Not sure which lesson to do first? Find your level and get a plan →

Statistics · Grade 9-12

Control Group

A control group is the comparison group in an experiment that does not receive the main treatment being tested.


Quick Answer

A control group is the comparison group in an experiment that does not receive the main treatment being tested. It provides a baseline for deciding whether the treatment changes the outcome. Use one whenever you want to know what a treatment adds: the control group shows what happens anyway, from natural recovery, practice or the passing of time. If a tutored group improves by 12 points and a matching control group by 8, the tutoring's effect is about 4 points, not 12. The control must match the treatment group in every other way, ideally because both were formed by random assignment.

Why This Matters

Without a control group, every before-and-after improvement looks like a treatment effect, and almost anything seems to work: colds clear up, test scores rise with practice, plants grow in spring. The control group subtracts that background change, so a 12-point gain can turn out to be a 4-point effect. A poorly matched control is nearly as bad, because it brings its own differences into the comparison.

Intuitive Explanation

Twenty-five students in a tutoring program raise their test scores from 64 to 76, a 12-point jump. Impressive, until you ask what would have happened without tutoring. They were in school all term, taking the same classes and practising for the same test. A control group of 25 similar students who did not get tutoring went from 65 to 73, up 8 points. The tutoring's effect is the extra improvement: 12−8=4 points.

A control group is the experiment's answer to "compared with what?" It must be like the treatment group in every way except the treatment: the same kind of people, the same time period, the same measurements. The best way to get that is random assignment, so that both groups come from one pool split by chance. Controls do not always get nothing. In medicine they often get a placebo, a dummy pill that looks real, or the current standard treatment, so the question becomes whether the new treatment beats what patients would otherwise receive.

An untreated group is not automatically a good control. If students at one school try a new toothpaste and students at another school serve as the control, the schools may differ in water fluoridation, diet or dental visits, and those differences travel into the comparison. The cavity gap then mixes the toothpaste with everything else that differs between the schools. A control that differs systematically from the treatment group reintroduces confounding.

Think of it like this. A control group works like a twin who skipped the treatment: whatever happens to that twin, from growth to practice to recovery, would also have happened to the treated twin, so the difference between them is what the treatment added. Real groups are twins only on average, which is why they are formed by chance.

Core idea. A treatment's effect is measured against a comparable group that did not get it: subtract the control group's change from the treatment group's change, and what remains is what the treatment added.


When to Use

Plan a control group whenever an experiment asks whether a treatment works or how much it changes an outcome, especially when the outcome would change anyway: recovery, growth, learning, practice effects. Give it the same conditions minus the treatment, a placebo when expectations matter, or the standard treatment when withholding care would be unfair. If the task only describes one group, no control is needed; if the groups chose themselves, the comparison is observational, not controlled.

Tip. Ask: Does the study need a group that skips the new treatment, so you can see what would have happened without it?

How to Recognize It

Before crediting a treatment for an improvement, find the group that shows how much would have changed anyway.

  1. Is there a group that did not get the new treatment but was measured the same way and at the same time?

    Yes, that is the control group. If every subject got the treatment, natural recovery or practice could explain the change, so the design is missing this piece.

  2. Were the control and treatment groups formed the same way, ideally by chance?

    Yes makes the comparison fair. If the control group volunteered or was chosen differently, Confounding Variables can explain the gap, and the needed step is Random Assignment.

  3. Does the control group receive something, such as a sugar pill or the standard treatment?

    A sugar pill handles expectations, which is the Placebo Effect; the standard treatment asks whether the new one beats current practice. Receiving nothing only shows whether it beats nothing.

Control Group vs Placebo Effect vs Random Assignment vs Blinding

These four often appear in the same description of a drug trial. Ask whether the term names a group of subjects, a reaction to belief, a way of forming groups, or a way of hiding which group is which.

Control Group

Meaning
A control group is the comparison group that does not get the treatment being tested, so it shows the baseline.
Key test
Names a group of subjects and what it receives.
Example
In a vocabulary-app study, the students who keep using paper flashcards.

Placebo Effect

Meaning
The placebo effect is a real change in how people feel or report, caused by believing they are being treated.
Key test
Names a reaction caused by belief, often seen inside the control group.
Example
Headache patients given a sugar pill report less pain than patients given nothing.

Random Assignment

Meaning
Random assignment uses chance to decide which subjects go to the treatment group and which to the control group.
Key test
Describes how the groups are formed, not what they receive.
Example
Shuffled name cards decide which 25 of 50 students get the app.

Blinding

Meaning
Blinding keeps subjects or researchers from knowing who is in the treatment group and who is in the control group.
Key test
Describes who knows the group labels, not the groups themselves.
Example
Pills for both groups look identical and carry only code numbers.

Worked Examples

Example 1 — Subtract the baseline

Easy

Fifty students are randomly split into two groups of 25. The tutored group's average score goes from 64 to 76 over a term; the control group, with no tutoring, goes from 65 to 73. What is the effect of the tutoring?
Show the steps and the answer
  1. Tutored group's change: 76−64=12 points.

    This change includes the tutoring and everything else that happened during the term.

  2. Control group's change: 73−65=8 points.

    The control group had the same term, classes and practice, so its gain shows what happens without tutoring.

  3. Effect: 12−8=4 points.

    Subtracting the baseline removes the improvement that would have happened anyway.

  4. Check that the groups were comparable: random assignment, and starting averages of 64 and 65.

    The subtraction is fair only if the groups were alike at the start, which chance assignment makes likely.

Answer: The tutoring raised scores by about 4 points, not 12; the other 8 points would have come anyway.

Takeaway: A treatment's effect is the treated group's change minus the control group's change.

Example 2 — An untreated group that isn't a fair control

Standard

A company gives its new toothpaste to 300 students at School A for a year; 300 students at School B use their usual toothpaste as the control. School A averages 0.8 new cavities per student and School B 1.4. Town A's water is fluoridated and Town B's is not. Does the study show that the toothpaste prevents 0.6 cavities per student?
Show the steps and the answer
  1. Compute the gap: 1.4−0.8=0.6 cavities per student.

    The gap is real; what matters is whether the toothpaste is the only systematic difference behind it.

  2. List other differences: fluoridated water in Town A, and possibly different diets and dental care.

    Fluoride itself prevents cavities, so it is tied to both the group and the outcome.

  3. Judge the control: it differs from the treatment group in more than the toothpaste, so the comparison is confounded.

    A control group is useful only if it matches the treatment group in everything except the treatment.

  4. Redesign: within each school, randomly assign half the students to the new toothpaste and half to the usual one.

    Then water, diet and dental care are shared by both groups, and only the toothpaste differs.

Answer: No. The 0.6-cavity gap mixes the toothpaste with fluoridated water and other school differences; a fair test assigns toothpastes at random within each school.

Takeaway: A control group must be formed the same way as the treatment group; a group that merely skipped the treatment is not enough.

Example 3 — Find the mistake: no comparison at all

Application

An ad says: "After 8 weeks on our memory supplement, users' memory test scores rose from 70 to 78." Rosa writes: "The supplement raised scores by 8 points, an 11% improvement." A later randomized trial gave half its volunteers the supplement and half a placebo: the supplement group again rose from 70 to 78, and the placebo group rose from 70 to 76. Find the mistake and correct it.
Show the steps and the answer
  1. Check Rosa's arithmetic: 78−70=8 and 870≈11%.

    The numbers are right; the mistake is treating the whole rise as the supplement's effect.

  2. Name what raises scores anyway: taking the same kind of test twice gives practice, and believing in a pill can improve effort.

    Without a control group, these effects are counted as if the supplement caused them.

  3. Subtract the placebo group's rise: (78−70)−(76−70)=8−6=2 points.

    The placebo group shows the gain from practice and expectation alone, so only the extra 2 points could be due to the supplement.

Answer: The supplement's effect is about 2 points, not 8; most of the rise came from practice and expectation, which the placebo group also showed.

Takeaway: A before-and-after gain is not an effect until a comparable control group's gain has been subtracted.

Where You'll See It

  • Plant fertilizer baseline

    To know if fertilizer helps, you grow some plants with no fertilizer (control) under the same conditions; the control shows how much they'd grow anyway, so you can measure the fertilizer's extra effect.

  • Does a new toothpaste reduce cavities

    One group uses the new toothpaste and a control group uses ordinary toothpaste; comparing cavity counts reveals whether the new formula does better than the everyday baseline.

  • Testing a focus playlist

    To see if a study playlist boosts test scores, a control group studies in silence under the same rules — without that baseline, you can't tell if the playlist or the studying caused any gain.

Common Mistakes

  1. The slip: You can judge a treatment's effect just by looking at the treated group's before-and-after change.

    Why it's tempting: A clear improvement in the treated group feels like proof the treatment worked.

    Instead: Without a control group you can't separate the treatment's effect from natural change, practice, or healing. The control shows what happens with no treatment.

  2. The slip: Any group that doesn't get the treatment is a good control group.

    Why it's tempting: If they're untreated, they seem to fit the role of control.

    Instead: A valid control group must be comparable to the treatment group in every other way (ideally formed by random assignment). A control that differs systematically reintroduces confounding.

  3. The slip: A control group means doing nothing at all to those subjects.

    Why it's tempting: 'Control = no treatment' is taken literally as no intervention.

    Instead: Controls often get a placebo or the standard/current treatment so they have the same experience minus the active ingredient — that keeps expectations and routines comparable.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. A coach has his 25-person track team do a new stretching routine for 6 weeks. Their average 400 m time drops by 1.8 seconds. What is missing before he can credit the routine?

    Hint: Would runners get faster over 6 weeks of practice anyway?

  2. In an acne trial, 42 of 60 patients using a new cream improve, and 30 of 60 using an identical cream with no active ingredient improve. Find each improvement rate and the difference.

    Hint: Turn each fraction into a percent and subtract.

  3. In a 4-week trial, 20 fertilized plants grow from a mean of 10 cm to 18 cm, and 20 control plants grow from 10 cm to 15 cm. How much of the growth can be credited to the fertilizer?

    Hint: Find each group's growth. The control group shows growth that happens without fertilizer.

  4. A tutoring study compares 30 students who signed up for tutoring with 30 students who did not. Why is the second group a weak control group?

    Hint: Were the two groups alike before tutoring began?

  5. In a headache trial, Group A gets a new pill and Group B gets nothing, and patients know their group. Is Group B a placebo group or a no-treatment control, and what problem could this cause?

    Hint: Did Group B receive anything that looks like a treatment?

  6. A student writes: "A school using the new math curriculum averaged 85, and a school in another state without it averaged 79. That school is our control group, so the curriculum works." Correct the reasoning.

    Hint: Were the two schools alike in every way except the curriculum?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Can a person serve as their own control?

Sometimes. In a crossover design, each subject tries both treatments in random order, say a new allergy spray for two weeks and a placebo spray for two weeks, with a break in between. Each person's results under one condition are compared with their own results under the other, which removes differences between people. It works only when the effect wears off during the break and the condition stays steady.

Is it fair to give sick people in a control group a fake pill?

When a proven treatment already exists, withholding it would be unethical, so the control group usually gets the standard treatment and the trial asks whether the new one does better. Placebo controls are used when no proven treatment exists or the condition is mild, and participants agree in advance knowing they may get a placebo. Ethics boards review these choices before a trial starts.

What is a historical control, and why is it weaker?

A historical control compares today's treated patients with patients treated in the past, instead of a control group studied at the same time. It is cheaper, but much else may have changed in between, such as how the illness is diagnosed, nursing care, or the kinds of patients coming in, and the new treatment gets credit for all of it. Trials with historical controls tend to make new treatments look better than randomized trials do.

Learning Path

Control Group

You are here

You need Experimental Design (comparing treatments fairly) and Random Assignment (forming groups by chance). The control group adds the baseline, what happens without the treatment, so the effect is the difference between groups rather than a before-and-after change. Next, Placebo Effect explains why controls often get a dummy treatment, and Blinding keeps expectations from tilting either group's results.

See Also