Since grade school, we've all been familiar with the term “hypothesis.” A hypothesis is a fact-based guess or prediction that has not been proven. It is an essential step of the scientific method. The hypothesis of a study is a drive for experimentation to either prove the hypothesis or dispute it.
A research hypothesis is more specific than a general hypothesis. It is an educated, expected prediction of the outcome of a study that is testable.
A good research hypothesis is a clear statement of the relationship between a dependent variable(s) and independent variable(s) relevant to the study that can be disproven.
Once you've written a possible hypothesis, make sure it checks the following boxes:
Start your research hypothesis from a journalistic approach. Ask one of the five W's: Who, what, when, where, or why.
A possible initial question could be: Why is the sky blue?
Once you have a question in mind, read research around your topic. Collect research from academic journals.
If you're looking for information about the sky and why it is blue, research information about the atmosphere, weather, space, the sun, etc.
Once you're comfortable with your subject and have preliminary knowledge, create a working hypothesis. Don't stress much over this. Your first hypothesis is not permanent. Look at it as a draft.
Your first draft of a hypothesis could be: Certain molecules in the Earth's atmosphere are responsive to the sky being the color blue.
Take your working hypothesis and make it perfect. Narrow it down to include only the information listed in the “Research hypothesis checklist” above.
Now that you've written your working hypothesis, narrow it down. Your new hypothesis could be: Light from the sun hitting oxygen molecules in the sky makes the color of the sky appear blue.
Your null hypothesis should be the opposite of your research hypothesis. It should be able to be disproven by your research.
In this example, your null hypothesis would be: Light from the sun hitting oxygen molecules in the sky does not make the color of the sky appear blue.
One of the main reasons a manuscript can be rejected from a journal is because of a weak hypothesis. “Poor hypothesis, study design, methodology, and improper use of statistics are other reasons for rejection of a manuscript,” says Dr. Ish Kumar Dhammi and Dr. Rehan-Ul-Haq in Indian Journal of Orthopaedics.
According to Dr. James M. Provenzale in American Journal of Roentgenology, “The clear declaration of a research question (or hypothesis) in the Introduction is critical for reviewers to understand the intent of the research study. It is best to clearly state the study goal in plain language (for example, “We set out to determine whether condition x produces condition y.”) An insufficient problem statement is one of the more common reasons for manuscript rejection.”
Characteristics that make a hypothesis weak include:
A weak hypothesis leads to weak research and methods. The goal of a paper is to prove or disprove a hypothesis - or to prove or disprove a null hypothesis. If the hypothesis is not a dependent variable of what is being studied, the paper's methods should come into question.
A strong hypothesis is essential to the scientific method. A hypothesis states an assumed relationship between at least two variables and the experiment then proves or disproves that relationship with statistical significance. Without a proven and reproducible relationship, the paper feeds into the reproducibility crisis. Learn more about writing for reproducibility.
In a study published in The Journal of Obstetrics and Gynecology of India by Dr. Suvarna Satish Khadilkar, she reviewed 400 rejected manuscripts to see why they were rejected. Her studies revealed that poor methodology was a top reason for the submission having a final disposition of rejection.
Aside from publication chances, Dr. Gareth Dyke believes a clear hypothesis helps efficiency.
“Developing a clear and testable hypothesis for your research project means that you will not waste time, energy, and money with your work,” said Dyke. “Refining a hypothesis that is both meaningful, interesting, attainable, and testable is the goal of all effective research.”
There can be overlap in these types of hypotheses.
A simple hypothesis is a hypothesis at its most basic form. It shows the relationship of one independent and one independent variable.
Example: Drinking soda (independent variable) every day leads to obesity (dependent variable).
A complex hypothesis shows the relationship of two or more independent and dependent variables.
Example: Drinking soda (independent variable) every day leads to obesity (dependent variable) and heart disease (dependent variable).
A directional hypothesis guesses which way the results of an experiment will go. It uses words like increase, decrease, higher, lower, positive, negative, more, or less. It is also frequently used in statistics.
Example: Humans exposed to radiation have a higher risk of cancer than humans not exposed to radiation.
A non-directional hypothesis says there will be an effect on the dependent variable, but it does not say which direction.
An associative hypothesis says that when one variable changes, so does the other variable.
An alternative hypothesis states that the variables have a relationship.
Example: An apple a day keeps the doctor away.
A null hypothesis states that there is no relationship between the two variables. It is posed as the opposite of what the alternative hypothesis states.
Researchers use a null hypothesis to work to be able to reject it. A null hypothesis:
Example: An apple a day does not keep the doctor away.
A logical hypothesis is a suggested explanation while using limited evidence.
Example: Bats can navigate in the dark better than tigers.
In this hypothesis, the researcher knows that tigers cannot see in the dark, and bats mostly live in darkness.
An empirical hypothesis is also called a “working hypothesis.” It uses the trial and error method and changes around the independent variables.
Examples:
In this case, the research changes the hypothesis as the researcher learns more about his/her research.
A statistical hypothesis is a look of a part of a population or statistical model. This type of hypothesis is especially useful if you are making a statement about a large population. Instead of having to test the entire population of Illinois, you could just use a smaller sample of people who live there.
Example: 70% of people who live in Illinois are iron deficient.
A causal hypothesis states that the independent variable will have an effect on the dependent variable.
Example: Using tobacco products causes cancer.
Make sure your research is error-free before you send it to your preferred journal. Check our our English Editing services to avoid your chances of desk rejection.