Biology · Introductory biology · Concept
Designing a controlled experiment
An experiment tests a hypothesis by changing one factor and measuring the effect. The independent variable is the factor the experimenter changes; the dependent variable is what is measured; controlled variables are held the same so they cannot explain the result. A control group, treated the same except for the independent variable, gives the baseline. Random assignment and replication guard against confounders and chance, so a difference between groups can be credited to the treatment.
Hypothesis and prediction
A hypothesis is a testable explanation; a prediction states what you should observe if it is right. The hypothesis that light intensity limits photosynthesis predicts that pondweed in brighter light gives off oxygen bubbles faster.
Variables
Every experiment names its variables before it begins.
| Variable | Meaning | In this experiment |
|---|---|---|
| Independent | The factor changed on purpose | Light intensity |
| Dependent | The outcome measured | Oxygen bubbles per minute |
| Controlled | Factors held the same in every group | Temperature, CO₂, species and plant size |
Controls
A negative control shows what happens without the treatment, such as pondweed kept in the dark; a positive control shows the method can detect an effect when one should occur. Without a control there is nothing to compare the treatment with.
Replication and random assignment
One plant per group cannot separate the treatment’s effect from one plant’s quirks. Use several independent experimental units, called biological replicates, in each group, and assign them to groups at random so that differences you did not measure spread evenly between the groups. Measuring the same plant three times is a technical replicate: it shows how precise the measurement is, not how much plants vary.
| Part | Plan |
|---|---|
| Question | Does fertilizer make bean seedlings grow taller? |
| Independent variable | Fertilizer: water only or fertilizer |
| Dependent variable | Plant height after three weeks (cm) |
| Controlled variables | Bean variety, light, water, soil and pot |
| Groups | 10 water only (control), 10 fertilizer |
| Confounder addressed | Bench position: assigned at random |
Correlation is not causation
In an observational study the researcher only records what happens. A confounder, a factor that affects both variables, can link them without one causing the other. Only an experiment with random assignment supports a claim that one causes the other; Statistics can then test whether a difference between groups is larger than chance.
Common mistakes
- Changing two factors at once, such as light and temperature: the result cannot be credited to either.
- Describing the control group as having no variables: it is the group without the treatment, treated the same in every other way.
- Counting repeated measurements of one subject as replicates: they show measurement error, not variation between subjects.
- Concluding cause from a correlation in observational data: a confounder may explain both variables.
Key terms
- Hypothesis
- A testable explanation for an observation. It leads to predictions that an experiment can support or contradict; support makes it more credible but never proves it for good.
- Independent variable
- The factor an experimenter changes on purpose to test its effect. In a test of light on photosynthesis, light intensity is the independent variable.
- Dependent variable
- The outcome measured to see whether it depends on the independent variable, such as the rate of oxygen release in a test of light on photosynthesis.
- Controlled variable
- A factor kept the same in every group so that it cannot explain a difference between them, such as temperature and species in a test of light on photosynthesis.
- Experimental control
- A group or condition used for comparison. A negative control should show no effect; a positive control should show a known effect, proving the method works.
- Experimental unit
- The thing a treatment is applied to independently, such as one plant or one dish. Measuring the same unit twice does not make two units.
- Biological replicate
- An independent biological sample, such as a separate plant or mouse, that captures the natural variation between individuals.
- Technical replicate
- A repeated measurement of the same sample. It shows measurement error, not variation between individuals, so it does not add to the sample size.
- Random assignment
- Using chance to decide which units get which treatment, so other differences spread evenly across the groups. Random sampling, choosing who is studied, is a different step.
- Confounder
- A variable linked to both the treatment and the outcome, which could explain a difference instead of the treatment.
- Observational study
- A study that records what happens without assigning treatments. It can show an association, but not on its own that one thing causes another.
- Causal claim
- A claim that changing one thing changes another. A controlled, randomized experiment can support it; a correlation alone cannot.
Work through an example
A student tests whether a fertilizer makes bean plants grow taller. She gives fertilizer to one plant on a sunny windowsill and none to one plant in a shaded corner. After three weeks the fertilized plant is 4 cm taller. Name the independent and dependent variables, find two flaws, and redesign the experiment.
Fix a flawed experiment →Sources and scope
Authored study material. Tool results depend on the stated inputs and model assumptions.
Study it, then try it
Practice the key terms as flashcards, then open the example in its tool, change a value and keep a useful result on your board.
Study the 12 key terms Plan it in the Experiment planner Open worked example on a board Experimental design in Biology ReferenceYour existing work stays on this device. Examples open as editable copies.