Skip to content
PSAT/NMSQT · Problem Solving and Data Analysis

Experimental design

Experimental design appears in 3 reviewed PSAT/NMSQT questions and sits in Problem Solving and Data Analysis. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Experimental design is tested

Example 1Easier
In an experiment, scientists randomly assign participants to either receive a new drug or a placebo. The group that receives the placebo is called the:
  1. AConfounding group
  2. BSample group
  3. CControl groupCorrect
  4. DExperimental group

Why C is right

In experimental design, the control group receives the placebo or no treatment. The experimental group receives the treatment being tested. Comparing the two groups isolates the effect of the drug.
Why the other 3 choices miss
A
Incorrect because a confounding variable is an outside factor that may influence results, not a group label. There is no group called the 'confounding group' in experimental design.
B
Incorrect because 'sample group' is not a standard experimental design term. Both the control and experimental groups together form the sample, but neither is specifically called the 'sample group.'
D
Incorrect because the experimental group is the one receiving the new drug (the treatment), not the placebo.
Example 2Medium
Researchers randomly assigned 60 volunteers with headaches into two groups. Group A took an herbal supplement and Group B took a placebo. After two weeks, Group A reported a 40% reduction in headache frequency compared to 15% in Group B. What conclusion is best supported?
  1. AThe herbal supplement likely caused greater headache reduction than the placebo in this experiment.Correct
  2. BThe herbal supplement causes a reduction in headaches in all people.
  3. CThe placebo had no effect on headache frequency.
  4. DHerbal supplements are more effective than all prescription drugs for headaches.

Why A is right

Because participants were randomly assigned to groups, a causal conclusion is supported — the supplement likely caused the greater reduction in headaches. However, the conclusion should be limited to this experiment; it cannot be generalized to all people.
Why the other 3 choices miss
B
Incorrect because the study cannot be generalized to 'all people' — only to people similar to the volunteers in this study. The random assignment supports causation but not unlimited generalization.
C
Incorrect because the placebo group did show a 15% reduction, meaning the placebo had some effect. The statement that it 'had no effect' is contradicted by the data.
D
Incorrect because the study does not compare herbal supplements to prescription drugs. No such comparison is made, so this conclusion goes far beyond the data.
Example 3Medium
A researcher designs a study to test whether a new fertilizer improves crop yield. She randomly assigns 40 identical plots of land to either receive the fertilizer or not, and measures each plot's yield. Which of the following features is most important for concluding that the fertilizer caused yield differences?
  1. APublishing results in a scientific journal
  2. BUsing a large number of yield measurements per plot
  3. CHaving a researcher observe and record each plot daily
  4. DRandom assignment of plots to treatment groupsCorrect

Why D is right

Random assignment of subjects to treatment groups is the key design feature that allows researchers to draw causal conclusions. It distributes confounding variables evenly between groups, isolating the effect of the treatment.
Why the other 3 choices miss
A
Incorrect because publication validates and disseminates research but is not a feature of study design. It does not address whether the fertilizer caused yield differences.
B
Incorrect because the number of yield measurements per plot increases precision but does not establish causation. The key to causation is random assignment, not the quantity of measurements.
C
Incorrect because daily observation and recording does not control for confounding variables. If plots are not randomly assigned to groups, observer bias or other factors could explain differences.

Studied alongside

Other concepts from Problem Solving and Data Analysis.