How to Write a Hypothesis | Explanation & Examples
Writing a hypothesis is a cornerstone of virtually any quantitative research. A strong hypothesis should be testable, allowing you to collect evidence that will support or refute it. However, hypotheses can be complex, and writing a good one takes time and effort. Below is a breakdown of steps you can follow to set yourself up for hypothetical success!
- People will retain more information when text is written in an easy-to-read font compared to a hard-to-read font.
- The amount of snow that falls in December will be correlated with daily average temperatures the following April.
- A plant watered with grape soda will grow faster than a plant watered with cream soda.
Clearly, some of these hypotheses are more likely to be true than others, but they share a key feature: they all make a prediction that can be tested by conducting some form of research.
If you feel stuck writing a hypothesis, QuillBot’s free AI writer can help you refine and better express your ideas.
What is a hypothesis?
A hypothesis (plural: hypotheses) is a clear and testable prediction about a research question. Hypotheses are generally statements (rather than questions) and are common in quantitative research. Hypotheses are written before running experiments and collecting data. They are evaluated using scientific evidence and statistical testing.
A hypothesis outlines a relationship between two or more variables.
- In experimental research, the experimenter typically measures the dependent variable, which is the outcome of interest or the thing under impact. The independent variable is the thing that’s driving that impact, and it is generally manipulated by the experimenter.
- In correlational research, rather than manipulating an independent variable and measuring a dependent variable, you measure two variables and see how they change with each other. Though you cannot infer causality with correlational research, it is helpful when your variables of interest cannot be directly manipulated.
Null hypothesis vs. alternative hypothesis
When you’re reading about hypotheses, you’re bound to come across the terms “null hypothesis” and “alternative hypothesis.” These two constructs form the basis of statistical hypothesis testing.
The alternative hypothesis is your prediction about an effect or relationship between two variables. It basically states that something is happening. For example, your alternate hypothesis may be that drinking green tea increases mood.
The null hypothesis, on the other hand, states that no effect or relationship exists. For example, the null hypothesis corresponding to the alternative hypothesis above would be that drinking green tea has no effect on mood. Note that the null hypothesis is not the opposite of an alternate hypothesis; it’s more so the absence of one.
Hypothesis testing is the process of evaluating your alternative hypothesis. To do so, you collect data. You then analyze these data using statistical tests to see if there’s an effect. If your results are statistically significant, you can reject the null in favor of the alternative hypothesis.
The experiment and data collection: Participants are randomly assigned to one of two groups: tea drinking and warm water drinking. Participants first complete a questionnaire to assess their happiness, then spend a week drinking a cup of their assigned beverage every day. At the end of this week, they complete the happiness questionnaire a second time.
The statistical analysis: We might compare changes in happiness between the water and tea drinking groups using a t test. This statistical test would tell us if a significant difference exists between the two groups. Generally the p value is used to assess significance (i.e., p < .05).
The conclusion:If we find a significantly greater increase in happiness for the tea drinkers compared to the warm water drinkers, we would reject the null hypothesis in favor of our alternative hypothesis.
Note that you don’t “prove” or “disprove” a hypothesis; you simply determine whether you have sufficient evidence to support or reject it.
What makes a good hypothesis?
A strong hypothesis should be clear, testable, and based on existing research. We break down each of these qualities below.
Clear: Vagueness is the enemy of science. An unclear hypothesis can dilute the impact of your findings and open your work up to criticism from a grader or reviewer. Be sure to clearly specify the relationship or effect you expect to see to guard against this.
Testable: If you start by writing a hypothesis that is not testable, your research will come to a shuddering halt! A testable hypothesis is one that can be proven or disproven through measurement and experimentation. Think about the variables relevant to your hypothesis and how/if they can be measured.
Based on existing research: A hypothesis is a prediction, not a guess. Before writing your hypothesis, you should read relevant literature to gain a better understanding of the question you’re asking and the outcome you expect. If your hypothesis is supported by existing research, it will be more readily accepted by your fellow researchers and more likely to result in compelling data and results.
How do you write a hypothesis?
We now know what a hypothesis is, but how should you go about writing one? Use the steps outlined below as a guide to get started.
1. Determine your research question
You can’t write a hypothesis if you don’t know what you’re studying! Consider your project and the focused, specific, and researchable question you want to pursue.
2. Do some background research
In order to create a strong hypothesis, you have to have a good sense of the topic you’re studying. Look for existing studies or theories that might help you make predictions about your own findings (if you’re conducting original research, this is also a great time to make sure no one else has done what you’re doing!).
Think of how your work will extend or add to what’s out there, and be sure to save relevant sources for writing the introduction and discussion of your paper.
3. Create a rough draft of your hypothesis
Sometimes, perfectionism can get in the way of progress. Start by writing a rough hypothesis that you can think about and refine.
4. Write your null hypothesis
Once you have a rough draft of your hypothesis, writing your null hypothesis is fairly straightforward—it simply states that no effect exists. Writing your null hypothesis early on can also help you identify any loopholes in your alternate hypothesis.
5. Write your hypothesis three ways
To ensure that your (alternate) hypothesis is as strong as possible, try creating different versions of it. Be sure to specify any variables of interest and the directionality of any effects you’re expecting. It can be helpful to use an if…then or when…then framework to structure your ideas.
- If a person reads a digital document in Arial, then their words-per-minute rate will be higher than if they read in Times New Roman.
- When the font legibility is increased by removing decorative serifs, then the time required to process the text will decrease.
- There is a positive correlation between font simplicity (sans-serif) and the speed of digital text processing.
6. Refine!
By following the steps above, you’ve already started the process of refining your hypothesis. Now that you have a strong working draft, ask yourself whether your hypothesis is clear, specific, and measurable. Consider getting feedback from your peers or a supervisor, or use tools like QuillBot’s free AI Chat to identify any weaknesses that need to be improved.
Scientific hypothesis examples
The table below outlines some alternate hypotheses and their corresponding null hypotheses.
| Alternate hypothesis | Null hypothesis |
| Students who use gamified learning apps will score higher on math tests than students using traditional textbooks. | There is no difference in math test scores between students using gamified apps and those using traditional textbooks. |
| Increased exposure to blue light before bedtime leads to a decrease in total REM sleep duration. | Exposure to blue light before bedtime has no effect on the duration of REM sleep. |
| Patients taking “Drug X” will show a significant reduction in systolic blood pressure compared to those taking a placebo. | There is no significant difference in systolic blood pressure between patients taking “Drug X” and those taking a placebo. |
| Higher concentrations of microplastics in soil will result in slower germination rates for clover seeds. | The concentration of microplastics in soil has no effect on the germination rate of clover seeds. |
| Text generated by AI will be rated as “less authentic” by human readers compared to text written by professional copywriters. | There is no difference in perceived authenticity between AI-generated text and human-written text. |
Frequently asked questions about writing a hypothesis
- What is the difference between a thesis statement, a research question, and a hypothesis?
-
A thesis statement, a research question, and a hypothesis are all related concepts. Together, they form the backbone of any academic research study.
A research question is generally the starting point of any academic research. It should capture what you are trying to learn about. Good research questions are specific, feasible, and relevant. For example, “Does the use of sans-serif fonts improve reading speed in low-light conditions?”
A hypothesis is a clear and testable prediction about the research question. Hypotheses are an essential component of experimental or theoretical work but may not be present in more qualitative work. For example, “Participants reading sans-serif fonts in low light will demonstrate significantly faster reading speeds than those reading serif fonts in the same conditions.”
The thesis statement is the main argument of a paper. It synthesizes a paper’s position/results in the context of existing research. Though it usually appears in an introduction, it is often written towards the end of the research process. For example, “Sans-serif fonts significantly improve legibility in low-light environments.”
If you need help defining a term, QuillBot’s AI Chat is a great way to discuss and dissect new concepts.
- What is a null hypothesis vs an alternative hypothesis?
-
The null hypothesis (H0) and alternative hypothesis (H1) are used in statistical hypothesis testing.
The null hypothesis states that there is no relationship between two variables nor effect of a manipulation (i.e., nothing is happening). For example, “Drinking a cup of coffee does not increase reading speed.”
The alternative hypothesis predicts that there is an effect; for example, “Drinking a cup of coffee does increase reading speed.”
If your data provide sufficient evidence (i.e., yield statistically significant results), you can “reject” the null hypothesis in favor of the alternative hypothesis. In other words, you can conclude that your results support the alternative hypothesis.
Note that you can never prove or disprove a hypothesis; we can only determine whether the data provide enough evidence to support or reject it.
Terminology related to hypothesis testing is very precise and can be confusing. QuillBot’s free AI writer is a great tool to help you perfect your wording.
- What is an example of a hypothesis?
-
A hypothesis is a testable prediction of an outcome or relationship between variables.
Many take the form of “if… then…” For example, “If I drink caffeine before bed, then it will take me longer to fall asleep.”
Hypotheses can also outline predicted relationships. For example, “Increased caffeine consumption before bed reduces the number of hours of restful sleep.”
If you need more examples, QuillBot’s AI Chat can provide more illustrative examples specific to your own area of research.
- What is hypothesis testing?
-
Hypothesis testing is how you formally test research predictions (hypotheses) using statistics.
Hypothesis testing involves creating a null hypothesis and an alternative hypothesis.
- The alternative hypothesis outlines the effect you expect to find (e.g., “Getting 30 minutes of sunshine within an hour of waking up improves mood”)
- The null hypothesis states that there is no effect (e.g., “Getting 30 minutes of sunshine within an hour of waking up has no impact on mood).
Generally, you use statistical tests to obtain a p value, which lets you decide whether to reject or accept the null hypothesis (i.e., decide whether there is or isn’t an effect). You never “prove” a hypothesis; you simply determine whether you have enough evidence to accept or reject the null hypothesis.
Terminology related to hypothesis testing is very precise and can be confusing. QuillBot’s free AI writer is a great tool to help you perfect your wording.
Cite this Quillbot article
We encourage the use of reliable sources in all types of writing. You can copy and paste the citation or click the "Cite this article" button to automatically add it to our free Citation Generator.
Heffernan, E. (2026, January 12). How to Write a Hypothesis | Explanation & Examples. Quillbot. Retrieved January 13, 2026, from https://quillbot.com/blog/academic-writing/how-to-write-a-hypothesis/
