and how the dependent variable can be measured in a certain population so as to depict the relational links thereof. You see, a good hypothesis will state the cause-effect implications between the predictor variable and the dependent or criterion variables which are inferentially evaluated to prove their validity.

2.7 Consistency:

A good hypothesis should portray an element of constancy, steadiness or uniformity in every way. Such as being underpinned by a body of theories, research findings and other inter-linked hypothesis. It should further show a connection with existing knowledge.

2.8 Simplicity

A hypothesis should be well understood for it is simple to the users. Therefore, the assumptions and conditions operationalizing the hypothesis should be simple.

2.9 Availability of Technique(s)

A scientific method should be identifiable that can be used to test the proposed hypothesis.

2.10 Purposiveness

The established hypothesis should be for a specific purpose. That is, it should be formulated to answer the research problem and meet the specific research objective. 

2.11 Verifiability

It should be possible to practically verify or validate the hypothesis if it represents good features.

2.12 Profundity of Effect

It means a good quality hypothesis should portray philosophical effect upon a variety of study variables. In other words it should portray a level of understandability of a concept.

2.13 Economical

A hypothesis should show cost-benefit picture. Such that the benefits that accrue to the research at hand is more than the financial or non-financial resources used to facilitate the research exercise.

Types of hypothesis

Hypotheses may be of diverse nature based on the criteria used. The various types of hypotheses are categorized as per several distinct criteria such as;

i). Number of independent and dependent variables in an empirical model Criteria

Under this classification, we have simple and complex hypothesis.

1. Simple Hypothesis

This type of hypothesis is also referred to as composite hypothesis. It is characterized by specification of all the parameters used. This hypothesis forecasts the link between two variables, namely; independent and dependent variables.

Example

1. If you overfeed at night you will have stomach upsets in the following day.
2. Early morning simple body exercises makes one feel fresh the whole day.
3. Frequent medical checkups prepare you psychologically for body health warnings.

2. Complex Hypothesis:

As the name implies, this type of hypothesis portrays the link between more than two independent variables on one side and two or more than two dependent variables. You see, for the simple one as it has been discussed earlier on, it only involves one independent variable and one dependent variable. But for complex case, the independent and dependent variables on both sides of the empirical model are more than one.

Example

1. Students who have both tuition fees payment and academic performance challenges end up dropping from school in the early and committing suicide.
2. Highly favored employees by employers at the workplace and in their resident places get frequent promotions and shift to their own occupier houses.

ii). Level and nature of outcome assurance Criteria

under this perspective, we have research hypothesis and logical hypothesis

3. Research Hypothesis

It is also referred to as empirical or working hypothesis and it represents a more than sure event occurring hence it is a specific, clear predictor of the event to occur in the future. The hypothesis is based on specific factors of the population.

Example

1. A vehicle will consume the same amount of fuel in Litres even if the prices increase.
2. Daily brushing of teeth will reduce gum health problems.

4. Logical Hypothesis

It is a hypothesis which represents a planned elucidation with inadequate proof. In other words, the suggestion has no actual evidence. The argument is only pegged on reasoning or deduction for there is no actual data.

Example

1. Whoever is faithful with little will be faithful with more resources.
2. Those who do not fear at night when alone will remain fearless during the day.

iii). Basis of either population or sample Criteria

Under this perspective, there is only one class known as statistical hypothesis as explained in number 5.

5. Statistical Hypothesis

It is a hypothesis whose basis suggestion is on the sample tested which represent the population. In this case, the statistical evidence (sample data available) is relied upon to make general conclusion of the population. It is a type of hypothesis, which is used for confirmation purposes and that is why it is referred to as confirmatory data analysis, which carry the population parameter assumption.

Example

1. On average, the former forth year Sociology students of California University had passed their examination.
2. 70% of the owners of small and medium enterprises had difficulties in accessing loan facilities from commercial banks after Covid-19 Pandemic episode.

iv). Direction of the Association/relationship between the variables Criteria

The possible link between the variables can be witnessed in the type of hypothesis used by a researcher. Under this classification, there are both directional and non-directional hypothesis.

6. Directional Hypothesis

This is a type of hypothesis which is biased to a particular direction of a relationship of the participating variables. Such that the hypothesis supports a more than (>) or less than (<)direction of the relationship of the variables of interest. For instance, the researcher postulates that;

“Those students who actively participate in religious activities in their respective religious affiliations perform better in their religious subjects respectively than those who does not. This is a greater than hypothesis which is assessed using right hand tailed test”.

Also,

“Those children who does not feed on starch foods are less active in school extra curriculum activities as compared to those who feed on starch foods. This is a less than (<) hypothesis which is assessed using left hand tailed test”.

7. Non-directional Hypothesis

This is a type of hypothesis which is not biased to any particular direction of a relationship of the participating variables. Such that the hypothesis does not support a more than (>) or less than (<) direction of the relationship of the variables of interest. For instance, the two examples used in the case of directional hypothesis can be postulated as;

Case one: “There is no significant difference in respective religious subject performance between those students who actively participate in religious activities in their respective religious affiliations and those who do not. This is a non-directional hypothesis which is assessed using two tailed tests”.

Case two:

“Those children who does not feed on starch foods and those who does are statistically the same as far as active participation in school extra curriculum activities is concerned.” This is a non-directional hypothesis which is assessed using two tailed tests”.

v). Causal and Correlational Criteria

Under this perspective, there are two common types of hypotheses, namely; associative and causal hypotheses.

8. Associative Hypothesis

This is a hypothesis which portrays the level of strength or weakness of the association between two variables. It postulates the direction and level of strength two variables only. The direction can be a positive or a negative while the level of association can be weak or strong. This type of hypothesis is measured using Pearson product-moment correlation coefficient approach.

Example

a). There exists a strong positive relationship between the weather changes and the level of mango productivity. Implying that the expected relationship is that when the weather is conducive, more mangoes are produced or increase in a large way.

b). The association between customer price and demand level for night dresses is weak and negative. Implying that the expected relationship is that when the price increase, night dresses demanded decrease slightly.

9. Causal Hypothesis

This is a hypothesis which portrays the cause-effect relationship between or amongst two or more variables. It postulates the significant level of the influence of the variables used in the empirical model.  The cause-effect relationship can either be statistically significant or not statistically significant. This type of hypothesis is measured using inferential statistics such as F-test, R-Squared (R2) or t-statistics.

vi). Theory Development Criteria

Majorly, suppositions can be based on how the researcher infer the relationship based on general or specific theory development. Hence we have the inductive and deductive hypothesis.

10. Inductive Hypothesis

This is a type of hypothesis which is postulated by moving from specific to general argument or conclusion. In other words, this hypothesis advocate for Generalizing from the specific observations to the general conclusion. At the end of it all, the outcome is building of a new theory. Inductive hypothesis works well where there is no theory in existence.

11. Deductive Hypothesis

This is a type of hypothesis which is postulated by moving from to general to specific argument or conclusion. In other words, this hypothesis advocate for Generalizing from the specific observations to the general conclusion. At the end of it all, the outcome is falsification or verification of existing theory. Deductive hypothesis works well where there exists a theory which the researcher is either trying to approve or disapprove.

vii). Statistical Significance Criteria

This criteria focuses on the level of statistical significance of the cause-effect relationship between two or more variables in an empirical model. Under this guideline, we have two commonly used hypotheses, that is Null and Alternative Hypotheses.

12. Null Hypothesis

The hypothesis is expressed in a general way with connotation that there exists no relationship between or amongst some variables. It is symbolically expressed as H0.

Now, in a study, a prediction on the event to occur is expressed in a manner that no link exists. There are two ways in which the statement should be expressed.

One;

“There is no significant relationship between X and Z”.

This is the most common expression used to imply a null hypothesis by most of the researchers or scholars. Caution is needed here for this expression has a shortcoming for it implies that there is an element of pre-emptying the expectations of the researcher.

You see, if you say there is no significant relationship, it means you already know that there are no reliable results expected at the end of it all. In other words, you know your proposition or supposition is not appropriate. Such a statement can be likened with a scenario like that of your friend telling you that he is wishing to visit the boss in his/her office in a certain locality and at the same time he/she has enough information that he is NOT IN THE OFFICE. So, he puts it this way, “The boss is not in his office and I am going there to meet him.” This means your friend is simply wasting his resources to make a visit to an office when he pretty knows that the occupant is absent. This takes us to the second expression of null hypothesis.

Two;

To rectify this common expression in research, the researcher needs to express null hypothesis as follows;

“The relationship between X and Z is not statistically significant.”

This implies that the researcher has first of all appreciated that there is a relationship that exist between X and Z. Then he/she goes ahead and declares that it is not significant. You see, putting it this way is also away of implying relationship aspect is what is taking you to the field to collect data.

This is the way to set the null hypothesis although sometimes it is dictated by the institution’s or sponsor’s format adopted.

Example

1. There is no significant relationship between Variable P and Variable Q.
2. The relationship between the IQ level of University students and their academic performance is not statistically significant.

13. Alternative Hypothesis

An alternative hypothesis is the opposite of null hypothesis for it is a statement which portrays some statistical significance between two variables and it is usually expressed as H1 or HA.

In any study, the researcher is after rejection of the null hypothesis to accept the alternative hypothesis. When this happens, implies that the researcher’s proposition or proposal was correct. When testing hypothesis in research, the null and alternative hypothesis are commonly utilized.

Example

1. Variable P is better in performance as compared to Variable Q.
2. IQ level does not imply better academic performance of University students.

So, what is the difference between null and alternative hypothesis? The differences are as portrayed by Table1.1 below

Table 1.1: Difference between Null & Alternative Hypothesis

Importance of hypothesis

There are numerous roles that a hypothesis plays in a study. This include and not limited to;

1. Ensure that the researcher adheres to scientific approach in all research activities.
2. It guides on the level of success or failure of a research for the hypothesis expression can tell it all.
3. It backs the theoretical foundation forming the study-hypothesis is a statement that guides the researcher towards the theories which underpin the study.
4. The researcher will comfortably rely on the hypothesis to analyze data and also measure the validity and reliability of the research.
5. Hypothesis eliminates all ambiguity for the researcher is focusing on the point enquiry such that enquiry of problem is straight forward.
6. Hypothesis builds up the necessary techniques or methods required in a research process. Almost in every stage of research, there are specific methods required, for instance, sampling techniques, data analysis and even operationalization methods used on variables. Hypothesis guide in to these methods in research.
7. Hypothesis help in streamlining the focus of the researcher such that he/she concentrates on the relevant topic, facts and observations from his study.
8. Hypothesis guide the researcher towards the correct factors or variables being studied. This is achieved when the right formulation of hypothesis is done.
9. It is through the set hypothesis that the researcher arrives at new knowledge and research finding.
10. Hypothesis increases accuracy and correctness which is required or expected in scientific investigation.
11. Hypothesis is the bridge between theory and research activity. This is because it prompts scientific investigation.
12. Time saving and economical benefits-hypothesis ensures that the right thing is done in a study leading to optimal utilization of the research resources.  
13. Hypothesis helps during data collection for the researcher is well guided to go for the relevant data or the right information needed in the study
14. Hypothesis is the yard stick to proper conclusions in research. You see, the hypothesis acts as a pointer to the keynote issues in a study so as to be able to make the right conclusions.