How to write a hypothesis

A hypothesis is a suggested explanation that can be tested, turned into a prediction in the form "If..., then...". A good one has four qualities:

A small glass vial holding a green plant sprig in liquid.
A plant cutting in a vial, set up for a testable hypothesis.Photo: katetsui / Pixabay
  1. 1

    Testable

    An experiment or study can check it.

  2. 2

    Falsifiable

    Some result could prove it wrong.

  3. 3

    Specific variables

    What you change and what you measure.

  4. 4

    Grounded

    Based on observation or earlier research.

Write your paper in Evernote. The AI Assistant builds your table, finds the exact line and plans the work.

Start your paper in Evernote

How to write a hypothesis in 5 steps

  1. 1

    Start with a question

    Start with an observation that leads to a question: "Why do plants near the window grow taller?"

  2. 2

    Propose an explanation

    Write a suggested answer you can test: "Plants grow taller with more light."

  3. 3

    Name the variables

    The independent variable is what you change (hours of light). The dependent variable is what you measure (height).

  4. 4

    Turn it into an if-then prediction

    "If bean plants get 12 hours of light a day instead of 6, then they will be taller after 3 weeks."

  5. 5

    Check it can be tested and disproven

    Ask what result would prove it wrong. If none could, rewrite it. For a statistical test, also write the null hypothesis.

Hypothesis example

An introductory biology lab experiment

Light and plant growth

1Question

Do bean plants grow taller with more light each day?

2Explanation

More light gives the plants more energy for growth.

3Variables

  • Independent: hours of light per day (6 or 12)
  • Dependent: plant height after 3 weeks, in cm
  • Controlled: soil, water, pot size, temperature

4Prediction

If bean plants get 12 hours of light a day instead of 6, then they will be taller after 3 weeks.

5Null and alternative

  • H₀: the mean height is the same for 6 and 12 hours of light.
  • Hₐ: the mean height is greater with 12 hours of light.

Types of hypothesis

You may need more than one, depending on the class and the analysis.

TypeWhat it saysExample
Research hypothesisYour proposed explanationPlants grow taller with more light.
Prediction (if-then)What you expect to observeIf plants get 12 hours of light, then they will be taller.
Null (H₀)No effect or no differenceMean height is the same in both groups.
Alternative (Hₐ)The effect you test forMean height is greater with 12 hours.

OpenStax Concepts of Biology, section 1.2 "The Process of Science"; OpenStax Introductory Statistics 2e, chapter 9 (null and alternative hypotheses); UNC Writing Center, "Scientific Reports".

Free hypothesis template

Fill in the brackets for your experiment. Copy it, or save it to your Evernote account with your lab notes.

Question: [what you want to find out]

Explanation: [your proposed answer]

Independent variable: [what you change]

Dependent variable: [what you measure, with units]

Prediction: If [change to the independent variable], then [effect on the dependent variable].

It would be wrong if: [result that would disprove it]

Trusted by Millions Worldwide

4.4

2,100+ reviews on G2

4.4

8,200+ reviews on Capterra

4.4

73,000+ reviews on App Store

248M

Registered users

Hypothesis questions

How do I start a hypothesis?

Start from your question, then write a testable answer. You can phrase the prediction with "If..., then...".

What is the difference between a hypothesis and a prediction?

A hypothesis is a suggested explanation. A prediction is what you expect to observe if the hypothesis is right, written "If..., then...".

What makes a hypothesis testable?

You can design an experiment or study whose results could support it or show it is wrong. A claim about what a bear is thinking is not testable.

What does falsifiable mean?

That some possible result would disprove it. "This painting is beautiful" cannot be disproven, so it is not a scientific hypothesis.

What is a null hypothesis?

In statistics, the claim of no effect or no difference, written H₀. You test whether the data give enough evidence to reject it.

Can a hypothesis be proven true?

Not with certainty. Results can support a hypothesis or reject it; repeated support from many tests builds confidence.

Where does the hypothesis go in a lab report?

In the introduction, after the background that explains why the hypothesis is reasonable.

How to plan an experiment with Evernote

Keep the lab handout, your notes and your data in one notebook. The AI Assistant helps you lay out the variables and turn your question into a testable prediction.

  1. 1

    Table the variables.

    Ask the AI Assistant for a table of the independent, dependent and controlled variables and how you measure each. It reads your notes and adds the table to your hypothesis note.

  2. 2

    Find the hypothesis rule.

    Ask what the handout says the hypothesis must include. The Assistant answers with a link to the handout, and the matching words are highlighted on the PDF page.

  3. 3

    Plan the experiment.

    Ask it to add a task for each experiment step, with due dates before the lab. You get a reminder for each one.

  • You can also photograph your data sheet; search finds the words in the photo.
  • You can also annotate the handout PDF with the variables you chose.
  • You can also ask the AI Assistant to draw your experiment as a diagram.

Save your sources in one notebook

Save your articles and PDFs in one notebook, let the AI Assistant build the table you need, and get a task for each step.

Start for free