Three Parts of an Experiment
In Experiments and Why They Can Show Cause, you learned that an experiment deliberately imposes treatments on experimental units and measures responses. The next step is to identify exactly what those words refer to in a description of a study. This tutorial focuses on three connected terms: experimental unit, factor, and treatment.
A useful way to read an experiment is to ask three questions in order: What receives an assigned condition? What feature is deliberately changed? Which specific condition does each unit receive? Keeping those questions separate helps prevent a common mix-up: the people in an experiment are not the treatments, and a treatment is not the measurement researchers record afterward.
For a study comparing two study methods, the students who receive the methods are the experimental units. The study method is the factor. The particular methods being compared are the treatments. The quiz score measured afterward is an outcome of the experiment, not a treatment. The next tutorial develops the distinction between variables used to explain and variables recorded as outcomes.
From Factor to Treatment
A factor describes what is varied; its levels describe the possible settings or categories of that factor. In a study comparing two study methods, the factor might be “study method,” with the levels “guided practice questions” and “rereading notes.” Each level is one condition the researchers could assign.
When an experiment has just one factor, each treatment is one of that factor’s levels. A treatment is the actual condition assigned, while the factor is the broader feature being varied. For example, “type of practice activity” is a factor; “practice questions” and “flashcards” are its levels and, in this one-factor experiment, its treatments.
Some experiments have more than one factor. In that case, a treatment can be a particular combination of levels, one from each factor. Suppose researchers vary both study method and study location. “Practice questions at a quiet desk” and “flashcards at a quiet desk” would be two different combinations of conditions. It is still useful to identify each factor separately before listing the treatments.
Do not assume the experimental unit must be a person. It could be a seed, a manufactured item, a classroom, or another individual unit that receives a treatment. The key is to identify what the researchers assign a condition to. If a whole classroom is assigned one study method, for example, the classroom—not each student separately—is the experimental unit for that assignment, even though individual students may later have measurements recorded.
Worked Example: Comparing Two Study Methods
Worked Example: Practice Questions or Rereading?
A fictional teacher recruits 36 students for a study of two ways to prepare for a vocabulary quiz. The teacher assigns 18 students to complete a set of practice questions and assigns the other 18 to reread a set of notes. The next day, every student takes the same quiz.
Identify the experimental units: The individual students are the experimental units. The teacher assigns a study condition to each student, so the students are the units receiving treatments. Because the units are people, they may also be called subjects.
Identify the factor: The factor is the study method. It is the feature deliberately varied by the teacher.
Identify the treatments: The two treatments are completing practice questions and rereading notes. These are the two levels of the study-method factor, and because this experiment has one factor, each level is one treatment.
Separate the outcome from the treatment: The vocabulary quiz score is recorded after the treatment. It is a measured outcome, not one of the study methods and not a treatment.
Conclusion: The units are the 36 students; the factor is study method; and the treatments are practice questions and rereading notes. The number assigned to each treatment tells us how many students receive it, not how many treatments or factors the experiment has.
A Quick Map Keeps the Terms Straight
For a short study description, make a simple map before writing a conclusion. A table is especially helpful when the treatments look similar or when a scenario includes a measured result that could be mistaken for a treatment. Put the assigned units in one column, the deliberately varied factor in another, and the treatment conditions in a third.
| Question to ask | Term | Study-method example |
|---|---|---|
| What receives an assigned condition? | Experimental unit | Each student |
| What feature is deliberately varied? | Factor | Study method |
| Which specific condition is assigned? | Treatment | Practice questions or rereading notes |
| What is recorded afterward? | Measured outcome | Vocabulary quiz score |
The last row is included as a contrast: identifying a measured outcome can help you avoid calling it a treatment. In this example, the quiz does not get assigned to students as a study condition; it is used to record information after the students have used their assigned methods.
Worked Example: When the Unit Is a Group
Worked Example: Two Review Formats in Classrooms
A fictional school compares two review formats before a science assessment. Researchers include 8 classrooms, with 4 assigned to a teacher-led review and 4 assigned to an independent worksheet review. Students in each classroom take the same assessment afterward.
Identify the experimental units: The classrooms are the experimental units because the researchers assign a review format to each whole classroom. There are 8 units receiving treatments. Although students take the assessment, the assignment described in the study is made to classrooms, not separately to each student.
Identify the factor: The factor is review format, the feature researchers deliberately vary.
Identify the treatments: The treatments are teacher-led review and independent worksheet review. These are the two levels of the review-format factor.
Identify what is not a treatment: The assessment score is recorded after the review. It is an outcome measurement, not the condition assigned to a classroom.
Conclusion: The unit receiving the condition determines the experimental unit. Here it is the classroom, even though the assessment produces scores for individual students. A response should not call each student an experimental unit unless the study actually assigns treatments to students individually.
Worked Example: Several Factors, Combined Treatments
Worked Example: Study Method and Study Location
A fictional learning center works with 24 students. Each student is assigned one of two study methods—flashcards or practice questions—and one of two study locations—a quiet room or a room with typical background conversation. The center records each student’s score on a later practice quiz.
Identify the experimental units: The individual students are the experimental units because each student receives an assigned study method and location.
Identify the factors: There are two factors: study method and study location. Each is deliberately varied by the researchers.
List the treatments: A treatment combines one level of each factor. The four treatments are flashcards in a quiet room, flashcards in a room with background conversation, practice questions in a quiet room, and practice questions in a room with background conversation.
Check the count: There are 2 study-method levels and 2 location levels, so the full set of combinations contains \(2 \times 2=4\) treatments. Each treatment is a distinct condition, even though some conditions share the same method or location.
Conclusion: The students are the units, the two deliberately varied features are the factors, and the four method-and-location combinations are the treatments. The quiz score is recorded after the conditions are assigned; it is not another treatment.
Worked Example: Identify the Parts in a Non-School Setting
Worked Example: Two Seed-Starting Conditions
A fictional garden club places 40 tomato seeds in individual pots. It assigns 20 seeds to a standard potting mix and 20 seeds to the same mix with a compost addition. After three weeks, members record whether each seed has produced a sprout.
Identify the experimental units: Each seed is an experimental unit because the researchers assign a growing condition to each seed in its own pot.
Identify the factor: The factor is potting-mix condition—the feature deliberately changed.
Identify the treatments: The treatments are standard potting mix and potting mix with compost. They are the two levels of the factor in this experiment.
Separate assignment from measurement: Whether a seed has produced a sprout after three weeks is recorded after the growing condition is assigned. It is an outcome, not a treatment. The 20 seeds in each group are counts of units assigned to each condition; they do not change the number of factor levels.
Conclusion: The 40 seeds are the experimental units, potting-mix condition is the factor, and the two mix conditions are the treatments. This identification works even though the units are objects rather than people.
Common Mistakes and AP Exam Tips
- Calling the people the factor. People usually serve as experimental units or subjects. The factor is the feature researchers deliberately vary, such as the study method assigned to those people.
- Calling the measured result a treatment. A quiz score or sprouting outcome is recorded after a condition is assigned. Name the condition as the treatment and the recorded result as an outcome.
- Listing a broad factor instead of its treatments. “Study method” names the factor, not the specific treatments. A complete answer lists the methods being compared, such as practice questions and rereading notes.
- Assuming each person is always an experimental unit. Identify the level at which the condition is assigned. If an entire classroom receives one format, the classroom is the unit for that assignment.
- Missing combinations when there is more than one factor. Name each factor and its levels, then list the distinct combinations that make up the treatments. Do not list only the levels of one factor.
- Using “treatment group” as if it meant the treatment itself. A treatment is the condition; a treatment group is the set of experimental units assigned to that condition. For example, “practice questions” is a treatment, while the 18 students assigned to practice questions form its group.
For full credit, make each identification explicit and connect it to the scenario: “The students are the experimental units because each student is assigned a method. Study method is the factor, and practice questions and rereading are the two treatments.” This gives the term and the evidence rather than relying on a label alone.
Check Your Understanding
For each scenario, identify the experimental units, factor or factors, and treatments. Explain which detail in the scenario supports your identifications.
- A teacher assigns 30 students to either a vocabulary app or paper flashcards and records their scores on a later quiz. Identify the units, factor, and treatments.
- A researcher assigns 6 entire classrooms to one of two warm-up activities. Students then complete a short math assessment. What are the experimental units, and why?
- A fictional greenhouse assigns individual seedlings to either full sunlight or partial shade. Name the factor and its treatments. Is the recorded number of leaves a treatment?
- A study varies both the length of a study session (20 minutes or 40 minutes) and the study method (notes or practice questions). How many treatment combinations are possible? List them.
- A club assigns 12 members to a guided review and 12 members to independent practice. It records how many questions each person answers correctly. Distinguish the factor, treatments, experimental units, and recorded outcome.