A research hypothesis defines the theory or problem your research intends to test. It is the “educated guess” of what the final results of your research or experiment will be. Before you can write the research hypothesis and its alternative, you must focus and define your research problem.
Start with a broad subject matter that excites you, so your curiosity will motivate your work. Conduct a literature search to determine the range of questions already addressed and spot any holes in the existing research.
Narrow the topics that interest you and determine your research question. Rather than focusing on a hole in the research, you might choose to challenge an existing assumption, a process called problematization. You may also find yourself with a short list of questions or related topics.
Use the FINER method to determine the single problem you'll address with your research. FINER stands for:
Feasible
Interesting
Novel
Ethical
Relevant
You need a feasible research question, meaning that there is a way to address the question. You should find it interesting, but so should a larger audience. Rather than repeating research that others have already conducted, your research hypothesis should test something novel or unique.
The research must fall into accepted ethical parameters as defined by the government of your country and your university or college if you're an academic. You'll also need to come up with a relevant question since your research should provide a contribution to the existing research area.
This process typically narrows your shortlist down to a single problem you'd like to study and the variable you want to test. You're ready to write your hypothesis statements.
Dovetail streamlines research to help you uncover and share actionable insights
It is important to narrow your topic down to one idea before trying to write your research hypothesis. You'll only test one problem at a time. To do this, you'll write two hypotheses – a null hypothesis (H0) and an alternative hypothesis (Ha).
You'll come across many terms related to developing a research hypothesis or referring to a specific type of hypothesis. Let's take a quick look at these terms.
The term null hypothesis refers to a research hypothesis type that assumes no statistically significant relationship exists within a set of observations or data. It represents a claim that assumes that any observed relationship is due to chance. Represented as H0, the null represents the conjecture of the research.
The alternative hypothesis accompanies the null hypothesis. It states that the situation presented in the null hypothesis is false or untrue, and claims an observed effect in your test. This is typically denoted by Ha or H(n), where “n” stands for the number of alternative hypotheses. You can have more than one alternative hypothesis.
The term simple hypothesis refers to a hypothesis or theory that predicts the relationship between two variables - the independent (predictor) and the dependent (predicted).
The term complex hypothesis refers to a model – either quantitative (mathematical) or qualitative. A complex hypothesis states the surmised relationship between two or more potentially related variables.
When creating a statistical hypothesis, the directional hypothesis (the null hypothesis) states an assumption regarding one parameter of a population. Some academics call this the “one-sided” hypothesis. The alternative hypothesis indicates whether the researcher tests for a positive or negative effect by including either the greater than (">") or less than ("<") sign.
We refer to the alternative hypothesis in a statistical research question as a non-directional hypothesis. It includes the not equal ("≠") sign to show that the research tests whether or not an effect exists without specifying the effect's direction (positive or negative).
The term associative hypothesis assumes a link between two variables but stops short of stating that one variable impacts the other. Academic statistical literature asserts in this sense that correlation does not imply causation. So, although the hypothesis notes the correlation between two variables – the independent and dependent - it does not predict how the two interact.
Typically used in philosophy rather than science, researchers can't test a logical hypothesis because the technology or data set doesn't yet exist. A logical hypothesis uses logic as the basis of its assumptions.
In some cases, a logical hypothesis can become an empirical hypothesis once technology provides an opportunity for testing. Until that time, the question remains too expensive or complex to address. Note that a logical hypothesis is not a statistical hypothesis.
When we consider the opposite of a logical hypothesis, we call this an empirical or working hypothesis. This type of hypothesis considers a scientifically measurable question. A researcher can consider and test an empirical hypothesis through replicable tests, observations, and measurements.
The term statistical hypothesis refers to a test of a theory that uses representative statistical models to test relationships between variables to draw conclusions regarding a large population. This requires an existing large data set, commonly referred to as big data, or implementing a survey to obtain original statistical information to form a data set for the study.
Testing this type of hypothesis requires the use of random samples. Note that the null and alternative hypotheses are used in statistical hypothesis testing.
The term causal hypothesis refers to a research hypothesis that tests a cause-and-effect relationship. A causal hypothesis is utilized when conducting experimental or quasi-experimental research.
The term descriptive hypothesis refers to a research hypothesis used in non-experimental research, specifying an influence in the relationship between two variables.
An effective research hypothesis offers a clearly defined, specific statement, using simple wording that contains no assumptions or generalizations, and that you can test. A well-written hypothesis should predict the tested relationship and its outcome. It contains zero ambiguity and offers results you can observe and test.
The research hypothesis should address a question relevant to a research area. Overall, your research hypothesis needs the following essentials:
Hypothesis Essential #1: Specificity & Clarity
Hypothesis Essential #2: Testability (Provability)
In developing your hypothesis statements, you must pre-plan some of your statistical analysis. Once you decide on your problem to examine, determine three aspects:
the parameter you'll test
the test's direction (left-tailed, right-tailed, or non-directional)
the hypothesized parameter value
Any quantitative research includes a hypothesized parameter value of a mean, a proportion, or the difference between two proportions. Here's how to note each parameter:
Single mean (μ)
Paired means (μd)
Single proportion (p)
Difference between two independent means (μ1−μ2)
Difference between two proportions (p1−p2)
Simple linear regression slope (β)
Correlation (ρ)
Defining these parameters and determining whether you want to test the mean, proportion, or differences helps you determine the statistical tests you'll conduct to analyze your data. When writing your hypothesis, you only need to decide which parameter to test and in what overarching way.
The null research hypothesis must include everyday language, in a single sentence, stating the problem you want to solve. Write it as an if-then statement with defined variables. Write an alternative research hypothesis that states the opposite.
The following example shows the proper format and textual content of a hypothesis. It follows commonly accepted academic standards.
Null hypothesis (H0): High school students who participate in varsity sports as opposed to those who do not, fail to score higher on leadership tests than students who do not participate.
Alternative hypothesis (H1): High school students who play a varsity sport as opposed to those who do not participate in team athletics will score higher on leadership tests than students who do not participate in athletics.
The research question tests the correlation between varsity sports participation and leadership qualities expressed as a score on leadership tests. It compares the population of athletes to non-athletes.
Once you decide on the specific problem or question you want to address, you can write your research hypothesis. Use this five-step system to hone your null hypothesis and generate your alternative hypothesis.
Step 1: Create your research question. This topic should interest and excite you; answering it provides relevant information to an industry or academic area.
Step 2: Conduct a literature review to gather essential existing research.
Step 3: Write a clear, strong, simply worded sentence that explains your test parameter, test direction, and hypothesized parameter.
Step 4: Read it a few times. Have others read it and ask them what they think it means. Refine your statement accordingly until it becomes understandable to everyone. While not everyone can or will comprehend every research study conducted, any person from the general population should be able to read your hypothesis and alternative hypothesis and understand the essential question you want to answer.
Step 5: Re-write your null hypothesis until it reads simply and understandably. Write your alternative hypothesis.
Some hypotheses are well-known, such as the Red Queen hypothesis. Choose your wording carefully, since you could become like the famed scientist Dr. Leigh Van Valen. In 1973, Dr. Van Valen proposed the Red Queen hypothesis to describe coevolutionary activity, specifically reciprocal evolutionary effects between species to explain extinction rates in the fossil record.
Essentially, Van Valen theorized that to survive, each species remains in a constant state of adaptation, evolution, and proliferation, and constantly competes for survival alongside other species doing the same. Only by doing this can a species avoid extinction. Van Valen took the hypothesis title from the Lewis Carroll book, "Through the Looking Glass," which contains a key character named the Red Queen who explains to Alice that for all of her running, she's merely running in place.
In conclusion, once you write your null hypothesis (H0) and an alternative hypothesis (Ha), you’ve essentially authored the elevator pitch of your research. These two one-sentence statements describe your topic in simple, understandable terms that both professionals and laymen can understand. They provide the starting point of your research project.
Do you want to discover previous research faster?
Do you share your research findings with others?
Do you analyze research data?
Last updated: 5 September 2023
Last updated: 19 January 2023
Last updated: 11 September 2023
Last updated: 21 September 2023
Last updated: 21 June 2023
Last updated: 16 December 2023
Last updated: 19 January 2023
Last updated: 30 September 2024
Last updated: 11 January 2024
Last updated: 14 February 2024
Last updated: 27 January 2024
Last updated: 17 January 2024
Last updated: 13 May 2024
Last updated: 30 September 2024
Last updated: 13 May 2024
Last updated: 14 February 2024
Last updated: 27 January 2024
Last updated: 17 January 2024
Last updated: 11 January 2024
Last updated: 16 December 2023
Last updated: 21 September 2023
Last updated: 11 September 2023
Last updated: 5 September 2023
Last updated: 21 June 2023
Last updated: 19 January 2023
Last updated: 19 January 2023
Get started for free
or
By clicking “Continue with Google / Email” you agree to our User Terms of Service and Privacy Policy