AP Statistics · Unit 1 · Exploring One-Variable Data and Collecting Data
1.13 Experimental design
Describe the key elements of a well-designed experiment, compare completely randomized and blocked designs, and explain why random assignment supports cause-and-effect conclusions.
1/6An experiment assigns treatments to study cause and effect
- An experiment assigns treatments to study cause and effect
- Control keeps other influences from distorting the comparison
- Random assignment makes the treatment groups comparable
- Replication and blinding make the results more trustworthy
- Blocking separates a known source of variation
- Random assignment supports cause-and-effect conclusions
- Check your recall
- Practice: 10 questions
Part 1 of 6: An experiment assigns treatments to study cause and effect
A teacher wants to know whether a new study app raises quiz scores. She could compare students who already use it with students who do not. That comparison only observes what happened.
In an experiment, researchers assign a condition to each unit and then measure an outcome. Each condition is a treatment. Each individual that receives a treatment is an experimental unit. When the units are people, they are often called subjects.
Key terms · tap one for its definition
experiment
An experiment is a study in which researchers assign treatments to experimental units and then measure a response.
treatment
A treatment is a specific condition that researchers assign to the experimental units in an experiment.
experimental unit
An experimental unit is a single individual, such as a person or plant, to which a treatment is assigned.
Part 2 of 6: Control keeps other influences from distorting the comparison
Suppose the app group scores higher. Something other than the app could explain the gap. Students who choose the app may simply study harder.
That hidden factor is a confounding variable. It is related to the treatment and also affects the response, so their effects cannot be separated. Control is the main defense. Researchers compare treatments while holding other conditions the same.
A control group gives a baseline for comparison. Many studies give this group a placebo, an inactive treatment that looks real. It keeps expectations about improvement equal across the groups.
Key terms · tap one for its definition
confounding variable
A confounding variable is a variable related to the explanatory variable that also affects the response, so their effects cannot be separated.
control group
A control group is a group of experimental units that receives no treatment, a placebo or a standard treatment, for comparison.
placebo
A placebo is an inactive treatment designed to look and feel like the real treatment.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: B · Worked explanation
Part 3 of 6: Random assignment makes the treatment groups comparable
Control handles differences that researchers can see. Chance must handle the rest. In random assignment, a chance process decides each unit’s treatment. Examples are drawing names or using a random number generator.
Because chance does the sorting, the groups should be alike before treatment. This holds even for variables nobody measured. A large difference afterward is unlikely to come from how the groups started.
A completely randomized design assigns all units to treatments at random, with no grouping first. For two treatments, the steps are:
- Number the units.
- Use a random device to pick half of them for the first treatment.
- Give the remaining units the second treatment.
- Measure the response for every unit.
Key terms · tap one for its definition
random assignment
Random assignment is the use of a chance process to decide which treatment each experimental unit receives.
completely randomized design
A completely randomized design is an experiment in which all experimental units are randomly assigned to the treatments.
Part 4 of 6: Replication and blinding make the results more trustworthy
One unit per treatment would prove very little, even with random assignment. Good designs rely on replication, which means using many experimental units in each group. Chance differences between individuals then tend to average out, so a real treatment effect is easier to see.
People’s beliefs can also change results. Researchers use blinding to keep subjects, or those measuring the response, from knowing who received which treatment. In a double-blind study, neither group knows.
Key terms · tap one for its definition
replication
Replication is the use of enough experimental units in each treatment group to reduce the effect of chance variation.
blinding
Blinding is keeping subjects, those measuring the response, or both unaware of which treatment each unit receives.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: A · Worked explanation
Part 5 of 6: Blocking separates a known source of variation
Sometimes a variable is known to affect the response strongly. Random assignment might leave the groups unbalanced on it by chance. A block is a group of experimental units that are alike in a way expected to affect the response.
In a randomized block design, researchers first sort the units into blocks. They then assign treatments at random within each block. Treatments are compared among similar units, so variation between blocks does not cloud the comparison.
Two designs compared
- Feature
- First step
- Completely randomized
- Number all units
- Randomized block
- Sort units into blocks
- Feature
- Random assignment
- Completely randomized
- Across all units
- Randomized block
- Within each block
- Feature
- Best when
- Completely randomized
- Units are similar
- Randomized block
- A known variable affects the response
- Feature
- Variation from the blocking variable
- Completely randomized
- Left to chance
- Randomized block
- Removed from the comparison
Key terms · tap one for its definition
block
A block is a group of experimental units that are similar in a way expected to affect the response.
randomized block design
A randomized block design is an experiment in which units are sorted into blocks and treatments are randomly assigned within each block.
Part 6 of 6: Random assignment supports cause-and-effect conclusions
We can now see why an experiment supports a stronger claim than observation. Random assignment spreads confounding variables across the groups by chance. If the groups differ by more than chance alone would likely produce, the treatment is the most reasonable explanation. Without random assignment, a difference might come from the groups themselves rather than the treatment.
The conclusion applies to the units in the study. Random sampling, which uses chance to choose the units, supports generalizing to a larger population. Random assignment supports cause and effect. Random sampling supports generalization.
Worked example
Describing a complete design
- Treatments: the new fertilizer and no fertilizer, which serves as the control group.
- Block by bench, because light is expected to affect yield.
- Within each bench, randomly assign 10 plants to each treatment, for example by drawing numbered tags.
- Give every plant the same soil and watering so other conditions are controlled.
- Measure the yield of every plant and compare the treatments.
Answer: A randomized block design with two blocks, random assignment within each block, and a comparison of yield between the treatments.
Check your recall[13 marks]
5 written questions. Answer each one in your head or on paper, then open the mark scheme.
- 1.Explain[2 marks]
Explain why an experiment often includes a control group.
Show mark schemeHide mark scheme
Model answer
A control group gives a baseline for comparison. Differences in response can then be linked to the treatment rather than to other influences.
- 1 mark for stating that the control group provides a baseline for comparison
- 1 mark for linking differences in response to the treatment rather than other influences
- 2.Describe[3 marks]
Describe how to carry out random assignment in a completely randomized design with 30 subjects and two treatments.
Show mark schemeHide mark scheme
Model answer
Number the subjects from 1 to 30. Use a random device to pick 15 subjects for the first treatment. The remaining 15 receive the second treatment.
- 1 mark for numbering or listing the subjects
- 1 mark for using a random device to choose 15 subjects for one treatment
- 1 mark for giving the other 15 subjects the second treatment
- 3.Explain[3 marks]
Explain why random assignment allows cause-and-effect conclusions.
Show mark schemeHide mark scheme
Model answer
Random assignment makes the treatment groups alike before treatment, including in variables that were not measured. Confounding variables are then spread across the groups by chance. A difference larger than chance would likely produce can be attributed to the treatment.
- 1 mark for stating that random assignment makes the groups similar before treatment
- 1 mark for explaining that confounding variables are spread across the groups by chance
- 1 mark for concluding that a difference larger than chance would likely produce is attributed to the treatment
- 4.Describe[2 marks]
Describe one advantage of a randomized block design over a completely randomized design.
Show mark schemeHide mark scheme
Model answer
Units are sorted into blocks that are alike on a variable that affects the response. Treatments are compared within those blocks, so variation between blocks does not hide the treatment effect.
- 1 mark for stating that units are grouped into blocks by a variable expected to affect the response
- 1 mark for explaining that treatments are compared within similar units, removing variation between blocks
- 5.Justify[3 marks]
A researcher compares two training plans using 40 runners, 20 beginners and 20 experienced. Justify blocking by experience level.
Show mark schemeHide mark scheme
Model answer
Experience probably affects the response, such as running time. Blocking by experience compares the plans among runners who are alike. Random assignment within each block keeps the comparison fair.
- 1 mark for stating that experience is expected to affect the response
- 1 mark for explaining that blocking compares plans within similar runners and removes variation between blocks
- 1 mark for mentioning random assignment of plans within each block

Practice this topic
10 quick-fire multiple-choice questions, easy to hard. Tap an answer to see if you're right and the worked explanation. (The written questions with mark schemes are under “Check your recall” above.)
0 right0 of 10 answered
Includes the 2 quick checks in the note above.
EasyIdentify[1 mark]
Show answer and explanation
Answer: C · Worked explanation
MediumIdentify[2 marks]
Show answer and explanation
Answer: C · Worked explanation
MediumDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation
MediumCalculate[2 marks]
Show answer and explanation
Answer: B · Worked explanation
Medium[2 marks]
Show answer and explanation
Answer: C · Worked explanation
MediumIdentify[2 marks]
Show answer and explanation
Answer: C · Worked explanation
HardDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation
HardIdentify[2 marks]
Show answer and explanation
Answer: C · Worked explanation