01 / 20

Experimental
Design

Master the principles of controlled experiments, learn to identify variables, and build the skills to analyze scientific studies on the ACT.

ACT Science

What Is Experimental Design?

Experimental design is the blueprint for a scientific investigation. It determines how data are collected, which factors are manipulated, and how results are interpreted. On the ACT Science section, roughly 6 out of 40 questions test your ability to evaluate the design of an experiment.

The Three Types of Variables

Independent Variable
The factor the experimenter deliberately changes between trials. Also called the "manipulated variable."
Dependent Variable
The factor that is measured as a result. Its value depends on the independent variable.
Controlled Variables
All factors kept the same across trials so they do not interfere with the results.

How Variables Connect

Independent Variable Experiment Controlled conditions held constant Dependent Variable Controlled Variables Kept the same across trials "What I change" "What I measure" Control Group: baseline with no treatment

The independent variable flows through the experiment (with controlled conditions held constant) and produces changes in the dependent variable.

The Role of the Control Group

A control group receives no treatment (or a placebo). It serves as a baseline for comparison: without it, you cannot determine whether the independent variable actually caused the observed change.

Experimental Group
Receives the treatment or manipulation of the independent variable.
Control Group
Receives no treatment (or a placebo). Everything else is identical to the experimental group.

Core Principle: One Variable at a Time

✓ Valid Experiment Change: Temperature Constant: Volume, pH, Time Measure: Reaction Rate Any change in rate is due to temperature alone. ✕ Flawed Experiment Change: Temperature Also Changed: pH, Volume Measure: Reaction Rate Cannot tell which factor caused the change!

Correlation vs. Causation

Correlation
Two variables change together, but one does not necessarily cause the other. A lurking variable might be responsible.
Causation
Changing the independent variable directly produces a change in the dependent variable, confirmed through a controlled experiment.

Sample Size and Replication

Larger sample sizes reduce the effect of random variation and produce more reliable results. Replication: repeating the experiment to verify results are consistent.

Small Sample (\( n = 3 \)) High variability, unreliable Large Sample (\( n = 30 \)) Tight cluster, reliable results

Analyzing an Experiment: Step by Step

1
Identify the purpose. What question is the experiment trying to answer?
2
Find the independent variable. What did the experimenter change between groups or trials?
3
Find the dependent variable. What was measured or observed as a result?
4
Check controlled variables. Were all other factors held constant? If not, the experiment is flawed.
5
Locate the control group. Which group had no treatment or the default condition?
6
Evaluate the conclusion. Does the data actually support the claim? Watch for overreach.

Worked Example: Plant Growth

Independent
Amount of water (mL/day)
Dependent
Plant height after 30 days
Controlled
Soil type, sunlight, temperature, pot size, plant species

Worked Example: Enzyme Activity

\( \text{Temperature } (^{\circ}\text{C}) \)
Reaction Rate 10 25 37 50 70
\( \text{Peak at } 37^{\circ}\text{C} \)
Enzyme denatures

Independent: temperature. Dependent: reaction rate. The enzyme works fastest at \(37^{\circ}\text{C}\) (body temperature) and slows dramatically at higher temperatures due to denaturation.

Worked Example: Confounding Variables

!
Problem 1: This is an observational study, not a controlled experiment. Students were not randomly assigned to eat or skip breakfast.
!
Problem 2: Confounding variables exist: students who eat breakfast may also sleep more, study harder, or come from more structured households.
✓
Correct conclusion: The data shows a correlation between breakfast and scores, but cannot establish causation without a randomized controlled experiment.

Common Mistakes on the ACT

✕ Confusing IV and DV
The independent variable is what the experimenter changes, not what is measured. If in doubt, ask: "Which variable did the scientist set before the experiment began?"
✕ Claiming Causation from Correlation
Observational data (surveys, population studies) can show trends, but only controlled experiments with random assignment can prove one factor causes another.
✕ Ignoring the Control Group
Every valid experiment needs a control group. Without it, there is no baseline to compare results against, and any conclusion is unsupported.

Bias and Validity

Selection Bias
Groups differ systematically before the experiment begins. Fix: random assignment to groups.
Placebo Effect
Subjects improve because they expect to improve, not from the treatment. Fix: use a placebo control and blinding.
Observer Bias
The researcher's expectations influence data recording. Fix: double-blind design (neither subject nor observer knows group assignment).

Summary Comparison Table

Controlled Experiment Observational Study
Manipulation Researcher changes the IV Researcher observes naturally
Control Group Yes, required No, not always present
Random Assignment Yes, subjects assigned randomly No, groups occur naturally
Can Establish Causation? Yes No, only correlation
Example Drug trial with placebo group Survey linking diet and health

Question 1 of 5

A scientist tests the effect of fertilizer concentration on plant growth by growing 5 groups of plants with different fertilizer amounts. All plants receive the same amount of sunlight, water, and soil. What is the independent variable?

Question 2 of 5

In an experiment testing whether a new drug reduces blood pressure, Group A receives the drug and Group B receives a sugar pill. Group B is best described as the:

Question 3 of 5

A study finds that people who exercise more tend to have lower resting heart rates. The researchers surveyed 500 adults but did not assign anyone to exercise programs. Which conclusion is best supported?

Question 4 of 5

A student tests whether music affects memorization. Group 1 studies in silence, Group 2 studies with classical music, and Group 3 studies with rock music. However, Group 3 also studies in a different room with brighter lighting. What is the primary flaw?

Question 5 of 5

Which of the following changes would most improve the reliability of an experiment that tested only 5 subjects per group?