Skip to content
AP Psychology · Mental and Physical Health

Third variable

Third variable appears in 9 reviewed AP Psychology questions and sits in Mental and Physical Health. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

A third variable is an outside factor that affects both of two measured variables. It makes them look linked even when neither one causes the other. This is the main reason a correlation cannot prove causation.

Why it matters

Many health headlines rest on correlations, such as "more sleep, better grades." Spotting a third variable stops you from accepting a causal claim the data cannot support.

How the exam tests it

A typical stem describes a survey that finds two things go together, and then someone claims one causes the other. You pick the answer that names the error and a plausible third variable. For example, long hours at an after-school job could cut both sleep and grades.

Don’t confuse it with

Students confuse a third variable with reverse causation, but a third variable is a separate factor driving both, while reverse causation means the supposed effect is actually the cause, such as anxiety driving more social media use.

Third variable in real questions

Example 1Medium
A high school surveys 600 students and finds that those who report sleeping at least 8 hours on school nights have higher grade point averages than those who report sleeping less. The principal concludes that getting more sleep causes students to earn better grades. Which thinking error does this conclusion MOST clearly illustrate?
  1. AIllusory correlation, because the principal perceives a link between sleep and grades that does not actually exist
  2. BInferring causation from correlation, because a third variable such as after-school job hours could affect both sleep and gradesCorrect
  3. COperational definition error, because hours of sleep and grades were never stated in precise, observable, measurable terms before the survey began
  4. DSampling bias, because students who choose to report their sleep habits may not represent the school's whole student body

Why B is right

The survey is correlational: it measured sleep duration (Unit 1, Topic 1.5 Sleep) and GPA as they naturally occurred, without randomly assigning students to sleep more or less. Correlation does not establish causation. A third variable, such as long after-school work hours or heavy family responsibilities, could both shorten sleep and lower grades, and the direction could also run the other way (students struggling academically may stay up late studying or worrying). Only an experiment that manipulates sleep and randomly assigns students to conditions could support the principal's causal claim.
Why the other 3 choices miss
A
Incorrect because an illusory correlation is perceiving a relationship that is not actually present in the data; here the survey did find a real relationship between sleep duration and GPA, and the error is the causal conclusion drawn from it.
C
Incorrect because self-reported hours of sleep and GPA are concrete, measurable variables; the flaw is not how the variables were defined but the leap from a correlational survey to a cause-and-effect claim.
D
Incorrect because even a perfectly representative sample would still yield only correlational data; the conclusion's error is inferring causation, not a problem with who was surveyed.
Example 2Medium
A psychologist notices that teenagers who spend more hours on social media report higher levels of anxiety. She concludes that social media use causes anxiety. Which major flaw undermines this conclusion?
  1. AShe should have operationalized anxiety using behavioral observation rather than self-report measures
  2. BHer sample of teenagers lacks population validity because teenagers are not representative of all social media users
  3. CShe should have used a within-subjects design to control for individual differences among the teenagers
  4. DCorrelation does not imply causation; a third variable such as pre-existing stress could explain both behaviorsCorrect

Why D is right

The fundamental limitation of correlational research is that it cannot establish cause and effect. Even if two variables are strongly correlated, a third variable (confound) could be causing both, or the direction of causation could be reversed. Here, pre-existing anxiety could drive social media use, or a third factor like family conflict could produce both.
Why the other 3 choices miss
A
Incorrect because while self-report measures have limitations, the core flaw is inferring causation from correlation, not the measurement instrument used.
B
Incorrect because population validity is a concern for all research, but it is not the specific flaw in concluding causation from a correlational finding.
C
Incorrect because a within-subjects design addresses order effects and individual differences in experimental research; the issue here is that correlational research cannot establish causation regardless of design type.

Studied alongside

Other concepts from Mental and Physical Health.