Correlation Coefficient in AP Psychology

Step 1: Spot the correlational design behind an r value

An r value usually comes from a correlational study, where researchers measure two variables as they naturally occur and don't manipulate either one. If a study summary says participants reported their hours of sleep and their anxiety scores, and the results give an r, that's correlational research.

This term isn't tied to one content topic. It belongs to the course's research methods and data skills, which apply to every unit. Determining the research design (skill 2.A) and evaluating non-experimental methods (skill 2.C) both start with noticing that nothing was manipulated and no one was randomly assigned.

That identification counts in the Article Analysis question, which asks for the study's method and variables. If the researchers only measured the variables and did not manipulate one, identify the study as correlational. Don't infer the overall design from r alone, since an experiment can also report correlations between measured variables. For a correlational study, don't name an independent and dependent variable as if it were an experiment.

Step 2: Read the sign of r for direction

A positive r means the two variables rise and fall together. For example, more hours spent studying go with higher test scores.

A negative r means one variable goes up as the other goes down. For example, more hours of screen time before bed go with fewer hours of sleep.

The sign only shows direction. A negative correlation isn't a "bad" or weak relationship. Don't mix up "negative" with "weak" on data-reading questions.

Step 3: Measure strength by distance from zero

Strength depends on how close r is to -1 or +1, whatever the sign. So -0.85 is stronger than +0.40, because 0.85 is farther from zero. An r near 0 means there's little or no linear pattern.

On a scatterplot, a strong r looks like dots packed tightly around a straight line. A weak r looks like a loose cloud with only a slight tilt. Multiple-choice questions may give you a graph and ask which r value fits it. They may also list several values and ask which shows the strongest relationship. Find the number with the largest size and ignore the sign.

Calculators aren't allowed on this exam, so interpreting r is the skill that matters here (skills 3.A and 3.C). Don't confuse r with the [[Mean]] or [[Standard Deviation]]. Those describe a single variable's center and spread. r describes how two variables relate. A correlation can also be statistically significant and still weak, especially in a large sample. Statistical significance indicates that a result would be unlikely if a specified null hypothesis were true. It doesn't measure how strong the relationship is.

Step 4: Check that the relationship is a straight line

r only captures linear relationships. If two variables have a curved relationship, r can land near 0 even when the variables are clearly connected. Arousal and performance are a good example. Performance tends to be best at moderate arousal and worse at both low and high arousal. That inverted-U shape can produce a small r.

An outlier can also push r up or down a lot, especially in a small sample. If a question shows a scatterplot, look at its shape before trusting the number.

Step 5: Stop at "related" and never write "causes"

This distinction is important. A correlation can't establish causation for two reasons:

• Third-variable problem: something else may drive both variables. Ice cream sales and drowning rates rise together because hot weather increases both.
• Directionality problem: even if one variable does affect the other, the correlation can't show which direction it runs. Depression and poor sleep are correlated, but the data alone can't tell which comes first.

Only an experiment, with a manipulated independent variable and random assignment, supports a cause-and-effect claim. This holds even for a perfect r of +1.00.

If an Article Analysis study reports a correlation, interpret its direction and strength as the question asks, for example "a moderate positive correlation between X and Y." Don't infer causation from the correlation alone. The same distinction can matter when you evaluate evidence in the Evidence-Based question. A correlational source can support a claim that two variables are related, but claiming that one causes the other would overreach, and the reasoning should say so (skill 4.B). Write "is associated with" or "is correlated with." Avoid "leads to," "lowers," or "increases."